Now that the Wikimedia software supports alt attributes in images, and the WP:ACCESSIBILITY and WP:ALT guidelines recommend alt text, should the featured article criteria also mention this? Something like the following change, perhaps?
This isn't a well-watched page; perhaps you should bring this up at WT:FA or WT:FAC. For now, in the absence of consensus, I'm going to revert you, although I really have no opinion either way on the matter. —Ed(Talk • Contribs)21:48, 23 June 2009 (UTC)[reply]
Now that the Wikimedia software supports alt attributes in images, and the WP:ACCESSIBILITY and WP:ALT guidelines recommend alt text, I propose that the featured article criteria should also mention this, with the following change (insertion underlined):
The problem is... what constitutes useful alt text? And in some cases, wouldn't specifying good alt text verge into OR? We already specify captions, which is the default, as I understand WP:ALT. I'm not against the idea, although I have to admit adding yet another bit of stuff that must be jumped through for images just makes me want to use less images. Ealdgyth - Talk23:01, 23 June 2009 (UTC)[reply]
WP:ALT, the wikilink in the proposed change, describes what constitutes useful alt text. The default alt text is blank—it doesn't default to the caption, as captions and alt text are for different purposes and captions often make for bad alt text. You can check this out by visiting today's featured article Major depressive disorder. Its first illustration Image:Vincent Willem van Gogh 002.jpg has a caption "Vincent van Gogh's 1890 painting At Eternity's Gate" which is not that useful to a blind reader (unless they already know the art work in question, which is highly unlikely). The image's alt text is blank. Its title text is merely "Vincent Willem van Gogh 002.jpg", which is even less useful than the caption. Useful alt text (for the blind, etc.) would be something like "Man dressed in blue and seated in wooden chair, leaning forward with his face buried in his hands." Eubulides (talk) 23:20, 23 June 2009 (UTC)[reply]
Joys. More stuff to jump through hoops for images. (I know, I know, it's important, but it's already a major hassle to deal with images). HOw in the Heck am I supposed to explain an image like the lead map in Gregorian mission? Or the various family trees there? Ealdgyth - Talk23:23, 23 June 2009 (UTC)[reply]
I really disagree with this. Sure, add alt text if it's useful, but the whole point of images on most articles is that they illustrate aspects that would be hard to explain in text form. For something like Chelsea Bridge, the alt text for most of the images would be as long or longer than the accompanying paragraph. – iridescent23:43, 23 June 2009 (UTC)[reply]
As a screen reader user, I think alt text should be encouraged, but not required on every image. If there's no alt text in math equations, screen readers will read out the raw LaTeX, which is never easy to parse. Long alt text is hard to read for me; a screen reader says "link graphic" before each line. Therefore alt text should be succinct; Eubulides' alt text for At Eternity's Gate is a good length. However, a picture paints a thousand words, and there are some images which probably can't be described succinctly, like maps or photos. Sometimes I find that all I want is a tactile diagram of an image, so I can properly understand it. Graham8701:08, 24 June 2009 (UTC)[reply]
"HOw in the Heck am I supposed to explain an image like the lead map in Gregorian mission?" It's not that hard. The alt text does not need to explain everything about the image, only its gist. I just now added alt text to every image in Gregorian mission; for the lead map the alt text is "Map of England and Wales. Britons are in the west, Northumbrians in the north, Mercians in the center, Saxons in the south, and Angles in the middle east. Many smaller groups are present." It took me about 20 minutes to view all the article's images and write alt text for each one.
"Or the various family trees there?" Sorry, I didn't see any family trees in that article.
"the whole point of images on most articles is that they illustrate aspects that would be hard to explain in text form" It's true that alt text is typically a poor substitute for an image; but alt text is much better than nothing, for people who can't see the image.
"For something like Chelsea Bridge, the alt text for most of the images would be as long or longer than the accompanying paragraph" No, the alt text doesn't have to be that long. I just now added alt text to Chelsea Bridge. Again, it wasn't that hard (another 20 minutes). In some cases the alt text was shorter even than the caption. I don't see any case where the alt text was longer than the accompanying paragraph.
In summary, let's please not overestimate the amount of effort it takes to add alt text to an article's images. Adding alt text is a tiny, tiny fraction of the work needed to write a featured article.
Wow! Good work. I checked out the Chelsea Bridge article; it was interesting to hear descriptions of things that sighted people would take for granted, such as the Battersea Shield. I made a tweak to the alt text to make it more readable with speech synthesizers. I agree with sticking to the most important details about an image in the alt text; I don't want information overload, just a rough mental idea of what's in an image. Graham8712:29, 24 June 2009 (UTC)[reply]
I am in favor of this improvement to our criteria. Note that our image criterion is extremely low - we require "professional" writing and "well-researched" articles, but we only require properly licensed media (we, apparently, could care less about their quality or presentation). This is just one step towards what should ultimately be our goal: bringing criterion 3 in line with the others! Awadewit (talk) 12:10, 24 June 2009 (UTC)[reply]
I am very much against insisting on alt text as part of the FAC criteria. As other have said, images are quite enough of a PITA already. --MalleusFatuorum12:15, 24 June 2009 (UTC)[reply]
Would this invalidate nonfree images, since they are being used to illustrate a concept that words are insufficient in doing, if words are used in the alt text, does that not pretty much prove that words do illustrate the concept in the image? --Moni3 (talk) 12:44, 24 June 2009 (UTC)[reply]
Not really: the alt text is meant to describe the image but does not need to apply critical commentary or the like to justify the image; that needs to be done in the text and caption. That said, this may make it clear for a handful of NFC cases that the image may be redundant to text, requiring further justification for its use - given the typical uses of NFC, this is a small minority of cases and nothing requiring a major review of all FAs to validate their images again. --MASEM (t) 14:10, 24 June 2009 (UTC)[reply]
I didn't say it's not going to happen - but it's not going to be a huge impact on the FA process already where image inclusion is already highly scrutinized; it might impact, what, 5% of the images that pass through the process? And even though, it is a good call for FA submitters to find something more to justify borderline images. Say, for example, the image is a tableau of characters from a movie - say File:Reservoirdog.jpg (This is not a featured article, just an example that came to mind). The alt text for that image would be easy: "The cast of Reservoir Dogs, from left to right..." and without any other comment, is replaceable by text. However, in this case, that tableau and that specific scene is part of the iconic nature of the movie which is further described by the article text and something that cannot be described in the alt text without it itself becoming a novel. Thus, while the alt text technically replaces the image to the point that a screen-reader user can understand what's in the image, the image would still meet NFC because there's additional commentary on it.
In other words, by requiring alt text for FAs, it will cause more images to be scrutinized, but at the same time in order to meet that scrutiny, that will likely further improve reasons why certain NFC images are used and strength their non-free rationale. --MASEM (t) 14:29, 24 June 2009 (UTC)[reply]
I can't imagine a case where an image that would otherwise satisfy WP:NFC would be rejected simply because alt text was added to it. Alt text is almost always a poor substitute for the image; alt text is useful only because it is much better than nothing for readers who can't see the image. Eubulides (talk) 16:01, 24 June 2009 (UTC)[reply]
I support this proposal wholeheartedly. Some issues of ALT text at FAC were discussed in December 2008 at Wikipedia_talk:Manual_of_Style/Archive_106#ALT_text and the idea of requiring ALT text for all math-mode equations was discussed in January 2009 at Wikipedia_talk:Accessibility#Image_alt.2Fcaption_confusion. To help people, I wrote a script that checks for ALT text in images which you can install by adding "importScript('User:Proteins/checkALTtext.js');" to the monobook.js subpage under your user name, e.g., User:Proteins/monobook.js. In the recently promoted mathematical FA, Euclidean algorithm, we avoided math-mode equations wherever possible and gave ALT text for the remainder. Tim Vickers and I added ALT text to all the images of Acid dissociation constant. Based on that experience, I expect that it should be feasible for other articles, if editors are willing. Proteins (talk) 08:15, 25 June 2009 (UTC)[reply]
I certainly agree that the addition of alt-text to acid dissociation constant was an improvement, and that the improvement could go further with some guidelines for users. In chemistry, for example, we would probably need some sort of formal ontology to avoid duplicated work. I'm not sure what this discussion is doing here: surely it is an issue which needs to be discussed more widely in the various editing communities which would be affected. Or is this just a suggestion for FAs and to hell with the rest of the encyclopedia? Physchim62(talk)12:39, 25 June 2009 (UTC)[reply]
The discussion for the whole encyclopedia has been done (see WP:ACCESSIBILITY and WP:ALT). This discussion is just whether it's helpful to remind and/or educate editors about this for FAs. Also, FAs often serve as examples for the rest of the encyclopedia. Eubulides (talk) 14:52, 25 June 2009 (UTC)[reply]
meta-comment (which probably belongs at ALT, but posting here for now since this is a wider audience) If having no alt text isn't good for accessibility, surely defaulting the alt text to be the caption when no alt text is specified should be the default fall-back, no? I presume this wouldn't be hard to achieve technologically. TwilligToves (talk) 06:04, 26 June 2009 (UTC)[reply]
The Philosopher (c. 250–200 BC) from the Antikythera wreck illustrates the style used by Hecataeus in his bronze of Philitas.
The Philosopher (c. 250–200 BC) from the Antikythera wreck illustrates the style used by Hecataeus in his bronze of Philitas.
No, because screen readers etc. already have access to the caption. There is no point to duplicating the caption, so it's better to have no alt text than to have it default to the caption. To the right of this comment I've placed a simulation of the useless repetition inherent in what an image would "look" like to a screen reader, if alt text defaulted to the caption. Contrast this with the more-useful example in WP:ALT #What people see with images turned off. Eubulides (talk) 06:29, 26 June 2009 (UTC)[reply]
I am reluctant to see use of alt-text being made mandatory, not because I am opposed to the goal of making wikipedia more accessible, but because I often wouldn't know how to appropriately describe an image without, at least partially, repeating the contents of the caption (I tried this for the images at India as a mental exercise). I assume that I am not alone on this.
I would therefore suggest that instead of trying to "force" alt-text use by fiat, it would be more useful if a wikiproject was set up add alt-text to current vital and Featured articles, while encouraging FA candidates to use alt-text too. That way we will have useful set of well-written exemplars of alt-text descriptions for maps, portraits, chemical formulas, plant, flowers, animals, album and book covers, logos, anatomical images, micrographs, color palettes, etc, which can then be used as rough templates by all editors. Such an exercise would also help establish the level of recommended detail and the boundaries of permitted OR and referencing requirements. Finally, the wikirpoject would serve as a useful forum for editors to ask for help when they (inevitably) run into problems or experience a writer's block. Once we have a large enough corp of editors familiar and comfortable with alt-text descriptions, we can make their use mandatory in featured content, in a few months time. Abecedare (talk) 07:37, 26 June 2009 (UTC)[reply]
It's OK on occassion to partly repeat the caption, when the caption describes what the image looks like. For India which images did you find hard to describe? I just now added alt text to all of India's images: it took maybe 30 minutes, and I didn't run into many problems. In my mind I briefly described each image over the telephone to a friend; then I wrote down what I said. Alt text shouldn't be an exhaustive description; it should merely say the gist of an image. Eubulides (talk) 08:58, 26 June 2009 (UTC)[reply]
Thanks, such examples are useful. Since you asked, the images I had problem thinking descriptions for were:
File:Aero-Sukhoi1.JPG, for which I was thinking of "grey coloured military plane ascending against a blue sky", which turned out to be not too different from what your choice.
File:BSE.jpg for which I only had, "a tall building viewed from street-level", which sounded inadequate, but I guess that's better than nothing.
Image:IPL T20 Chennai vs Kolkata.JPG for which I had a much longer description in mind, mentioning the batsman's pose and the visible fielding positions (obviously I have heard too much cricket commentary :) )
Image:India Geographic Map.jpg, I noticed that you mentioned the colors in the topographical map, but without the key, the colors are not meaningful at all. Still not sure what the appropriate description would be ("mountainous region bordering China ... higher altitudes in the central plains ..." ?).
Also I had planned to include the name of Nehru, Gandhi, Manmohan Singh and Medvedev in the alt-text ("Gandhi dressed in a dhoti..."), which would then have been partially redundant with the caption; I noticed that you avoided that.
Anyway, coming back to the larger point: what do you think about my wikiproject proposal above ? Since the FA criteria applies not only to FA candidates but also to current FAs, it would be good to deal with them preemptively instead of threatening them with, and overburdening, FAR. Abecedare (talk) 09:39, 26 June 2009 (UTC)[reply]
A wikiproject would make sense, yes. (Not that I'd join it; I'm not much of a project guy.) Adding an alt text requirement to the criteria shouldn't threaten or overburden FAR. We needn't add alt text to all featured articles right away, and if some article is "threatened" (presumably for political reasons?) with an FAR it's only a matter of a few minutes to add alt text to it, which would remove the "threat". Eubulides (talk) 09:55, 26 June 2009 (UTC)[reply]
I fully support this and can't understand the PITA argument at all. If our featured articles "exemplif[y] our very best work" then it shames us if we say this is possible while ignoring the needs of some of our readers. I guess for many of us, writing good alt-text involves learning, but it doesn't look that hard. I would hope many FA nominees, when told "please could you add some alt-text for the images", would react positively that they had learned something important that they would continue to apply in their article-writing. Colin°Talk08:14, 26 June 2009 (UTC)[reply]
As the one who made the PITA comment, let me clarify that I was really talking about the already onerous licensing requirements, particularly for images loaded to Commons, which are offputting enough to many editors without adding yet another hoop to get through. Having said that though I'm all in favour of encouraging editors to add alt text, particularly since I've seen Graham's evident enthusiasm for the idea, and have made a start on doing that in some of the articles I've written. This discussion has been interesting, as it's not really an issue I'd given much thought to before. --MalleusFatuorum12:23, 26 June 2009 (UTC)[reply]
It's worse than you think. {{Infobox Museum}} doesn't allow alt text (check the page and see the logo has no alt text). I found a problem earlier with with {{Infobox disease}} which didn't accept "|alt=" but was happy with "|Alt=", sadly that fix didn't work for museum. There is a can of worms in infobox templates and somebody who knows what they are doing needs to go through and make sure they all accept |alt= parameters, if we're going to be taken seriously. --RexxS (talk) 16:54, 26 June 2009 (UTC)[reply]
Thanks for that Museum of Bad Art example! It's beautiful! I trimmed some of its alt text to remove unnecessary phrases like "A color photograph of ...". I also added alt text to the {{Infobox Museum}} use, and asked for the trivial (and tested) changes to be installed into the template so that it supports alt text. I expect it to be installed shortly by an admin. Protected templates are admittedly a hassle if you're not an admin, but I've added alt text support to several infobox templates now, and they've all been simple and fast to fix. (I'm not an admin, thank goodness.) Eubulides (talk) 18:50, 26 June 2009 (UTC)[reply]
Curious: why is "Color photograph of" and such not preferable? Wouldn't visually impaired people be interested to know what the image actually is? --Moni3 (talk) 18:53, 26 June 2009 (UTC)[reply]
People who can't see an image primarily want to know what it looks like. How the image was prepared is typically unimportant, as is provenance. In a few cases (for example, this mirror view in a diascope of an autochrome of Percy MacKaye), a sighted viewer will immediately see that it's a color photograph, and for these images it's important to say so in the alt text; but normally the phrase "Color photograph of" is useless baggage in alt text. Eubulides (talk) 19:09, 26 June 2009 (UTC)[reply]
Small family trees can be converted entirely to text. I used parentheses to indicate subfamilies and added text like "Family tree. Ann Whittle (Chattox)'s children were Elizabeth (Bessie) Whittle and Anne Redfene (who married Thomas Redferne, with daughter Mary Redferne)" for Image:ChattoxFamily.png. Perhaps there's a standard for this in genealogy; that might be better than my ad hoc notation. Eubulides (talk) 22:30, 26 June 2009 (UTC)[reply]
Black and white photographs have historical markers in them. Color photographs do as well. Black and white photography, particularly in recent decades, is often used for artistic effect. I added alt text to 13 FAs yesterday explaining in every one what the image actually was. I used references to colors knowing that fully blind readers may not have any idea what they refer to. A digital representation of a map is different than a hand-drawn map. I have to say I raise my eyebrows at statements that all visually impaired people deem the same details as important or unimportant. --Moni3 (talk) 19:15, 26 June 2009 (UTC)[reply]
Well, this is a judgment-call area and reasonable editors can disagree. I'm not visually impaired, but I do use text browsers like Lynx that benefit heavily from alt text, and when doing so I want to see a brief description of the main point of the image.
For Image:João Zeferino da Costa - Moisés recebendo as tábuas da lei - 1868.jpg, for example, I find A (at right) to be much less useful than B. This is a deliberately extreme example of course (nobody is seriously proposing the A style) and it is not meant to be a fair comparison, but I hope it illustrates the issues involved.
I think we agree that when color (or black-and-white) is significant to the appearance of the image, it should be mentioned in alt text; when not, then not. Where we might disagree is how frequently it is in Wikipedia that the colors are significant.
Part of this may have to do with the fact that I rarely edit articles on art. In art, colors typically matter; but in other fields, they often don't. For Image:US-autism-6-17-1996-2007.png, for example, colors are irrelevant, so it's perfectly reasonable to use alt text that doesn't mention color or way the image was prepared.
I use a Firefox Add-on called Web Developer toolbar that allows me (among many other things) to display alt attributes above images. That lets me check at a glance that the alt description matches the image. Others may find it useful. --RexxS (talk) 00:15, 27 June 2009 (UTC)[reply]
To summarize the above discussion: Awadewit, Colin, Physchim62, and Proteins were in favor, some strongly so. Malleus Fatuorum was initially very much against, but later said was in favor. Iridescent really disagreed, citing Chelsea Bridge as an article that wouldn't be helped with alt text, but after alt text was added to Chelsea Bridge Graham87 thought it was good work and Iridescent didn't follow up. Ealdgyth wasn't opposed (though not happy). Abecedare thought the time was not ripe yet, citing India as an example; and after I added alt text to India suggested again the idea of a WikiProject.
Graham87 uses alt text and had an opinion that weighed heavily with us, namely that alt text should be encouraged but not required on every image, and should be brief. In an attempt to summarize the consensus on all the above, I added "brief and useful alt text when feasible". Eubulides (talk) 22:58, 1 July 2009 (UTC)[reply]
Question on the wording Why "brief"? The examples given at WT:FAC #Alt text in images seem to suggest that good alt text need not be exactly brief. Also what does "when feasible" mean? "feasible" implies that there may be cases where one is not capable of putting in alt text, but I don't think this is what the intent is here though, is it? Do we mean "where appropriate" or "where applicable"? TwilligToves (talk) 05:29, 3 July 2009 (UTC)[reply]
"Brief" because Graham87, who (I guess) can't see the images and actually uses the alt text, requested "brief" in WT:FAC#Alt text in images. "Brief" is a relative term; I just now reworded the longest alt text on that page to make it briefer, and I hope that suffices. "When feasible" is intended to cover the case (mentioned in that discussion) when a template won't let you specify the alt text. Removing "when feasible" would be OK, I expect, as it's not that hard to fix the templates. Eubulides (talk) 06:34, 3 July 2009 (UTC)[reply]
I am a little late to the show but I wanted to chime in here. I understand the need to add alt text to allow a max number of users to be able to access WP content and applying this to an article is one thing. I have several lists that have a large number of images and this makes me rethink putting images on the lists the way I have been. See List of Medal of Honor recipients and associated lists in hte template (Veracruz or African American are a good example and I am working on WWI now). Korea is another good example. In my opinion if I was using technology which allowed me to hear or feel the text because I could not see the image I think I would feel overwhelmed by the info. IMO. --Kumioko (talk) 01:57, 10 July 2009 (UTC)[reply]
Thinking on this a little more I also think thay before we implenent this and potentially delist a huge number articles ad lists we should give folks a little while to get used to the idea of using it. I used awb to do a scan of 14000+ articles related to US military biographies over night and only 9 had alt text and after looking at them not every image had it.-Kumioko (talk) 11:45, 10 July 2009 (UTC)[reply]
I added alt text to List of Medal of Honor recipients. I don't think the result is overwhelming in the context of a list of all the Medal of Honor recipients (a list that is already overwhelming, at least to this naive reader). On the contary, the change is a big improvemen. For example, for the Charles Brown entry a screen reader will now say something like "Large foreign flag behind three 1870s soldiers or sailors on a shipdeck" where it would have said "one eight seven one ess you jay a gee eye dot jay pee gee"; this should be an immense improvement for the visually impaired. This text is not overwhelming in the context of an entry that already says "Charles Brown ... Brown at right USMC Corporal aboard the USS Colorado June 11, 1871 USS Colorado Assisted in capturing the Korean standard from the citadel of the fort." Eubulides (talk) 18:28, 10 July 2009 (UTC)[reply]
As far as I know, this does not apply to featured list candidates, the change only applies to featured article candidates. Ealdgyth - Talk13:02, 10 July 2009 (UTC)[reply]
Criterion 5 of the FL criteria has been updated to say that images should have "'alt" text if necessary", whatever that's supposed to mean. As opposed to the FA criterion of "alt text where feasible". --MalleusFatuorum13:21, 10 July 2009 (UTC)[reply]
I tend to think that featured is featured whether it be an article or a list and the criteria should be basically the same and this still doesn't solve the problem of what to do about all the featured articles that currently do not have the alt text. Do we unfeature them or give some timeframe for them to be fixed before its implemented. By all means make it law for all new ones being submitted but at some point the old ones will need to be considered as well. Also, the FL folks appear to be taking the stand (as they should in my opiion) that if its required for FA then FL will do it also. --Kumioko (talk) 13:51, 10 July 2009 (UTC)[reply]
Nobody is suggesting immediately delisting thousands of articles. As articles and lists come up as candidates or for review, this requirement (along with the other requirements) can be checked. In the FA and FL context, alt text is a fairly minor requirement; it doesn't take that much time to do compared to all the other stuff one has to do. It took me about 30 minutes to add alt text to all the images in [[List of Medal of Honor recipients, and it would have taken less if I had known anything about military uniforms and their styles. Eubulides (talk) 18:28, 10 July 2009 (UTC)[reply]
I support a strong encouragement for alt-text, and I think it should be a requirement for images that appear on the main page excerpt. I'm going to drop a note at WT:WPACCESS pointing here as well. –xenotalk13:29, 10 July 2009 (UTC)[reply]
Ok I ran into another point of question in regards to the display of ranks and or ribbons for military folks. Such as the case of Michael P. Murphy or Smedley Butler the Medal of Honor recipients who have ribbons displayed. How does one describe these ribbons and should we have a standard alt text for each so that we don't end up with 50 different descriptions for the same medal or ribbon. This also applies to the rank insignia where displayed. Its relatively easy to describe a four star general but a bit more difficult to describe the insignia worn by the navy to designate a career field. --Kumioko (talk) 18:30, 10 July 2009 (UTC)[reply]
Recommendation: I would like to recommend that as we implement this we start with the image itself. If we create a standard alt text for an image then a bot, AWB or even a human can ensure that the alt text is 1)associated with the picture and 2)is consistent from article to article for that picture. I admit that it may require a modification to the way that comments are associated to images however I think this is the easiest way that allows us to present a clean and consistent alt text to the readers and editors while at the same time minimizing the manual editing required. If we can come up with a standard wau to document the alt text for an image it shoudl be relatively easy to create a script or bot to populate said image with alt text.--Kumioko (talk) 18:42, 10 July 2009 (UTC)[reply]
I've asked at VPT if it would be possible for such a solution to allow the alt-text to automatically be included if not provided when an image is used. See WP:VPT. –xenotalk19:06, 10 July 2009 (UTC)[reply]
Semms to me that this change to the criteria wasn't really well thought out. The alt text clearly ought to be associated with an image, not with the use of an image in an article. --MalleusFatuorum19:11, 10 July 2009 (UTC)[reply]
Disagree with this, whether or not the "alt text requirement" goes ahead. The appropriate alt-text for an image depends on context; this image of Lindsay Lohan wearing a Gucci hat would warrant totally different descriptions in its two different Wikipedia usages; in Gucci, the text would focus on the design of the hat and the person wearing it would be irrelevant, whereas in Lindsay Lohan it would describe the person wearing it with just a passing mention of the hat. – iridescent19:10, 10 July 2009 (UTC)[reply]
I don't think that's the case, as the alt text is simply describing the picture. It may require different captions, of course, but not different alt text. --MalleusFatuorum19:15, 10 July 2009 (UTC)[reply]
I agree with Malleus. The description of the photo would seemingly stay the same, while the caption would change. I don't think that means the alt text idea is ill-conceived, however. I think that means we are still learning to write alt text - a different point entirely. Some people are still learning to write "professionally" and that has been part of the criteria for a while now. :) Awadewit (talk) 19:23, 10 July 2009 (UTC)[reply]
Although I can see that in specific circumstances the image can depend on the context I think that more than 90% of the time a standard desciption of the image itself is sufficient. If the editor feels like it doesn't adequently describe the image based on the context then by all means change it but I think we should have a basis from which to implement this editorial nightmare. --Kumioko (talk) 19:19, 10 July 2009 (UTC)[reply]
Lindsay may not be the best example to use, as she was just the result my my clicking "random page" to find an image used on multiple articles. Take for example the multiple "terrace house" photos in Noel Park (which I went through and alt-textified as an exercise last week) – in the context of this article (showing the subtle differences between the five different styles of design) they need to include descriptions of architectural details, while in most other usage "row of red brick houses" would be sufficient. Additionally, the alt text is used to describe similarities (for instance, between the church and the village hall), which would make no sense if the images were used in separate articles. – iridescent19:25, 10 July 2009 (UTC)[reply]
As commented at the VPT page, if this functionality were available, it should also be able to be overridden by explicitly providing it when calling the image. I think the merits of the technical proposal should be discussed at VPT, fwiw. –xenotalk19:29, 10 July 2009 (UTC)[reply]
(bloody EC)Comment: I am late to this discussion because I have been very busy in real life (swine 'flu). But I have noticed that Eubulides has been requesting the incorporation of alt texts at FACs, and the change to the FA criteria. So, I added alternative texts to Rotavirus and, for some images, to Virus, (comments welcome). The problem that I discovered is that "a picture is worth a thousand words". So, whom are these alternative texts for; those readers with slow modems, those with less than good eyesight or those who use clever speech programmes? Perhaps, it is all of these. I think the community needs some guidance on how to write them. Graham ColmTalk19:22, 10 July 2009 (UTC)[reply]
They do not seem that difficult to write, and in esoteric science articles they might even be better than the legend! But, as Colin once wrote, the art of good writing is in knowing your audience. Graham ColmTalk19:46, 10 July 2009 (UTC)[reply]
On further reflection, I support this in principle because, (sorry for the POV) it will make scientists such as me look at our images more discerningly and not assume a knowledge of scale, anatomy, viruses, molecular biology, microbiology, chemistry and so on and so forth. But, I accept that it might be difficult to maintain WP:NPOV for photographs of people doing, well, what people do. Graham ColmTalk20:02, 10 July 2009 (UTC)[reply]
The difficulty in writing alt text is exactly that: making it descriptive and easy to understand without being verbose or using POV language and original research. Dabomb87 (talk) 23:03, 10 July 2009 (UTC)[reply]
Also jumping in late here... I strongly support this proposal. Alt text is very important for accessibility, and it's not that hard to add. I would be fine with the additional proposal above to specify default alt-text for an image that can be overridden in an article. --Alynna (talk) 12:45, 14 July 2009 (UTC)[reply]
I through the alt viewer quickly together after I noticed that Ruff had alt description but the syntax was wrong (|alt = instead of |alt=). It works like the other tools. Could throw together a short description for unfamiliar editors to use the tool, to replace the placeholder text. I'd do it myself, but typically confusion arises. — Dispenser17:17, 14 July 2009 (UTC)[reply]
An average Wikipedia FA writer, not waving but drowning. Alt=A black and green man with a squished face and half his head missing holding his hands up against glass with squiggly green blotches on it. :)
I would like to remove the requirement to provide alt text in FAs. It is yet another stylistic burden on FA writers, and there are already too many. It is getting to the point where the prospect of submitting an article for FA is something of a depressing prospect, and anything that makes that even worse should be avoided. SlimVirgintalk|contribs08:38, 4 August 2009 (UTC)[reply]
As the above discussion illustrates, alt text typically requires little extra work compared to what else needs to be done to create an FA, and it can markedly improve accessibility to visually impaired readers. Do you have specific examples of why alt text would be an undue burden? Eubulides (talk) 09:04, 4 August 2009 (UTC)[reply]
When I added it at the recent FAR at your request, I found it an enormous pain, and that was only for three images. I would hate to have to go through an article with a lot of images figuring out how to describe certain types of architecture, or "man with a wig looking to his left" to describe Immanuel Kant. You described Clare College, Cambridge as a nondescript three-storey building :). I can't say I blame you because what else is there to say, but it's hardly accurate.
I really find these of very questionable value, and as I said, even if they were the best thing since sliced bread, I'd still say please don't add any more style issues to the FA criteria. The editors who want these can add them to FAs themselves, surely. SlimVirgintalk|contribs09:13, 4 August 2009 (UTC)[reply]
Alt text has little value to sighted readers with graphical browsers, of course; it's designed more for visually impaired people who use screen readers and the like. The alt text you added to Bernard Williams at FAR contained 104 words total, for four images. I'm sorry that you found this an enormous pain, but with a little practice most editors find that it's not that much work. Perhaps it would help to say that alt text is supposed to be functional rather than perfect, and that it should be brief? Eubulides (talk) 09:43, 4 August 2009 (UTC)[reply]
I also oppose making alt text mandatory at FA reviews or anywhere else. The discussion so far has focussed on "photographic" images, with the exception of one map. Alt text for diagrams would often be a nightmare, for example File:Arthropod head problem 01.png would need alt text almost as long and complex as a summary of the Arthropod head problem, File:Mollusc generalized.png would need alt text about as long as the description of basic mollusc anatomy, File:Porifera cell types 01.png would need alt text about as long as the description of sponges' cell types - and if it proved possible to make the alt text more concise in these diagrams, the effort required would exceed that required to compose the corresponding article content text. I'm sure there are other types of image and types of article where alt text is unlikely to be helpful at all unless it is very long or ridiculous amounts of effort are expended. --Philcha (talk) 09:55, 4 August 2009 (UTC)[reply]
Even for "photographic" images, alt text often cannot help users with visual difficulties, slow connections,etc. In the example Eubulides cited (09:43, 4 August 2009), alt text text for pic of Bernard Williams, the description of his clothes is irrelevant as this pic aims to identify an individual, and for that you need a few dozen features peculiar to an individual and often hard to describe in words rather than by measurements (e.g. "centres of eyes X cm apart and Ycm below crown of head, nose Z cm long, etc., and goodness knows how one handles colours of eyes, hair , complexion, etc.). The RNIB recommends "picture of X" for a simple example - whihc is mor erelevant, but not informative. As far as I can see, the proponents of alt text have not done the hard work of reviewing and commenting publicly on hundreds of diverse examples. --Philcha (talk) 10:08, 4 August 2009 (UTC)[reply]
(ec) There are two separable issues here, E. First, the issue of helping readers: is there any evidence that it actually helps visually impaired people? And secondly, the issue of helping editors. No matter how helpful another style addition might be, we are literally sinking under the weight of these requirements. That we only have 2,000 or so FAs out of however many million articles is illustrative of it. Seriously, my heart sinks when I think of bringing an article to FA, even though I love writing, I enjoy writing WP articles, I love improving them, I love to see them complete and professional-looking, but I feel myself losing what is left of my mind at the thought of bringing them to FA. So, really, this is a plea for mercy. SlimVirgintalk|contribs09:56, 4 August 2009 (UTC)[reply]
Please see #Alt text helps the visually impaired below. In practice, alt text is one of the easiest parts of writing an article; it's easier than the article text, for example, because it doesn't need sources and can be immediately verified. If the FA requirements are onerous I suggest looking at some of the other requirements, not this one, as this one is fairly minor. Eubulides (talk) 17:43, 5 August 2009 (UTC)[reply]
Regarding Philcha's point, I agree very much. I couldn't see the point of removing more informative descriptions of Cambridge colleges and a man punting on the River Cam, in favor of "three-storey building, man in a boat with a stick." I can't see how that helps anyone. It felt like a make-work project. I should add that I don't mean to sound disrespectful to the people who want this, because obviously you're trying to help readers, which is clearly commendable. I just question whether this really is helping, and I also don't want to see a small number of readers helped at the expense of the straw-that-broke-the-camel's-back for FA writers. SlimVirgintalk|contribs10:26, 4 August 2009 (UTC)[reply]
Slim, can you please stop bitching. If this stylistic requirement is causing you to "drown" figuratively, that's rather sad, but it's your issue. It takes no time to add halfway decent alt text, and I hardly expect FAR will be filled with articles that don't meet this criteria--anyone can add it, it requires no special knowledge of the topic. Trying to suggest that we should lower FA requirements isn't going to win over anyone here. If you don't want to do the work, no one's forcing you to. --Der Wohltemperierte Fuchs (talk)12:00, 4 August 2009 (UTC)[reply]
I tend to agree with David. I'm recently much less gung ho about getting articles to FA, but not because of the amount of work that goes into the article. I'm still constructing articles with FA criteria in mind. Every several months I go read over the articles that were promoted to FA previously, including the ones that have appeared on the main page. They need dusting off and prose tweaking occasionally because standards rise. I like that. I'm glad visually impaired folks will have an opportunity to understand the images in an article, even Museum of Bad Art. --Moni3 (talk) 12:16, 4 August 2009 (UTC)[reply]
(ec) But the point is that people are trying to force it. The editors who want the alt text are free to add it themselves, but what has happened here is that others are being required to add it whether they want to or not. That's the point I'm objecting to—objecting, mind you, not "bitching". No need for that kind of response, David. SlimVirgintalk|contribs12:17, 4 August 2009 (UTC)[reply]
When you're involved, Slimmy, it's basically my stock response. Guess what: the FA criteria have always been a set of standards many editors don't agree with. If you don't, then don't participate in the FAC process, because you're going to have people picking apart your articles and making sure they adhere to standards. News flash: this is nothing new. --Der Wohltemperierte Fuchs (talk)12:20, 4 August 2009 (UTC)[reply]
Is there a difference between, say, alt text and inline citations? Or strict image policy adherence? Is there not value in strengthening criteria and setting high standards? --Moni3 (talk) 12:26, 4 August 2009 (UTC)[reply]
(Response to David) In case you haven't noticed, many editors have already stopped trying to meet these criteria, many of which don't actually make a better article for the reader. It's been my own observation that FAs have become more formulaic and less interesting. While they aren't poorly written, one that has brilliant prose, is engaging and motivates the reader to click the links, is an exception rather than the rule - although they do tend to have their m-dashes in all the right places. I'd been looking forward to returning to work in this area now that my external obligations had lessened, but reading recent FACs and their reviews has left me feeling unmotivated. Risker (talk) 12:29, 4 August 2009 (UTC)[reply]
Many of us put a lot of a time into reviewing articles for sourcing, images, comprehensiveness, and prose. Good to know that this is being ignored and forgotten and that all people remember is the MOS review. Awadewit (talk) 12:34, 4 August 2009 (UTC)[reply]
Moni, it's just a question of it being yet one more style thing for FA writers to adhere to. Citing sources is obviously needed and is policy; ditto with images. But alt text isn't policy, and isn't required by the MoS. Yet somehow it has managed to be added here as a stipulation. It's also not clear that it actually helps anyone, which is why some evidence of that would be appreciated. I think it needs to be removed from here and discussed at MoS level. SlimVirgintalk|contribs12:34, 4 August 2009 (UTC)[reply]
Why do individual editors stop trying to get FAs? Is it because of endashes and alt text? Is it because we're starting to demand something better than GoogleBooks and the top 10 hits on a search engine? These are generalities that cannot be attributed to an alt text requirement. If the only thing editors had to do to get an FA was to add alt text, I figure a bunch of editors would be swarming to get FAs. SlimVirgin seems to be saying that alt text is a breaking straw over a foundation of many criteria. What the straw falls atop is a (what I hope is) a strong foundation of superior sourcing and brilliant writing. That takes weeks and months to do. I spent 2 days filling in alt text for 14 FAs. --Moni3 (talk) 12:38, 4 August 2009 (UTC)[reply]
I know that one of the reasons editors stop submitting FACs is because of the increasingly strict style requirements. I'm all for style consistency within an article, and good sourcing. But there is an intense focus on style at FA, contrasted with very little focus on the writing or the substance of the content; the latter is harder to do, so naturally there's less of it, but still people feel they have to inject something, so the style issues are often piled on. I'm not arguing against what's already in place; it's part of the culture now, and it probably can't be changed easily. What I am arguing is that we shouldn't make it worse. We especially shouldn't add criteria that aren't required by the MoS, aren't policy, and haven't been widely discussed. Alt text has nothing to do with superior sourcing and brilliant writing. It's exactly the kind of thing that might put a brilliant writer off.
In addition, and this is a separate issue, it's not at all clear that the way we're required to write the alt texts is at all helpful or informative. Who is helped, who is educated, by being told that, in Immanuel Kant, there is an image of a man in a wig looking downwards and to his right? SlimVirgintalk|contribs12:54, 4 August 2009 (UTC)[reply]
Multiple comments above say that alt text isn't policy, as if that settled the matter. First, this is technically incorrect, as alt text is recommended by Wikipedia:Image use policy. Second, even if alt text weren't mentioned by policy, it's still recommended by several guidelines, not only within Wikipedia but by the W3G. There are good reasons for these guidelines, even if individual Wikipedia editors may not agree with them.
There may well be too many requirements for FA. But alt text is by far one of the simpler and easier requirements to meet. In the FAs I've written alt text is an imperceptible part of the work, less than 0.1% I would estimate. If the goal is to make FAs easier to write, the focus should be on other FA requirements, not this one.
The claim that alt text has not been widely discussed is incorrect. It's been discussed in many forums, not just this talk page, and has wide support.
"It's exactly the kind of thing that might put a brilliant writer off." No, it's not. What most puts off brilliant writers are things like stubborn and poorly-supported disputants on talk pages, lack of access to good sources, etc., etc. There are dozens of things that are far more a barrier to good FAs than alt text is. And there is no evidence, other than unsupported opinion, that alt text is a significant barrier to FAs. Eubulides (talk) 17:43, 5 August 2009 (UTC)[reply]
<reset>Editors not trying for FAs: On one hand we should stop speculating why editors don't try for FAs and find out, but on the other hand the stringent criteria weeds out editors who are not devoted to maintaining and improving their articles. Until I see a report with data indicating why editors are not trying to get FAs, I'm not going to take FA talk page observations as Gospel. The objections I've seen to the FAC process either turns editors away forever, or strengthens them to improve themselves and their articles. I think that's a valuable fork and a decision editors have to make.
Alt text: I suppose it remains to be seen how visually impaired readers are assisted. I'm not (that) visually impaired, so I guess we would need some input from readers who are. --Moni3 (talk) 13:01, 4 August 2009 (UTC)[reply]
@Slim: If any "brilliant" writer is put off by having to describe an image or two to a disabled viewer, he or she is a lazy moron. Brilliance doesn't cut it if you aren't willing to do any work, and if you can't spend another five minutes on improving an article, you aren't going to survive the FAC process anyhow. Just let it drop, and go crawl back to whatever hole you spend your wikidays in, please. --Der Wohltemperierte Fuchs (talk)13:43, 4 August 2009 (UTC)[reply]
I really don't see any need to adopt that kind of insulting tone David. I've certainly become more reluctant to submit articles recently, but it's nothing to do with alt text. SlimVirgin asks a valid question nevertheless. Does alt text (as it's being written now) really help the visually impaired or not? --MalleusFatuorum13:47, 4 August 2009 (UTC)[reply]
I'm sorry to hear this tone taken, it's really not necessary to call folks "lazy morons". I objected (not very strongly, granted) to the adding of this requirement to FACs, does that make me lazy and a moron? In the last year or so, FAC has generally tightened up the requirements for sourcing, image sourcing, implemented a higher standard of sourcing and stepped up enforcement of the non-free image requirements. While I can see the point of ALT text, I'm not adverse to hearing more input on how we can write better ALT texts, and if it really does help? Ealdgyth - Talk13:54, 4 August 2009 (UTC)[reply]
I'm not insinuating anyone is foolish for disagreeing with the use of alt text. I'm saying that Slim's assertion that these "brilliant" writers will turn away from writing FAs because of alt text is bull, plain and simple. Many people above have testified that adding it took very little time, even to highly technical and scientific topics. What turns people off of writing FACs would be worth discussing, but it has nothing to do with alt text. --Der Wohltemperierte Fuchs (talk)14:07, 4 August 2009 (UTC)[reply]
I won't speak for SlimVirgin. What I am saying is that the ever-increasing focus on minutiae and stylistic matters that have no effect on the reader is demotivating quality writers, and reinforcing formulaic ones. Can anyone explain how the reader is benefitted when the writer must be facile at using more than a dozen different citation templates - most of which render essentially the same information in the footnote? There has just been a significant dustup about image sizes, but only when someone deeply invested in style suddenly realised they were a problem; it's something editors (and readers) have been saying for a long time. I've got half a dozen manuals of style sitting on my desk at home; ours is more than twice the size of any of them, and is far more convoluted and complex. It's time to go back and re-examine what a featured article should be: there has to be a midpoint between "anything goes as long as it's interesting" and "over-produced and dull", but the minority of articles that find that midpoint seem to have to fight for their life to survive this process. Risker (talk) 14:32, 4 August 2009 (UTC)[reply]
I can't explain the benefits of cite templates. I just refuse to use them anymore. One of these days I have to go through six or seven of my FAs and take out the templates. Cite templates are not a requirement. --Moni3 (talk) 14:38, 4 August 2009 (UTC)[reply]
If it's not such hard work to do, perhaps people should be constrained against opposing FACs on this basis, instead having to complete the ALT wording themselves. --Dweller (talk) 14:16, 4 August 2009 (UTC)[reply]
I would hope they would too, but that just bridges into another issue of SOFIXIT issues that are endemic to any part of the FA criteria. Alt text is certainly one of the easiest issues to fix as it requires a very superficial analysis of an image. I'm not sure we can constrain people, though. --Der Wohltemperierte Fuchs (talk)14:19, 4 August 2009 (UTC)[reply]
Der Wohltemperierte Fuchs, I'm put off from nominating for FA because of various time-consuming niggles - mostly WP:MOS:creep such as WP:DASH, but alt text has recently joined my reason for being averse. So you just called me "a lazy moron" (13:43, 4 August 2009). As an admin, you should be much more careful to keep a civil tongue in your head. I expect to see your apology soon. --Philcha (talk) 14:35, 4 August 2009 (UTC)[reply]
WP needs to get its priorities straight. My own are: (a) good coverage supported by good sources and explained clearly; (b) improving the prose. Requirements that deliver little benefit to readers should not be allowed to take up much time. Despite its theoretical benefits to users with visual difficulties, slow connections, etc., I see too many cases where alt text is of little help - fairly technical diagrams; images that identify persons, places, etc.; some for which there are no suitable terms in normal English, such as the box art some games (e.g. Master of Orion II: Battle at Antares); and good luck with alt text for the pics at Anal sex :-)
I'd also like to see examples of the kind of alt text that visually impaired readers actually find useful. Is "man in a wig"? really helpful, for instance? In general it's my conviction thast the requirement for alt text was rushed and botched. --MalleusFatuorum14:39, 4 August 2009 (UTC)[reply]
If you'd also handle the WP:MOS:creep I'd be sorely tempted, just pour épater les bourgeois like Gropecunt Lane did. However I think I think getting the ones I've targetted to GA would benefit readers and WP's reputation rather more, as a return on the work required. However if anyone else wants to take up Moni3's offer, I assign my rights in that offer to them :-) --Philcha (talk) 14:56, 4 August 2009 (UTC)[reply]
There are other style things I'm not a fan of, but alt text is easy to add (although I've seen some serious struggles with it, for some reason), and helping the blind is a noble thing to do. I bet blind people particularly like Wikipedia, and we should cultivate that. - Peregrine Fisher (talk) (contribs) 14:38, 4 August 2009 (UTC)[reply]
Maybe not that easy, or agreed upon, but I bet blind people are pretty happy with anything that doesn't just repeat the caption (which seems to be what people add on their first attempt). - Peregrine Fisher (talk) (contribs) 14:44, 4 August 2009 (UTC)[reply]
I'd like to see some feedback from the horse's mouth, so to speak. Also I'm curious, given that many people would prefer not to see some of the images accompanying many wp articles (Philcha mentioned Anal sex above), whether they'd be happy to have a graphical description read out to them. --MalleusFatuorum14:48, 4 August 2009 (UTC)[reply]
(ec) I found it quite difficult to write it in a way that would be helpful. It would be time consuming to do it well on an article with a lot of images. How about we change point 3 of the FA criteria from:
It has images that follow the image use policy and other media where appropriate, with succinct captions, brief and useful alt text when feasible, and acceptable copyright status.
to:
It has images that follow the image use policy and other media where appropriate, with succinct captions, and acceptable copyright status. Brief and useful alt text is encouraged, but not required.
I don't think that goes far enough. The FA criteria have lost sight of the fact that the only images which really must have alt text are those without which it is not possible to understand the article. A picture of Kant in a wig is in no way necessary for the understanding of his article. --MalleusFatuorum14:53, 4 August 2009 (UTC)[reply]
I like that. I'd support wording that made mandatory alt text for images which are necessary for comprehension of the article. I'd also suggest for those, like me, who feared an extra bar to FA production, that it'd reduce the proportion of images covered by the criterion to a very low level. --Dweller (talk) 14:57, 4 August 2009 (UTC)[reply]
It has images that follow the image use policy and other media where appropriate, with succinct captions, brief and useful alt text for images essential to a complete understanding of the article, and acceptable copyright status.
I could sign up to that. I still think the cart has been put before the horse though, as what ought to have been done first was to make a software change to allow images to have a default alt text. It seems absurd for every editor to keep reinventing the wheel for the same image. --MalleusFatuorum15:16, 4 August 2009 (UTC)[reply]
I prefer SlimVirgin's previous suggestion, "Brief and useful alt text is encouraged, but not required" (14:50, 4 August 2009). There's no proof that it's always possible to produce alt text that is both brief and useful. --Philcha (talk) 15:34, 4 August 2009 (UTC)[reply]
I'm a totally blind user of Wikipedia; I navigate it using the screen readerJAWS. I find alt text helpful when it describes properties of an image that are easy to discern for sighted people but aren't appropriate for a caption (e.g. roughly what a logo looks like, the text of a cartoon). I really don't know what I'm missing out on when I've never seen an image. I think alt text should be encouraged in FA's, but not required where their use would be impractical. They're certainly not worth bickering over. Graham8716:02, 4 August 2009 (UTC)[reply]
I'd be really interested to know whether you find alt text such as for the image I drew attention to below useful or not, and if not what you might find useful for a map, for instance, if anything. It's the very first image here. --MalleusFatuorum16:22, 4 August 2009 (UTC)[reply]
Yes, I find it useful, if only because it describes the seal as circular. That's the only detail that's included in the alt text but not in the caption. The alt text is a good general overview of the image, as far as I can tell. Graham8700:24, 5 August 2009 (UTC)[reply]
That is very interesting. It would be enlightening to know your opinion of the usefulness of various alt texts. Perhaps a sample could be submitted to you and you could provide feedback. Do you like long, detailed explanations, say of the subject of a painting? Regards, —Mattisse (Talk) 00:31, 5 August 2009 (UTC)[reply]
I prefer it if the alt text is short and succinct, about the same length as a typical caption. It's probably better if sighted people evaluate alt text, because as I've said before, I don't know what I'm missing because I can't see any images at all. IIRC, the alt text will display if you turn off images in your browser. Graham8712:05, 5 August 2009 (UTC)[reply]
I came here from the alt text discussion on Wikipedia talk:Featured article criteria, so I'm kind of jumping in feet first, but I've got to ask: are David's comments to SlimVirgin the normal tone on this page? I must ask you to cut it out, David. People who disagree with you are not by that token "bitching", nor are their arguments "bull". And your reply when she complains of your rudeness? "When you're involved, Slimmy [sic], it [="bitching"] is basically my stock response." Did you get up on the wrong side of the bed today or something, or is it really your stock response? If it is, you need to revise your stock response. Right now. Stop sneering, please. Think about it. Bishonen | talk16:25, 12 August 2009 (UTC).[reply]
I would argue that alt text like this: "Map of Carrick, sea and rivers in blue with land going from brown to green depending on topography (brown higher, green lower); parish churches, major settlements and the monastery of Crossraguel are depicted on their locations", which accompanies the image to the right is next to useless. What is the point of having any old text just to get through a new FA hoop? Thoughts? --MalleusFatuorum15:37, 4 August 2009 (UTC)[reply]
I agree that the alt text could be improved. Also, the caption "Carrick in the 13th and 14th centuries" is not needed, as it duplicates text that is in the image. I attempted to fix this by removing the caption and replacing the alt text with "Map of Carrick, 13th–14th century, on the east shore of a large body of water. Its center was Crossraguel Abbey about 5 km inland. Nine parish churches and eleven important settlements ranged from Ballantrae in the south at the mouth of the Stinchar, then 40 km north to Greenan at the mouth of the Doon, and east to Bennan about 20 km up the Girvan." which conveys more-useful info to someone who can't see the image. Eubulides (talk) 17:43, 5 August 2009 (UTC)[reply]
I didn't feel like my view on the matter fit into any of the sections above. Therefore, I will put it here, by itself. Here are my thoughts on the addition of alt text to the FA requirement: AHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH!!!!!!1111 Ottava Rima (talk) 15:57, 4 August 2009 (UTC)[reply]
I may be missing something here, but why can't/isn't alt text be tagged onto the description page of images on Wikimedia Commons and those images on Wikipedia. That way, the alt text for images that are to be used multiple times won't have conflicting and inferior alt text, surely? MasterOfHisOwnDomain (talk) 19:21, 4 August 2009 (UTC)[reply]
It would make sense if the person who uploads it were to add it to the image page, then everyone after that could simply copy it—though, again, my preference would be to encourage, not require. SlimVirgintalk|contribs19:55, 4 August 2009 (UTC)[reply]
Yes that did cross my mind, however I also considered that some images would have no alternative (forgive the pun) alternative text. In these cases especially (although I still think in the case you highlighted above, separation could be given for emphasis of different items in the image) I do think including alt text in the description page could save on a lot of time. I thought I'd bring up the topic early, before it becomes too dug in (to the FAC process). MasterOfHisOwnDomain (talk) 21:05, 4 August 2009 (UTC)[reply]
The Wiki software should pull the title from the image if there is no overriding alt text at the use of the image. This would mostly solve the problem without editor effort. --John Nagle (talk) 21:28, 4 August 2009 (UTC)[reply]
Not bothering to read the above discussion, but just a random point that hard-coding alt text might be difficult if the images were on commons (I'm not a technical guy, so I could be horribly wrong.) --Der Wohltemperierte Fuchs (talk)01:16, 5 August 2009 (UTC)[reply]
My reading of the sections above is that there's no consensus to require writers to add alt text to all images in FACs, but there is consensus that alt text be encouraged, or that it be required for certain types of images. The two suggestions I made above that got support were (I'm bolding the alt text parts):
1. It has images that follow the image use policy and other media where appropriate, with succinct captions, and acceptable copyright status. Brief and useful alt text is encouraged, but not required.
or:
2. It has images that follow the image use policy and other media where appropriate, with succinct captions, brief and useful alt text for images essential to a complete understanding of the article, and acceptable copyright status.
or (per Malleus):
3. It has images that follow the image use policy and other media where appropriate, with succinct captions, and acceptable copyright status. Brief and useful alt text is encouraged, but not required, except for images essential to a complete understanding of the article.
The current version says:
4. It has images that follow the image use policy and other media where appropriate, with succinct captions, brief and useful alt text when feasible, and acceptable copyright status.
Version 3 2, mainly because I think it's very likely that version 1 will simply encourage reviewers to continue to insist on alt text for all images. --MalleusFatuorum14:46, 5 August 2009 (UTC)[reply]
On reflection I'd prefer version 3, which would say 'Brief and useful alt text is encouraged, but not required except for images essential to a complete understanding of the article. --MalleusFatuorum14:49, 5 August 2009 (UTC)[reply]
As a cautionary note, the phrasing "images essential to a complete understanding of the article" starts rubbing against NFCC#8, and could be taken to mean that every non-free image must have alt text, as if alt text is not needed, the image is not essential to the complete understanding of the article, and will fail NFCC#8. I'm sure there's cases of NFC images where alt text isn't helpful (due to possible repetition in the text or an image that is easily identified in the caption) and I can see some more aggressive NFC patrolers using that langauge to assert its not important. --MASEM (t) 16:59, 5 August 2009 (UTC)[reply]
I completely disagree with the above claim of consensus. Many editors have spoken against the idea of omitting alt text from featured articles; I haven't done a head count but it could well be that more have spoken against than for. The argument against alt text is based in the mistaken assumption that alt text does not help visually impaired readers.
Proposals 1 through 3 represent an extreme position, in that they say that it's OK to omit alt text unless an image really, really needs it. That's not the consensus here, nor is it the consensus elsewhere at Wikipedia or at W3C. Instead, the overall consensus is that alt text helps the blind, and that images should have alt text unless they are purely decorative. We should not carve out an FA exemption to these general guidelines.
If the wording is to be changed, it should be changed to:
5. It has images that follow the image use policy and other media where appropriate, with succinct captions, brief and useful alt text except for purely decorative images, and acceptable copyright status.
4 (the current version) would also be acceptable. 1 through 3 are not acceptable, as they attempt to overturn guidelines and policy. Eubulides (talk) 17:43, 5 August 2009 (UTC)[reply]
WP:IUP is policy and it gives as a rule of thumb "Add a good alternative text for images." I'm not particularly interested in wikilawyering over whether this is policy or guideline; the point is that alt text is in several guidelines and is mentioned in policy, and this page shouldn't be attempting to undermine that widespread consensus. Eubulides (talk) 18:02, 5 August 2009 (UTC)[reply]
Just to make clear: IUP does not require alt text. It discusses adding it as a "rule of thumb," just as it recommends cropping images to highlight the subject. But these are not policy. SlimVirgintalk|contribs18:14, 5 August 2009 (UTC)[reply]
I'd be surprised if there's a widespread consensus to require alt text for all images. Also, you've acknowledged below (Wikipedia_talk:Featured_article_candidates#Alt_text_for_portraits) that, where an image is purely decorative and doesn't require alt text, your argument is that the image ought to be removed. You are therefore saying that all images require alt text, which makes the wording of your version 5 not quite as it seems. SlimVirgintalk|contribs18:10, 5 August 2009 (UTC)[reply]
Again, I'm not interested in wikilawyering whether alt text is formally required by policy. The point is that there's wide consensus support for alt text both in guidelines and policy.
I agree that there is no consensus to require alt text for all images. However, there is consensus and multiple guidelines (both on- and off-Wikipedia) to supply alt text for images that are not purely decorative.
I did not argue that all purely-decorative images should be removed. Quite the contrary, it's fine to have purely-decorative images in Wikipedia. I just now, for example, made an {{editprotected}} request for {{Geographic Location}}'s purely-decorative compass roses to have better accessibility support, and my request did not ask for these images to be removed or altered.
It appears that some of the disagreement in the above discussion come from a lack of consistent use of terminology. I am using "purely decorative" in the sense used by W3C (see Web Content Accessibility Guidelines) and in WP:ALT #When to specify.
The W3C link says purely decorative means "serving only an aesthetic purpose, providing no information, and having no functionality". That leaves a lot to interpretation, but probably means "purely decorative" images are not allowed on WP anyway - say I felt the urge to plaster pretty flowers or kitten pics on an economic article, and are certainly not likely to be found on FAC candidates. There should be no reference to them in the policy. I don't know how many people have seen the issue raised at Wikipedia:Featured article candidates/Pavel Bure/archive1. Except for diagrams, family trees etc, it seems to me that the Alt texts provided as exemplars are generally unhelpful and that to make them helpful would invoke calls of OR. At least one example on the policy page is clearly mis-informative, where an autistic child pointing at a fish in an aquarium (per the caption) is described as pointing at a "window". I think no-one has nominated a visual arts article in the 3 months since this edict came into force, and I'm not surprised. Art historians often have to describe images their publisher cannot afford to illustrate, and the "man in a wig looking left" is not the approach they adopt - nor that of the Moses, gold coin and philospher alt text examples provided here and at the policy. Of the options above I'd favour 3, and I generally support Slim Virgin's comments. Johnbod (talk) 04:30, 6 August 2009 (UTC)[reply]
"probably means "purely decorative" images are not allowed on WP anyway" This reflects a misunderstanding of the technical term "purely decorative" as used by W3C and WP:ALT. Purely decorative images are definitely allowed on Wikipedia and are quite common in some articles. For example, {{flag|France}} generates "France", which contains a purely decorative image: the image is not intended to provide any info that's not in the text, and nothing happens when you click on it. Although it is true that Wikipedia thumb images are not purely decorative in the W3C sense because they have a link, plain pictures such as the one generated by {{flag}} are often purely decorative in Wikipedia, and are often found in featured article candidates.
Could you please illustrate the art historian comment with a specific example of an artwork and how its visual appearance is briefly described by an art historian, so that this can be contrasted with what WP:ALT suggests? Quite possibly we can use the result to improve WP:ALT. (I assume by "visual-arts article" you are excluding cinema, video games, and TV shows; we've had plenty of those, with no significant problems that I'm aware of.)
Thanks for catching the "window"/"aquarium" typo (fixed) and the too-long Pavel Bure alt text (I shortened it and added this to WP:ALT #Examples).
(outdent) I see your point about "purely decorative", but as this covered in the WP:ALT guideline and is likely to mislead, it should not be in the FAC requirements. The Bure alt text is certainly an improvement. For Greek coins, see this link and this catalogue These (like the alt text) do not mention that heads are in profile, because all heads on Greek coins are, but the average reader is unlikely to know this. The alt text above does not mention the band diadem, a sign of kingship, or the inscriptions; "well-fed" seems open to accusations of OR to me. Generally in short descriptions, identifiable figures are always named - strictly speaking the example above is incorrect to describe God the Father as a "man". The Greek philosopher image discussed above is unrelated to the subject and only included to show artistic style, yet neither caption nor alt text make the slightest attempt to characterize the style - which might also lead to OR of course. That is a problem in the Moses text also. In art history terms, the alt text examples (of all types, not just artistic images) tend to describe the Composition (visual arts) very briefly, but ignore the subject (which the caption will usually cover), the iconography, the style, and the meaning of the image. There are times when this approach adds meaningful information, but often it does not. In the Kant portrait example, an informed summary including the correct terms for his clothing & type of wig, or that he is formally dressed, and his age at the time, might be helpful, but "man in a wig looking left" adds next to nothing to the caption. The reluctance to mention racial appearance in the Indian photo with the Russian, and SoS Condi, gives semi-comical results, but are we always to mention race? Conveying useful very brief verbal summaries of images is a very difficult thing. Johnbod (talk) 13:47, 6 August 2009 (UTC)[reply]
Thanks for your insigntful comments.
I redid the alt text for the gold coin to mention profile, diadem, and inscription, and changed "well-fed" (which is OR) to "plump", resulting in 'A gold coin shows paired, profiled busts of a plump man and woman. The man is in front and wears a diadem and drapery. At top is the word "ΑΔΕΛΦΩΝ".' Identifying the figures in the alt text would be counterproductive, as it would repeat the caption; in this sense alt text takes a different role than the single description that's often in coin catalogs, which contains a mixture of alt-text-like and caption-like text.
I improved the alt text for the philosopher image to add "with furrowed brow and unruly hair", which indicates the style of the statue (realistic rather than idealized) without involving OR.
The alt text for the Moses example was intended for that picture's use in Ten Commandments in Roman Catholicism. I agree that different alt text would be appropriate in an art-history article, alt text that would focus more on the painting's visual style than on the scene it depicts.
As far as I know there is no alt text "a man in a wig looking left" anywhere in Wikipedia; that is a rhetorical example being used in some arguments against alt text. The alt text actually used in Bernard Williams #Critique of Kantianism is:
"Head of a man viewed from his upper left so that his brightly lit left forehead, receding hairline, and sharp nose dominate the image. He looks downward with a serious expression. He wears a modestly sized 18th-century-style wig and something small and white at his throat; the rest of his clothing is so black that it merges into the black background."
In the context of a philosophical article this could well be trimmed down a bit, but certainly this alt text does not have the problem of adding "next to nothing to the caption".
Race is admittedly a sensitive issue, in both alt text and in main article text. This can also be true for other visual attributes, e.g., "plump". The basic idea is that if an aspect like race or weight is a striking component of the visual appearance, then it should be mentioned; otherwise not. Editorial judgment is obviously required here.
With all the above in mind, I still don't see how writing alt text is typicaaly a very difficult thing. Yes, there can be tough cases, but generally speaking it's not a serious problem. One can write the alt text for a typical image in two or three minutes; its more-difficult aspects (uploading the image, captioning it, sourcing the caption, and integrating it with the text) typically take far longer than this.
As my tuppenceworth, 2 requiring "brief and useful alt text for images essential to a complete understanding of the article" sets a reasonable minimum standard, so that use with a screen reader is at least feasible. With any other images, editors should be encouraged to either add concise and helpful alt text, or to use |link= to hide the image link from the screen reader, subject to image license attribution requirements if applicable. 3 has an added comma: to me, "Brief and useful alt text is encouraged, but not required, except for images essential to a complete understanding of the article." seems to imply that it's not encouraged where the image is essential for understanding. Malleus's phrasing as "Brief and useful alt text is encouraged, but not required except for images essential to a complete understanding of the article." makes better sense, but perhaps would be better put as "Brief and useful alt text is required for images essential to a complete understanding of the article, and is strongly encouraged but not required for other images, which where possible should incorporate |link= to hide these decorative images from a screen reader." . . dave souza, talk11:17, 6 August 2009 (UTC)[reply]
As with many MOS-related discussions, the discussion itself is likely to consume more bandwidth and energy than adding alt-text to every single FA ever will. And, as with other MOS issues at FAC, it is so easy to do that if the FA writer or nominator doesn't do it, it's always easy to find someone who will take a few minutes to do it. I have still never encountered an otherwise worthy FAC on which I (or someone else) couldn't and didn't easily fix MOS issues. So, I support Eubulides' effort to make Wikipedia more accessible to a wider audience, and concur with his choices of 4 or 5. SandyGeorgia (Talk) 14:58, 6 August 2009 (UTC)[reply]
If alt text is included in FAs, it should obviously meet the general FA standards. The examples mentioned here, on the Pavel Bure FAC, and those on the WP:ALT page, suggest that producing Alt text of FA standard is neither quick nor easy. Johnbod (talk) 15:23, 6 August 2009 (UTC)[reply]
Quite. That is easy to produce any old alt text is not the issue. as what has certainly been more than adequately demonstrated is that producing useful and accurate alt text is very far from easy. I find it curious that the FAC requirement is just for the presence of alt text, not even requiring good alt text. --MalleusFatuorum15:32, 6 August 2009 (UTC)[reply]
I looked at the Pavel Bure alt text and it is neither short nor concise. Are the ages e.g. "in their thirties" from reliable sources? Is the presence of a "carved wooden table" important? —Mattisse (Talk) 15:42, 6 August 2009 (UTC)[reply]
Quite so! The original issue there concerned the lead pic alt text, which after 10 days or so & the attention of 4 editors, has now been improved and included as an example at WP:ALT. You can see the first, much longer but less informative, text alongside. Johnbod (talk) 15:57, 6 August 2009 (UTC)[reply]
The previous comments greatly exaggerate how much work it takes to write alt text. It doesn't take four editors five days to write alt text for a FA! A more accurate estimate would be one editor and ten minutes. Sure, editors could quibble about the result afterward, just as editors can quibble about any part of an FA. But the actual work is not that great.
"I find it curious that the FAC requirement is just for the presence of alt text, not even requiring good alt text." Nor do the FA criteria explicitly require that images be good. Nor do they explicitly require that the citations be good. WP:FACR does not, and need not, explicitly call out "good" for every single aspect of FAs.
A phrase like "in their thirties" is meant to be an abbreviation about appearance, not a possibly-controversial assertion about chronological age. If this isn't clear in practice, the phrase can be replaced by "about thirty". It's not a big deal either way.
For almost any alt text, there will be judgment calls, such as whether Image:Vladimir Putin in the United States 13-16 November 2001-20.jpg should mention the carved oval wooden table. Some editors could easily say it should be there, to indicate to a visually impaired reader the level of formality of the meeting; others could just as easily argue that it should be absent. (I wrote the previous sentence before seeing Graham87's remark that he found the detail useful; I would weigh his comment heavily since he uses a screen reader.) Editorial judgment is obviously required in matters like these, and reasonable editors will disagree in some cases.
In practice there is rarely controversy over whether a particular descriptive phrase should be in alt text. I've already seen far more controversy in this talk page about this issue than I've seen in weeks of reviewing real FAs, and I've reviewed them all.
"the discussion itself is likely to consume more bandwidth and energy than adding alt-text to every single FA ever will" At first I thought this statment was sheer hyperbole but on doing the math I found that it's within an order of magnitude or so of being literally true. Currently this thread (including #Alt text helps the visually impaired) contains about 23,000 words. As a back of the envelope guess, that's enough for about 1,000 images' worth of alt text, or about 200 featured articles. If we few editors had written alt text instead of laboring to complain about it and to respond to those complaints, we'd already be nearing 10% done in covering all the existing FAs with alt text, thereby directly benefiting the visually impaired. Wouldn't that make more sense?
Supporting alt text for all images as mandatory at FA: Eubulides, David Fuchs, Moni3, YellowMonkey, Peregrine Fisher, Sandy Georgia
Objecting to alt text for all images as mandatory: SlimVirgin, Philcha, Risker, Malleus_Fatuorum, Dweller, Dabomb87, Matisse, Graham87 [1]
Unclear: Awadewit, Ealdgyth, Johnbod, dave souza
And that the alternative sentences favoured are:
2. "It has images that follow the image use policy and other media where appropriate, with succinct captions, brief and useful alt text for images essential to a complete understanding of the article, and acceptable copyright status.
or
3. "It has images that follow the image use policy and other media where appropriate, with succinct captions, and acceptable copyright status. Brief and useful alt text is encouraged, but not required, except for images essential to a complete understanding of the article."
After reading through the discussion, and thinking about my own views, I do not strictly object to requiring alt text for all images (EDIT: excluding those used only for decorative purposes); in fact, I think I lean toward that more than only "encouraging". However, I do think alt text should be made mandatory for images that are vitally important to the article. Of course, if we can't decide what "vital" means, then I think alt text should be mandatory for all images. One thing though, I think someone—probably Eubublides, as he is the expert, and Graham87, since he's the type of editor for whom we're writing alt text—should write an easy-to-understand tutorial about writing alt text (like WP:ALT without so much detail and technical information). I also heard that someone is writing a Signpost dispatch about this with an interview of Graham87; how is that coming along? Dabomb87 (talk) 03:20, 11 August 2009 (UTC)[reply]
SlimVirgin's summary does not accurately represent the comments made by editors. As far as I know, nobody other than Dabomb87 (just now) supported requiring alt text for all images. WP:ALT and the W3G guidelines clearly say that alt text is not needed for purely decorative images. (I expect that Dabomb87 would agree with this exception, once reminded of it.)
Eubulides, you said elsewhere that purely decorative images should be removed, so I don't understand how you're using the term. Can you show us an example of an image in a current FA that does not need alt text in your view because it's purely decorative? SlimVirgintalk|contribs15:15, 11 August 2009 (UTC)[reply]
There seems to be some confusion here, as I wrote "My argument did not say or imply that purely decorative images should be removed." (Perhaps you missed the "not" in that statement? That might explain the confusion.) Below, I explain what I mean by "purely decorative" and give an example featured article Australia at the Winter Olympics that contains seven purely decorative images. Eubulides (talk) 23:40, 11 August 2009 (UTC)[reply]
Neither (2) nor (3) fairly reflect what Wikipedia and W3C guidelines/policy say, or what editors supporting alt text are saying. The existing wording, possibly with "except for purely decorative images" appended, better reflects that consensus.
SlimVirgin's summary is not accurate as a head count. It listed Risker and Dweller as opposing alt text requirements, but Risker made no comments about alt text, and Dweller's comments, although not in favor of alt text, did not clearly object to the requirements either. If it's that easy to be counted as an opponent of alt text requirements, why weren't (for example) Xover and Graham87 listed as proponents?
The summary also omitted Graham87's repeated endorsement of alt text. The opinion of a blind reader of Wikipedia should carry considerable weight.
Multiple reliable sources, cited near the start of #Alt text helps the visually impaired, support the claim that alt text helps visually impaired readers. The claim that alt text does not help the visually impaired is not supported by reliable sources.
I'd be happy to write a tutorial if that would help. Currently, for tutorial material we have a couple of paragraphs in WP:PIC #Thumbnails, along with the lead to WP:ALT. If that's not enough, I suggest refactoring WP:ALT so that the tutorial material is more to the front (each point with an example), and the more detailed technical stuff is more to the end.
The way I read Graham's comments he was saying alt text wasn't useful much of the time. One of the problems with this situation, as I see it, is that alt text is being required for all images by stealth. The current text says "when feasible," but that's not defined. You say that images that are "purely decorative" don't need it, also not defined. So can we have some clear definitions from you, please, and examples of FA images that don't require alt text? SlimVirgintalk|contribs15:21, 11 August 2009 (UTC)[reply]
The term "purely decorative" is defined in WP:ALT as an image that has no function (for example, nothing happens when you click on it) and conveys no additional useful information. This is intended to reflect the common technical meaning of "purely decorative" used by the W3G. For details, please see WP:ALT and the sources cited in WP:ALT #Sources.
Featured articles routinely contain purely decorative images. For example, the content of Australia at the Winter Olympics contains eight images, seven of which are purely decorative; Image:Saguenay 500m.jpg is the only image in that article that has (or needs) alt text. The ratio of purely decorative images is unusually high in that example, but it's quite common for featured articles to contain some purely decorative images. For example, Autism, which I help maintain, contains one.
Which browser are you using? I can clearly see eight images. For example, the lead image, in the infobox, is purely decorative. It is Image:Flag_of_Australia.svg, and it's the default size for thumbnail images (180px wide). The other decorative images are much smaller. Eubulides (talk) 02:39, 12 August 2009 (UTC)[reply]
I don't recall Graham87 ever writing that alt text "wasn't useful much of the time". On the contrary, in his most recent comment expressing his preference, he wrote "I strongly support the use of alt text"[2]. His only qualm was that he dislikes conflict and doesn't want us arguing about it. Opposing alt text on the ground that the opposition causes Graham87 to mute his evident strong support for it isn't opposition based on evidence: it's opposition based on a circular argument. And listing Graham87 as an opponent of the alt text requirement is a misrepresentation of his position, just as listing Risker is, and underlines the fact that the above summary of the discussion is not at all a fair summary.
And alt text is not mandatory. It's not required for purely decorative images. The current guideline is entirely consistent with Graham's comments, and has been modified on multiple occasion to better reflect his suggestions. Eubulides (talk) 02:39, 12 August 2009 (UTC)[reply]
"required for all images by stealth" First, it's not being required for all images. Second, there is no "stealth" here; the requirement is stated plainly and has been thoroughly discussed.
It is being required for all images. When I ask you what type it's not being required for, you say "not decorative" ones. When I ask you what that means you say images that have no function. What does that mean? SlimVirgintalk|contribs01:08, 12 August 2009 (UTC)[reply]
No, alt text is not being required for all images. It is not required for purely decorative images. These are images that are not needed for the understanding or use of the article. By "no function" I meant, roughly, "nothing happens when you click on them". Please see WP:ALT and the sources that it cites for the gory details. Eubulides (talk) 02:39, 12 August 2009 (UTC)[reply]
The "when feasible" phrase was intended to allow for problems in the MediaWiki software, which as far as I know have all been overcome since it was added; it can be removed at our convenience (though I'd rather not do this in the middle of this unrelated dispute).
That's correct. Support has been added for infoboxes used by articles at FAC or FAR as they come up. So far, we haven't run into any infobox that can't be fixed. It's generally pretty easy to fix them. This problem is a real one, but it has been less than anticipated. Eubulides (talk) 02:39, 12 August 2009 (UTC)[reply]
Graham87 also said more than once that he liked alt text to be "concise" and "short and succinct, about the same length as a typical caption". He also said he preferred that colours not be used in alt text, and I thought he said some where that he wished decorative images to have alt text. Regards, —Mattisse (Talk) 23:50, 11 August 2009 (UTC)[reply]
Yes, and that's why WP:FACR and WP:ALT emphasize "brief", and WP:ALT gives examples of why alt text should not emphasize colors. I don't recall anywhere he suggested that purely decorative images should have alt text. Eubulides (talk) 02:39, 12 August 2009 (UTC)[reply]
Graham said: "I find alt text helpful when it describes properties of an image that are easy to discern for sighted people but aren't appropriate for a caption (e.g. roughly what a logo looks like, the text of a cartoon). I really don't know what I'm missing out on when I've never seen an image. I think alt text should be encouraged in FA's, but not required where their use would be impractical." [3]SlimVirgintalk|contribs01:08, 12 August 2009 (UTC)[reply]
Re "is there any evidence that it [alt text] actually helps visually impaired people?"above:
Surely it is obvious that describing an article's images to a visually impaired person makes it easier for them to understand an article, for the same reason that including a transcript of the article's voice recordings makes it easier for the hearing impaired. However, to help allay any concerns that alt text is well-intentioned but misguided, I did a Google Scholar search, and there is indeed solid support among reliable sources for the obvious proposition that alt text helps visually impaired readers, or (the equivalent) that lack of alt text makes things more difficult for them. Here are three examples:
A 2007 study found that absent or low-quality alt text was one of the top five causes of frustration for blind users of the web. See Lazar J, Allen A, Kleinman J, Malarkey C (2007). "What frustrates screen reader users on the web: a study of 100 blind users". International Journal of Human–Computer Interaction. 22 (3): 247–69. doi:10.1080/10447310709336964.{{cite journal}}: CS1 maint: multiple names: authors list (link)
A 2008 paper on blind access to Wikipedia listed lack of alt text as the first obstacle. See Buzzi M, Leporini B (2008). "Is Wikipedia usable for the blind?". Proc 2008 W4A. Beijing. pp. 15–22. doi:10.1145/1368044.1368049. {{cite conference}}: Unknown parameter |booktitle= ignored (|book-title= suggested) (help)
The lack of alt text was also emphasized in a followup paper published last month, which focused on making it easier for blind readers to edit Wikipedia. See Buzzi M, Leporini B (2009). "Editing Wikipedia content by screen reader: easier interaction with the Accessible Rich Internet Applications suite". Disabil Rehabil Assist Technol. 4 (4): 264–75. doi:10.1080/17483100902903457. PMID19565387.
I'm sure that I could find and cite other sources supporting the same point.
In short, there's no real question that good alt text helps visually impaired readers. The only question is whether editors should bother to aid the visually impaired. Eubulides (talk) 17:43, 5 August 2009 (UTC)[reply]
Graham87 was saying that alt text is useful in describing a logo, for example. But in many of the examples we've seen of alt text, if not most, the descriptions are often quite unhelpful. The example I gave above was describing Immanuel Kant as a man in a wig looking to his right, or Clare College Cambridge as a three-storey building. It's hard to see how that would help someone, and yet this is what you're arguing should be essential for FA. The issue is that it's yet another stylistic burden on FA writers, when people are arguably already overloaded, and in an article with a lot of images that are tricky to describe well, it adds quite a bit of extra work. SlimVirgintalk|contribs18:02, 5 August 2009 (UTC)[reply]
I disagree that alt text should be omitted for the Kant image in the article where this came up. Please see #Alt text for portraits below for more on this. I also disagree that alt text is a substantial burden on writing FAs. In my experience (I haven't measured precisely) it's like 0.1% of the burden. Maybe less. That image is not hard to describe, and one can write alt text for it in a couple of minutes. The Clare College image was also easy to write alt text for. Eubulides (talk) 21:40, 5 August 2009 (UTC)[reply]
As SlimVirgin says, for images that identify people / places / various other subject alt text that literally describes the subject is useless.
Maps and technical diagrams are even worse - in the case of diagrams alt text that adequately describes the subject may be as long as a section of main text. Having bored everyone with the same example a few times, I offer another one File:4-Stroke-Engine.gif - can anyone adequately describe the 4-stroke cycle in a reasonable lentgh for alt text? --Philcha (talk) 19:30, 5 August 2009 (UTC)[reply]
Good point, but it seems useful to me to thrash out options here. Guess these subsections could always be moved up. That's a great example, and for its use in Four-stroke engine the appropriate alt text could presumably be alt="sectional view of an engine, with the piston rising up and down as its connecting rod rotates the crankshaft". The rest of the description of the cycle is given in the text which follows the image. Agreed? . dave souza, talk19:42, 5 August 2009 (UTC)[reply]
Yes, that's good alt text for this example image use. Let me underscore the point: Alt text is not that hard to write. Much of the opposition to alt text on this talk page seems to come from an exaggerated fear of how much work is involved. It really isn't that hard, typically. Eubulides (talk) 21:40, 5 August 2009 (UTC)[reply]
That complicated diagram is an extreme case, but consider the alternative: a visually impaired reader of the article will be missing information that was considered to be so important that a complex diagram was drawn up to present it. I'm no expert in arthropods, so it took me about 15 minutes to research the topic and write alt text for that diagram, which I installed. Problem solved, and it wasn't that much work.
I expect that an editor who knows the topic well could be able to do it in somewhat less time, and could come up with something more succinct. I don't know what's important, so I converted everything to alt text, making it more verbose than it probably needs to be. Eubulides (talk) 17:43, 5 August 2009 (UTC)[reply]
Are you certain that alt text would really be useful to a visually impaired reader? I'm certainly not. Graham mae the point above that he found a short description helpful, not an essay attempting to duplicate the contents of the article. --MalleusFatuorum18:26, 5 August 2009 (UTC)[reply]
As I said, I would expect an editor who knows the material to be able to come up with briefer version. But yes, the alt text I wrote, if technical errors in it are corrected, should be better than nothing, for a visually impaired reader. How could it be otherwise? It explains an important image that the reader cannot see. Eubulides (talk) 21:40, 5 August 2009 (UTC)[reply]
From my reading of WP:ALT, there seem to be two ways round this. One is to provide alt text providing the missing information that is shown only in the diagram. Not easy to do concisely, but necessary for readers dependent on the text. The other option is to treat the diagram as a visual aid to text in the main body of the article explaining the essentials of the information, in which case that body text would not be duplicated in the alt text. In that option it would then be possible to have no alt text and treat the diagram as purely decorative, or a simple description referring to the body text. Presumably it would be appropriate to add |link= to remove the link to the image, but apparently that can't be used where imatge attribution is required. Not sure if that can be used when there's a caption. . . dave souza, talk19:16, 5 August 2009 (UTC)[reply]
The first option is fine, of course. The second is also fine if the article text adequately describes the image. On my list of things to do is to fix the image attribution problem, for CC-BY-SA images anyway, so that one can use "|link=| with them; in the meantime we can muddle through acceptably with alt text that says the equivalent of "see article body". Eubulides (talk) 21:40, 5 August 2009 (UTC)[reply]
I'm afraid Eubulides' alt text for File:Arthropod head problem 01.png has inaccuracies, starting with the first sentence. I'm afraid it's also not much help, since you need to be able to read the segments both lengthways for one creature and crosswise to compare the sub-phyla. If the W3C had had the sense to allow markup in alt text and screen readers can read tables both ways, an HTML table might have helpful - but W3C missed that trick. If straight sequential text were useful such situations, we'd seldom bother with diagrams. --Philcha (talk)
Not being an arthropod expert I'm not surprised there are errors in my alt text, but they can be easily corrected and are not an argument against alt text per se. I also agree that my alt text is not as good as the image (if it were, we wouldn't need the image at all!) but this doesn't affect the point that the alt text is better than nothing for the visually impaired reader. Eubulides (talk) 21:40, 5 August 2009 (UTC)[reply]
I'm afraid that to me that just again reveals your blind spot. Are you seriously suggesting that an image wouldn't be helpful if what it showed could be described textually? Really? Seems to me that in your unseemly haste to offer help to the visually impaired you are also trying to reduce the value of the work to those who (in your eyes unfairly) are not visually impaired. I find that attitude to be grossly offensive. --MalleusFatuorum15:20, 6 August 2009 (UTC)[reply]
If the idea behind an image is better rendered as text, then yes, it should be rendered as text. I don't see how that can be controversial. Obviously if image and text renditions are roughly equal in usefulness, then editorial judgment comes into play. In this case, though, my comment said that the alt text was worse than the image. I don't see how that can reasonably be interpreted as criticizing the image, nor is there anything whatsoever in my comment that supports your claim about my opinions about fairness. Eubulides (talk) 21:43, 6 August 2009 (UTC)[reply]
dave souza's "two ways round this" are missing at least one, and the use of File:Arthropod head problem 01.png at Arthropod#Segmentation illustrates an additional intermediate possibility - that article does not describe in detail the Arthropod head problem, it aims only a quick illustration of the Arthropod head problem's complexity, to reinforce the point made by "Working out the evolutionary stages by which all these different combinations could have appeared is so difficult that ...". On the other hand the same image used at Arthropod head problem would be dave souza's 2nd option, "to treat the diagram as a visual aid to text in the main body of the article explaining the essentials of the information". --Philcha (talk) 19:48, 5 August 2009 (UTC)[reply]
Guess that at at Arthropod#Segmentation the appropriate alt text would be "the various sequences of head segments in x, y and z present a complex matrix with no evident evolutionary sequence". In every case the alt text will relate to the use in the article. . . dave souza, talk20:11, 5 August 2009 (UTC)[reply]
Yes, this briefer alt text would be better in that context. This illustrates two points. First, the person who's writing the article is often much better at writing the alt text and can do a better job at writing brief alt text. Second, alt text often depends on the context, for the same reason that captions depend on the context. Eubulides (talk) 21:40, 5 August 2009 (UTC)[reply]
Re "Who is helped, who is educated, by being told that, in Immanuel Kant, there is an image of a man in a wig looking downwards and to his right?"above:
In the context of the Bernard Williams #Critique of Kantianism article, who is helped by seeing the image of Kant at all? If readers are typically not helped by seeing the image, then it should be removed, alt text and all. Conversely, if in the context of the article the image's visual appearance aids understanding of the personality or philosophy of Kant (and I think it clearly does), then the alt text should say so. No doubt the current alt text could be improved, but that's no argument for removing alt text entirely. Eubulides (talk) 17:43, 5 August 2009 (UTC)[reply]
Right, and this is the issue that opens a can of worms. If we say an image is purely decorative and doesn't need alt text, you will respond, "Then remove it." And right there is another obstacle course for FA writers to weave their way through. SlimVirgintalk|contribs18:07, 5 August 2009 (UTC)[reply]
There is no obstacle course here. My argument did not say or imply that purely decorative images should be removed. (I regularly help to improve the accessibility of such images.) My argument was purely about whether that particular image is useful in that particular article. I argued that it is useful, and I think you agree that it is useful. Let's not hypothesize cans of worms where the cans do not exist. Eubulides (talk) 21:40, 5 August 2009 (UTC)[reply]
It's not absurd. It's common consensus. What is absurd is how my position has been mischaracterized here, mostly I expect due to lack of understanding of standard technical phrases like "purely decorative". Eubulides (talk) 21:40, 5 August 2009 (UTC)[reply]
In which case, maybe we should accept more use of decorative images? Here it's arguable that a brief alt text can convey something of use to the the reader, as suggested at the link given below. The issue isn't that it's a hurdle for FA writers, but an extra task for all editors. A bit of reasonable leeway for FA writers while we all get used to the idea is a good thing, but it's certainly a worthwhile improvement extending the usefulness of the project. . . dave souza, talk19:29, 5 August 2009 (UTC)[reply]
One could fairly easily, I think, make a convincing case that all images are decorative, to which my reply would be "So what?" The position apparently being advanced by Eubulides seems to me very much like banning colour, because it's not essential and not everyone can see it. --MalleusFatuorum20:44, 5 August 2009 (UTC)[reply]
The previous comment is based on a misunderstanding of the technical term "purely decorative", which I am using in the sense of W3C and WP:ALT to mean (roughly) "the picture doesn't add significant info to the article, and nothing happens when you click on it". In this sense, many of the images in Wikipedia are purely decorative, but many are not.
I would certainly accept more use of decorative images in Wikipedia. It shouldn't be overdone, of course, but I haven't yet run into an article where it has been overdone, and more use would be nice.
Colour should be banned, in the sense that one should not convey information by colour alone, in order to be accessible to the ~10% or so of the population that are colour blind. You are quite right; and as was your edit summary. --Xover (talk) 20:53, 5 August 2009 (UTC)[reply]
I'm probably going to sound quite callous to some, but this needs saying anyway. Sure we should do everything that can reasonably be done to provide a prosthetic environment for those with some visual impairment, and so no article should depend on colour or images, alt text or no alt text—by which definition all images are "decorative" and ought to be removed according to Eubulides. But the key word here is "reasonable". What about deaf readers who are unable to hear fragments of music? Why are they being ignored for instance? --MalleusFatuorum21:05, 5 August 2009 (UTC)[reply]
Yes, the key is "reasonable". It is reasonable to say that our best work supports access by the visually impaired? Obviously the answer ought to be yes, if it's not too much work. The main issue here that I can see is whether it's too much work. So far, the evidence has been presented in that arena has indicated that it's not too much work. For example, one specific examples alleged to be a "nightmare" turned out, when examined, to have alt text that could be written in a few minutes.
I would also like to support deaf and hearing-impaired users better, and would welcome further discussion about that (for example, transcripts for voice recordings). However, that's a different topic. It's also a less-pressing topic, as sound is far less commonly used in Wikipedia compared to images. I assume that the argument that color should be banned is meant to be a joke; nobody is arguing for that, just as nobody is arguing for images to be removed simply because they are purely decorative.
No, I am quite earnest in suggesting using color alone to convey information should not be allowed. It is a fairly important principle of accessibility work, and a quite real problem for the ~8% of the male population of the US that are color blind. It is also not a significant burden: most tables use color only for emphasis (it adds to information already present), and most infographics already are or can easily be adapted to follow this principle (e.g. instead of two bars in a chart being only different colors, make them also have different fill patterns). This would improve accessibility for more users (~5% vs. ~2.5%) with less effort than alternative text; but, of course, alt text solves a much more critical problem (and visuallly impaired users do find alt text a critical issue!). I very much agree that we in all ways possible should encourage the use of proper alt text, and that in my experience this is not a significant burden. The debate over alt text has been repeated on the web since the W3C first started working on accessibility guidelines, and it always goes the same: some will argue that it's such a burden (cost), and that it's impossible to get perfect so why even try (defeatist), and besides there aren't really that many blind users so it's hardly worth the bother (profit). Not to mention the always popular “But nobody else is using alt text, why do I have to!” (do we have wikimarkup for a whiny tone of voice?). This is why the equivalent of the W3C Web Content Accessibility Guidelines have been made a requirement in 29 U.S.C. Section 508 (along with things like wheelchair access etc.): some will always argue against doing anything at all, rather than accept the (usually minimal) added burden of striving for the best feasible level of accessibility. Don't set an unattainable standard to be sure, and do provide the necessary tools, but do provide every incentive (including force-measures such as policy requirements) for editors to make articles accessible to the widest range of people possible. In any case, I get the feeling the discussion has an odd emphasis on the “This is too hard, we can't do it” angle, rather than “This is how and to what degree we are practically able to do it” (on which latter point reasonable editors may disagree). The question really shouldn't be whether to require accessibility in the FA criteria, it should be which specific measures should be required, to what degree, and what tools do editors need (e.g. shared alt text on image page) in order to comply. The reference to Malleus' edit summary (reductio ad absurdum) was to underline the point that all claimed arguments that “by that argument, all images should be removed” or your own “all color should be removed” are strawmen that badly miss the point; as either argument or refutation they represent rhetoric rather than reasoned argument, and we can do better than that. --Xover (talk) 22:32, 5 August 2009 (UTC)[reply]
Sorry, I did not read your earlier comment carefully enough and just now struck my out-of-touch response. I wholeheartedly agree with you that color alone should not be used to convey significant information. Eubulides (talk) 23:44, 5 August 2009 (UTC)[reply]
More generally, the question of good alt text for portraits came up above, with the suggestion using "Photo of Joe Schmoe" as the alt text for a photograph of Joe Schmoe. While that would be adequate for images of people like Einstein where most readers can be expected to know the appearance, and it may also be adequate for lower-quality work, I would think that any high-quality article on a person should devote some space to what that person looked like, when this is known. With an image, it's easy: just show the image. If the image isn't visible, though, alt text can be quite useful in giving a brief impression of what the person looked and dressed like. There is some guidance for this in WP:ALT #What to specify, which uses a portrait as its example and contains a paragraph about portraits. Eubulides (talk) 17:43, 5 August 2009 (UTC)[reply]
Eubulides, we seem to be discussing 2 types of user who might hypothetically benefit from alt text:
Visually disadvantaged users. For them the subject's name and possibly age and circumstances are best, e.g. "Nelson dying on the deck of Victory". The current official WP approach would be something like "A small, slight man with only one arm and with a patch over one eye, wearing an early 19th century senior naval officer's uniform, lies on the deck of an early 19th century battleship, with blood seeping from a wound in his chest". It's hard enough to guess what the significance of this is to the article for a visually disadvantaged user who saw pictures of Nelson's death before the onset of their visual difficulties, and totally impossible for those who have not seen such a picture in the past.
Users who have forbidden their browsers to show images. Since they mostly have eyesight within "normal" parameters, they do not have years' of experience of dealing with only verbal descriptions both online and offline, and probably do not have the highly-trained short term memory that come with such long practice. If they've seen pics of Nelson dying before, "Nelson dying" is useful to these users. If they haven't, a detailed physical description will probably provoke "so what?"--Philcha (talk)
The previous comment seems to be confusing caption (which is where "Nelson dying on the deck of Victory" would normally be) and alt text (which would describe the visual appearance). The appropriate alt text for such an image would depend on which image and what the context is. Perhaps you could give a specific article and image so that we can better understand the problem that you see for visually disadvantaged users? (As Nelson didn't die on the deck of the Victory, perhaps this particular hypothetical example is not the best choice....) Eubulides (talk) 21:40, 5 August 2009 (UTC)[reply]
I think you make a good point Philcha. The point that doesn't seem to be getting enough attention is what do the audience for these alt texts want? I got the very strong impression from the comments made by Graham that he wanted a short description to give him an idea of what was in the picture, in which case "middle aged man in wig" might be perfectly OK for Kant's picture. The problem is that nobody seems to be clear about what makes good alt text from the consumer's pov, which is why I remain convinced that this half-baked addition to FA criteria was rushed and botched. These repeated claims that alt text is easy to write just miss the point. Sure, stringing words together is easy enough, but to what benefit if they're the wrong words? --MalleusFatuorum20:50, 5 August 2009 (UTC)[reply]
The current rule says alt text should be "brief", which addresses the point about brevity made above. The rest of the comment seems to be ignoring the content of WP:ALT, which is quite clear (and if it's not clear, specific suggestions for clarifying it are welcome). Eubulides (talk) 21:40, 5 August 2009 (UTC)[reply]
Fragment of Delafosse's (1904) linguistic map highlighting Nafaanra ("Nafana") in the borderland of Côte d'Ivoire and Ghana. Bonduku is found on the left.
Perhaps some more thought is needed before alt text is unleashed. The points by Philcha are very interesting. We cannot write alt text helpfully if the audience of readers is not clear, i.e. those who have no sight versus sighted persons with images turned off at the extremes.
Here is an example of an image I just plucked from an article where the editor had been asked to add alt text.
The alt text: map in black and white filled with (apparent) place names and different kinds of lines (normal, consisting of small crosses, of small points, and of orange-colored points that form two straight lines
I would think this alt text would be meaningless to someone who cannot see the map. Even seeing the map, it is not helpful. Alt text is now being added to many articles. Are we creating chaos for text readers? Do they have the option to "turn off" the reading of alt text, if all of this becomes too much? Graham said he liked concise, helpful alt explanations.
I think editors are implementing the new alt text requirement as best as they can. They are trying to follow the directions. But although the directions may seem clear and simple enough to those who can do alt text for an image in a few minutes, it is not so clear and easy for others.
I agree that that particular alt text is not that good, and have replaced it with the following:
'Map of two dozen locations in half a dozen regions. The central region's name, "NAFANA", is magnified in an inset. Other region names include "ABRON" and "NTAKIMA"; location names include "Bondoukou", the largest location in the NAFANA region.'
No attempt is being made to implement alt text "helter skelter"; instead, we're trying to encourage it one article at a time, as they come up for FAC and FAR. That is a methodical and gradual approach, and is the exact opposite of helter skelter. One can expect alt text to be done poorly at first, before people gain experience with it. But it is unreasonable to say that alt text should not be required merely because mistakes are sometimes made in writing it. Eubulides (talk) 23:44, 5 August 2009 (UTC)[reply]
One of the problems with the guideline is that it thinks the image caption, e.g. "Nelson dying on HMS Victory", is part of the package of info available to the blind user as part of the pic. But it's just some text in a container div (try view source). A blind user who understand HTML structure and the specific way they're used in WP images might be abble to use the caption as well as the alt text - but not the average blind person.
But then the W3C's thinking on alt text was either fuzzy or tokenism - a limited amount of straight text is pretty useless for any complex subject. So why expect WP to think it out better? --Philcha (talk) 21:38, 5 August 2009 (UTC)[reply]
I don't follow the previous two paragraphs; they seem to be either misunderstanding the W3C recomendations, or misunderstanding how screen readers work, or both. (Certainly the W3C is not just a bunch of fuzzy tokenist hackers.) Perhaps you could explain the point by citing a specific example of a particular image in a real article, and saying why alt text would not be useful for that image? Eubulides (talk) 23:44, 5 August 2009 (UTC)[reply]
I believe screen readers read the captions as well as the alt text. That is why it is important not to repeat what is in the caption. Graham said his did. —Mattisse (Talk) 21:55, 5 August 2009 (UTC)[reply]
What's fairly clear to me is that none of us really knows what needs to be done or how to do it. I've never used screen-reading software, for instance, and listened to the description of an image I've never seen. All of the alt text I've written has been (obviously) for images that I've seen. How the Hell do I, or anyone else not visually impaired, know what such a reader would find useful? What I do know though is that much of the alt text I've seen is just useless padding added to get through FAC, and that needs to change. --MalleusFatuorum23:18, 5 August 2009 (UTC)[reply]
This is incorrect. I have used a screen reader, and Graham87 exclusively uses a screen reader. Also, I can very roughly approximate the feeling of a screen reader by configuring my browser to not download images; perhaps you could try that? It's easy.
I agree that some of the alt text described above has not been good, and have fixed all specific examples mentioned. There will always be some problems with alt text, just as there are problems with all other article aspects, but trotting out the worst examples of alt text that one can find is not a good argument against the idea of alt text in general.
It's not true that a large part of the alt text added for FAC is "useless padding"; it may be useless to a sighted reader, but it is not useless to the visually impaired.
As has happened so often in this discussion Eubulides you completely miss the point. I am not blind, but yes, I do switch off images to check alt text, and most of the time I find it to be a complete waste of space. I have to say that I am coming round to SlimVirgin's view, and that if alt text remains a requirement for FA then I will not be submitting any more articles at FAC. My view is that this proposal was ill-considered and rushed through without proper thought. Now you can argue amongst yourselves, because I'm done here. --MalleusFatuorum23:58, 5 August 2009 (UTC)[reply]
How do we know what is useful for someone like Graham? For example, he changed the alt text to remove a fraction, he said it garbled the speech.[4] He has also made it clear that he prefers the text to be brief and concise, to give him a "rough idea". On the description of a seal, he said he appreciated the information it is round. I think we should go slow in implementing it, until we understand what is needed. I rather agree with Malleus that most of the alt text appears useless to ridiculous. And much of it is not brief nor concise. But I am not as polite as Graham is. Regards, —Mattisse (Talk) 00:08, 6 August 2009 (UTC)[reply]
Graham87 has said multiple times that he finds alt text useful and has made suggestions for improving it. We have been tracking this closely. His advice to be brief and/or concise appears multiple times in WP:ALT and is part of WP:FACR. His advice to remove the fraction now appears in WP:ALT #What not to specify as "Alt text should avoid tricky typographical constructions...". Further suggestions from Graham87 and other visually impaired users are welcome. The opinion that most of the alt text is "useless to ridiculous" is contradicted by Graham87's remarks, and having just re-reviewed the recently featured article The Bartered Bride (alt text), it's simply not true that most of the alt text there is "useless to ridiculous". Eubulides (talk) 06:59, 6 August 2009 (UTC)[reply]
He also said, somewhere on this page, that although he was in favor of alt text, it was not important enough to him to "bicker" over. Also, I think switching off images, and seeing the alt text only is not the same experience as one who has never seen images. Further, some editors appear overly invested in describing the details of their images, and some of it may verge on OR, like the "well-trimmed" mustache of the police chief; to me it look bushy. —Mattisse (Talk) 15:30, 6 August 2009 (UTC)[reply]
I would be happy to stop bickering over this issue and write more alt text and help others write alt text.
I agree that switching off images is not the same as using a screen reader; it's only a very rough approximation, as I've said. A better approximation is to install and use a screen reader, with your eyes closed; this isn't that hard to do, though I am not saying every Wikipedia editor should do it!
If editors generate too-long alt text, we can easily ask them to trim it; I've done that and in practice this works well.
I don't care whether the policeman's mustache should be described as "well-trimmed" or as "bushy" (it's both). Please feel free to change it to "bushy" if you like. Issues like these are clearly, clearly, not worth bickering over.
Your second point exposes a very dangerous assumption: for a fully sighted user to turn off images, use a screen-reader with his eyes closed, or the perennial favourite, to put on a blindfold and attempt to navigate around the streets; these are all seductively tempting, but absolute hogwash! Don't get me wrong, they will help you appreciate what types of difficulties may ensue, but they're absolutely useless in giving you a true understanding of how a visually impaired user, say, navigates the web. Especially for people who have been blind since birth or very early age, the difference isn't in the mere lack of sensor input from the optical sensor array; it's in how they process and refer to information, and the very minute practical details of how available technology actually operates. For them as well as those with a visual impairment that isn't total, or those who've lost their vision later in life, you will never approximate just how much they've adapted to processing information without relying on optical input. In this I absolutely agree with Matisse: writing good alt text is hard! The only way to develop a real feel for this is by talking to visually impaired users and so begin to understand what their needs actually are. For instance—and this is based on previous experience with attempting to make a web service accessible to the visually impaired—I'm willing to bet that the chaotic and high-volume nature of this discussion thread, the various technology artifacts (such as parenthesis, dashes, timestamps, random abbreviations, etcetera) are a much bigger problem to Graham than lack of alternative text is in most actual articles. In email conversations, for instance, it turns out (or so I'm told) that the common convention of interspersing your reply with the quoted text of the original actually depends on the recipient's ability to visually locate the reply text relative to the quoted text. Whereas if you write your reply as an unbroken narrative above or below the quoted text it is much easier to read for those relying on text to speech or braille technology. All of this is not to proclaim myself to have "the answer", but rather to point out that making assumptions is extremely dangerous here, and that the way to figure out what visually impaired users need is to talk to visually impaired users. --Xover (talk) 06:59, 7 August 2009 (UTC)[reply]
I also think that one of the examples of useful alt text that Graham specifically mentioned—the Earl of Carrick's seal—was very telling. He obviously would not have known it was round, and that's clearly helpful, but it ought also to make us consider what concept's a blind person can relate to and find meaningful. Clearly shape is one, because that can be felt, but equalyy clearly colour isn't. So for instance what would be the purpose of describing an image as "an expanse of green grass under a cloudless blue sky" to someone who has seen neither. The oft-quoted advice to think of trying to explain the image to someone over the telephone misses the point by a very wide mark, as it presupposes that the person to whom you are describing the image can see just as well as you can, and that concepts such as "grey", "translucent", "ripple", or even "reflection" have any real meaning to a blind person, Or at least that they convey the same meaning to a blind person that they do to us. Those who are claiming that writing alt text is "quick and easy" are talking out of their arses. --MalleusFatuorum16:45, 6 August 2009 (UTC)[reply]
Yes, when I said that, I was disheartened by how long and convoluted the discussion was, considering that we're only talking about one or two-sentence image descriptions. Probably because I'm borderline autistic, I dislike conflict and usually try to avoid it. I'm a habitual [[lurker] and only participate when I have something *really* important to say. I kinda thought, "if it's causing so much angst for certain people, then maybe it'd be better to drop it as an absolute requirement". I didn't realise there were so many shades of grey to the issue. I strongly support the use of alt text, as long as it's concise and conveys important information for people who can't see the images. Graham8716:39, 6 August 2009 (UTC)[reply]
Wonderful to hear from you! Concise I completely understand. But what is important to you? For example, there is a portrait of some men in Pavel Bure, and the alt text is "Two middle-aged men sit on a couch at left, and two men in their thirties sit on another at right. At center is a carved oval wooden table." Is it important that at center is a carved oval wooden table? —Mattisse (Talk) 16:51, 6 August 2009 (UTC)[reply]
Yes, it helps me form a better mental picture of the room. The type of table would set the scene ... I'd have a different mental picture if it was, say, a plastic outdoor table that's been placed in the room for some reason. I can't quite explain it properly. Any more detail about the table would be overkill IMO. A sighted person should probably check this edit. Graham8717:10, 6 August 2009 (UTC)[reply]
Your edit was correct, but how did you know? Also, that is good information about the table, and a good contrast between types of tables and how that changes the setting. —Mattisse (Talk) 18:12, 6 August 2009 (UTC)[reply]
Because of the alt text mentioning two men in their thirties sitting opposite two men in their fifties. I knew that Putin would've probably been in his fifties, and the simplest explanation for the incongruence in my mental picture was that Putin was listed as being on the wrong side. Graham8700:44, 7 August 2009 (UTC)[reply]
I agree about the difficulty, & the usefulness of Graham's comments, but blindness confirms my impression that relatively few people classified as blind have been completely so from birth, & a large number had normal sight until old age, disease or an accident. In many cases, these are just the people who would benefit from well-written alt text. Johnbod (talk) 16:56, 6 August 2009 (UTC)[reply]
I would think we would aim for alt text that would be helpful to all levels you mention, and certainly not be an interference, as alt text that is not concise. I notice that sometimes the skin color (e.g. black) is mentioned in the alt text, but usually ethnicity or race is not.
Image one - Alt text is two black men with in white clothes with faces turned to a large blackboard with written text
Image two - Alt text is A young child points. Behind the child a woman smiles broadly and points in the same direction.
Is there a policy about racial/ethnic descriptions in alt text?
Also, should the alt text start with a capital letter and end with a period?
And, as Malleus points out, how relevant is colour? Should we name colours as in a flag, but leave out "green" grass and "blue" sky colours? Or "brown" grass should be called "dry" grass or "dead" grass? Should we describe weather rather than colour of the sky?
Mustache "bushy" or "trim"? And should we describe the furniture in the room in a portrait? How to describe age? Elderly? Old? Aged? Young? (How young is young?) Distinguished? How do we describe Augusta, Lady Gregory, as in the portrait she appears "distinguished" or "handsome" but appears to be roughly the same age as the image captioned "Lady Gregory in later life". Just some questions I am curious about.
I think descriptions should be used rather than colours, but that's not always possible with, for example, blonde hair, which can be important. Whether the race of a person should be mentioned would depend on the context; it's propbably worth mentioning it in articles about race riots or similar. I suppose "dark clouds" would be another example; I know, from people explaining it to me, that they mean that it's about to rain heavily. How else would you describe them without introducing original research? Graham8702:13, 7 August 2009 (UTC)[reply]
I agree with Graham87. The general rule of thumb is to describe the gist of the image's visual appearance, roughly in the order that you noticed it. Focus on the visual aspects of the image that are relevant to the article. If color or race is important, put it in; otherwise, don't. Another way of saying this is that alt text depends on context, the same way that what one immediately notices about an image depends on context. As for the more-detailed questions:
I've been suggesting to people to punctuate alt text the same way they'd punctuate a caption.
It's better to say "brown grass" than "dead grass"; the former is visual appearance (which is what alt text is all about), the latter interpretation.
Whether to describe a person in a portrait as "distingushed", "older", etc. is a judgment call. One cannot reasonably expect a policy or guideline to spell out to editors exactly when to use which adjective in alt text, just as one cannot reasonably expect it for ordinary article text. The existence of judgment calls in writing alt text is no more an argument against alt text than it is against ordinary article text.
Re the "brown grass" you suggest, Graham87 says descriptions should be used rather than colours. People used to speak of the "golden hills of California" until it was learned that dead grass (dry grass) is actually a fire hazard and the "golden grass" description faded out. —Mattisse (Talk) 13:09, 7 August 2009 (UTC)[reply]
In other web sites I'd agree with you that "dead grass" would be fine, but Wikipedia has another core principle, verifiability, which says that if the alt text makes a challengeable claim about the picture, it needs to be supported. Alt text cannot contain <ref>s or other links, so realistically it can't contain citations. Claims about what the picture looks like don't need a citation, as the picture serves as its own documentation of what it looks like, but claims about meaning (e.g., this grass is dead, and not merely dormant) require a citation and therefore can't occur in alt text. For this particular case the point is relatively minor, but the general point is important and we might as well be consistent about it. (Sorry, I don't follow that "golden hills" and "fire hazard" comment; I don't see how it applies here.) Eubulides (talk) 05:54, 8 August 2009 (UTC)[reply]
← My first experience of alt text was on a recent FAR, where I was asked to add it to the image on the right, which you would think would be easy to describe. My first attempt was:
1. King's College Chapel and the Gibbs building photographed from The Backs, looking east across the River Cam. Four people are punting on the river, the punter standing at the back of the boat on the deck or till, which is the style in Cambridge.
Too specific, too many names, so I tried again:
2. King's College Chapel and the Gibbs building photographed from the River Cam. Four people are punting on the river, the punter standing at the back of the boat on the deck or till.
Still not right. Third attempt:
3. "Several old buildings next to a river, including a chapel that dates to the 15th century. Four people are punting on the river, the punter standing at the back of the boat holding a long pole, while the others are seated."
This time Eubilides did it for me:
4. Chapel in late Gothic style with a large window between two spires about eight stories tall, behind water and a green. Four people are punting on the water; the punter stands at the back of the boat and holds a long pole. On either side of the chapel are relatively nondescript three-story buildings.
Now, I can see that his is better than mine in certain ways, but there are issues I don't understand. Why are we not allowed to say it's King's College Chapel? We're not supposed to say "river," which everyone would understand, or River Cam, which would be more informative, but we are allowed to say Gothic, which many people won't understand. And one of the nondescript three-story buildings is the very beautiful Clare College. Given these issues with a relatively simple image like this, I foresee real problems at FA with certain articles, and a serious drain on people's time. By all means, if people want to add alt text to FAs, let them do so, and let's even encourage FA writers to do it, but to require it of them would be a mistake. SlimVirgintalk|contribs17:03, 6 August 2009 (UTC)[reply]
You are not supposed to repeat information in the caption, which is reasonable. Presumably this included the name of the building, and perhaps mentioned the river. Those are not spires, and "stories" is not a helpful way to describe a chuch. Nor are the surrounding buildings at all "nondescript" - a wildly POV statement. My attempt:
5. "Small river in foreground, with four people punting. A wide empty lawn leads to a large Late Gothic chapel with many pinnacles, seen end on with a large arched window, and flanked on either side by buildings in neo-classical style." Johnbod (talk) 17:23, 6 August 2009 (UTC)[reply]
Yours is good, though it does depend on people knowing what late Gothic and neo-classical refer to. The caption, by the way, was "Williams spent nearly 20 years at Cambridge, eight of them as Provost of King's." SlimVirgintalk|contribs17:26, 6 August 2009 (UTC)[reply]
(5) is excellent; it's better than (3) or (4) in my opinion, as it's concise and reads the way that a reader's eye would follow the scene. The Chapel's name should be in the caption, though, not the alt text, because the name is also useful to the sighted reader. A typical sighted reader of Wikipedia cannot reasonably be expected to know the name of that building (not everybody hangs out in Cambridge). Eubulides (talk) 03:06, 7 August 2009 (UTC)[reply]
SlimVirgin's 4 examples illustrate what's wrong with alt text for "portraits" (in the loose sense, including places / buidings). If we consider 2 blind users, of whom A lost his sight after seeing that scene or pictures of it while B never saw the location or a pic of it:
Alt texts 1 & 2 probably work for A, who will look up his internal "photo album".
Alt texts 3 & 4 do not help to identify the scene, and user A cannot differentiate it from any of the thousands of similar scenes that can be found in Britain - unless the mention of a punt leads him to guess it's either Oxford or Cambridge.
None of them is any good for user B. He will have no mental picture of the buildings named in alt texts 1 & 2. But alt texts 3 & 4 are no better for B: the physical descriptions do not assert notability in the real-world sense, i.e. no reason why he should care; and the architectural terms are unlikely to mean much to him. --Philcha (talk) 17:20, 6 August 2009 (UTC)[reply]
In the vast majority of cases users are like user B, not like user A, so let's forget about A for now. (3), (4) and (5) are all good for B, as they provide useful info about the image's appearance, and they are all much better than nothing for a visually impaired reader. There is no requirement or guideline that alt text must assert notability (I don't know why notability is even an issue here). It's quite plausible that B knows architectural terms like "pinnacle" and "Gothic", in which case these terms are obviously useful; and even if B doesn't know them, B can look them up in a dictionary. The question about what level of expertise one should expect of a reader is the same as for alt text as it is for regular article text; there's nothing new here. Eubulides (talk) 03:06, 7 August 2009 (UTC)[reply]
I'm for the idea of alt text if it assists the visually impaired. I added alt text to all my FAs fairly soon in the discussion. I left one out of Mulholland Drive (film). As usual, I forgot about it. User:Erik brought it to my attention in this thread on my talk page.
So the issues involved are these:
The image is associated with the article. It is illustrative. It adds nothing really to the understanding of the film, but it accompanies a discussion of its reception at Cannes.
Erik has suggested adding a rather extensive alt text. I object to this. The caption describes the image. No alt text is necessary in my opinion.
Erik's suggested alt text is more descriptive than what is necessary, and much more descriptive than the alt text accompanying images that the article actually discusses. So the least informative image has the most alt text.
So the question arose: should the image be in the article at all if it adds nothing to the understanding of the article? What does this mean? If an image is so basic that it needs no alt text beyond the caption, does that automatically warrant its removal? Does this new criteria mandate that public domain or free license images should be removed if it has no alt text, even if the caption explains what is in the image? Do all images require alt text on top of captions, or are there appropriate cases where captions are sufficient?
I declined to add the alt text, using my best judgment. Erik added it himself.
This is starting to remind me of the haphazard way image policy is applied. No one understands what's going on. We don't have a clear idea of what is required. Gray areas should be discussed, and I hope we can come to a collective understanding before edit wars occur. I personally prefer mutual understanding over making edits to articles despite what I consider to be valid disagreements. I think we should err on the side of holding off of article space edits until we can say what should be there. --Moni3 (talk) 03:07, 7 August 2009 (UTC)[reply]
Alt text does not affect whether an image is appropriate for an article. If this image helps people understand the film's reception, it's a valid image for that article. Conversely, if it does not add any understanding to the article, it should be removed, regardless of the alt text issue.
That caption gave some details about the visual appearance of the image, but not very many. It wasn't terrible, but it wasn't a great description of what the image looks like.
It's fine for alt text to be briefer when an article discusses an image in detail. After all, alt text shouldn't duplicate material that the visually impaired reader can already read. So it's not at all paradoxical for the least-discussed image to have the most alt text.
I attempted to help out redoing the alt text along the lines suggested in the previous bullets.
For specific questions like this, WP:ALT suggests asking at WT:ALT; perhaps we can follow up there if there are further questions.
Conversely, if it does not add any understanding to the article, it should be removed, regardless of the alt text issue. I have a significant problem with this, or what I take it to mean. The image does not assist the reader in understanding the multiple interpretations of the film. It does not assist with grasping Lynch's characteristic style, the poisonous culture of Hollywood, or the emotions involved in the story and characters. But it is related to the article because the film was very well-received at Cannes and Lynch got its most prestigious award. If I added to the caption a bit more information, let's say making the current alt text unnecessary (I still think it is), would you still be recommending its removal? Why is one person interpreting or clarifying these issues? How do I know you are right? This is why it reminds me of image policy. Everyone tries to explain it, but everyone has a different explanation.
You changed the meaning of the alt text I provided for Image:BettyMulholland.jpg, making it more about Watts' clothing and hair than her demeanor and the light used to bathe her character, but these issues are specifically mentioned and cited in the article. Yes, I understand many blind readers will not understand the issues with light, but it's an integral part of filmmaking and I don't think we should sell readers short in discussing it.
Will we be neglecting sighted readers by removing images where the caption is sufficient for alt text, or creating confusing overzealous alt text just to keep an image before an alt text guard comes through FAs and removes free licensed images with no text? These are problematic issues that should be settled. --Moni3 (talk) 11:39, 7 August 2009 (UTC)[reply]
My main point is that alt text should not affect the decision whether to include the image; I assume we agree on that point. I'd rather that this discussion not get distracted by disagreements over rules of thumb over whether to include an image. That is an orthogonal issue, which can be taken up in a different thread.
It would be fine to move info from the alt text to the caption, if that improves the caption.
As per WP:ALT and the W3G guidelines, we cannot remove the alt text entirely; at the very least it should have a placeholder referring the reader to the visual description of the image in the caption or the main text. (This point has come up a lot in this thread, so I added it to WP:ALT.)
"How do I know you are right?" The previous bullet is not just my unsupported opinion: it follows directly from what's long been in WP:ALT (well before I volunteered in this area) and what's in the WCAG 2.0 guidelines (see guideline 1.1).
"Why is one person interpreting or clarifying these issues?" Because I volunteered and nobody else has. I'd be happy to have help. I'd be even happier to retire from the job and let someone else do it. I don't particularly want to be the alt text guy, but I feel that someone should do it.
It's fine to restore info about lighting into the alt text. As usual, it's a matter of editorial judgment whether a particular detail (in this case lighting) is worth mentioning. I was trying to fix gross problems with the alt text and keep it brief; whether that detail is included is relatively unimportant.
"Will we be neglecting sighted readers by removing images where the caption is sufficient for alt text, or creating confusing overzealous alt text just to keep an image ...". No. Absolutely not. Please see the first bullet in this comment.
(Xover made several good points in a comment indented so deeply in #Alt text for portraits above that I've started a new subsection. Naturally, I'm not going to comment on all the stuff I agree with, only on the places where I'll quibble.)
"I'm willing to bet that the chaotic and high-volume nature of this discussion thread ... are a much bigger problem to Graham than lack of alternative text is in most actual articles." Absolutely. And a blind person accessing Wikipedia faces other problems even more daunting. Alt text does not solve all of the accessibility problems a visually impaired person has, compared to a sighted reader. It doesn't even solve half of them. All that alt text does is help visually impaired readers grok important parts of articles that they can't see. This doesn't make Wikipedia perfect; it merely makes it better.
"for a fully sighted user to turn off images, use a screen-reader with his eyes closed ... are ... absolutely useless in giving you a true understanding of how a visually impaired user, say, navigates the web." It's quite true that these hacks are only approximations, but they are not absolutely useless, any more than taking driving lessons is absolutely useless for understanding what it's like to drive a car by yourself. For people who do not normally think about the subject, these hacks can provide valuable insight what sort of problems visually-impaired readers run into. The hacks are not the same thing as being blind, of course, and nobody has claimed that they are; but that is a far cry from being "absolutely useless".
For more on the bigger picture about problems that visually impaired people have with technology in general and with web access in particular, please see Shinohara K, Tenenberg J (2009). "A blind person's interactions with technology". Comm ACM. 52 (8): 58–66. doi:10.1145/1536616.1536636.
"writing good alt text is hard!" Actually, no, in my experience, it's typically not hard. I write the alt text, and Graham likes the result. And I am not some super-special alt text guru; I just look at the image and write down what I see, limiting myself to what's obvious and what's not in the caption.
I sense that a lot of the frustration with alt text comes from the desire to make alt text perfect. Please don't try to do that. Just spend a few minutes to make it as good as you can. Then stop. This won't take much time, and you're almost sure to improve the article.
Here's another way to put it. The audience for alt text is not the same as the audience for article text, and alt text does not follow the same standards as article text. It's much more important for alt text to be functional, than for it to be brilliant prose. Please don't waste time trying to make alt text be brilliant prose. (We have better things to do.) Just make it a concise summary of visual appearance. Eubulides (talk) 08:41, 7 August 2009 (UTC)[reply]
The formulation “absolutely useless” was of course somewhat hyperbolic, but I do think it's an important point to make: all experience shows that making assumptions about the actual needs of visually impaired users is more likely than not to lead one into making the wrong or at best suboptimal choices. The discussion here is informed by Graham's participation, but the issues are not clear cut; and while having Graham's input improves our understanding immensely, even relying on his' alone is not perfect. However confident we feel in our understanding now, we need to actively seek input from visually impaired users (the WAITooltip Web Accessibility Initiative xtech list or education and outreach may be able to help here) and expect to keep honing our skills and understanding over time. The main ting about “switching your eyes off” in order to learn what a blind user will "see" is that you then add the limitation without the coping ability. Your average blind user is infinitely more adept at using a screen reader and braille display than any amount of these sorts of exercises can teach you. If you perform this exercise without that fact firmly in mind you'll end up simulating the wrong thing. Not useless, but with potential to be quite misleading. As to your third point, writing good alt text is hard; but that's not an argument against writing minimally acceptable alt text right away. I understand that you're currently focussed on answering those who argue against making alt text a FA requirement, which colors the focus of your responses (or at least they would so color mine), but sweeping the very real challenges of writing good alt text under the carpet will not do visually impaired users any favours. We need to start as our ability allows, fail at it, learn from it, and then get better; rinse and repeat. Case in point, different users (as simply different people), using different assistive technologies (basic text to speech, vs. screen readers, vs. braille displays, vs. ...), or with different forms or levels of visual impairment, will all need different things from the alt text (ranging from negligible to significant differences). In any case, as best I can tell we're more or less in perfect agreement. I'm just trying to inject a little moderation into the “It's so easy, don't worry about it” argument in order to alleviate my concern that we may end up turning simple and easy into stupid and useless. The road to hell being, as we know, quite well paved and regularly maintained. --Xover (talk) 09:26, 7 August 2009 (UTC)[reply]
Hear, hear. I guess I'm overreacting to stories about it taking X editors Y days to add alt text to an article. (Next time I'll try to remember to write, "It's not that hard.") Eubulides (talk) 09:38, 7 August 2009 (UTC)[reply]
Eubulides's simplistic comment "Just make it a concise summary of visual appearance" (08:41, 7 August 2009) shows that WP has not through this issue and the most likely result is a requirement that is both too laborious and of little use to the blind. To summarise my comments in earlier sub-sections:
Four types of arthropods showing the acron and 9 head and/or body segments. Trilobites and chelicerates are shown with 7 head segments, and crustaceans and tracheates with 5 head segments. Of these, the first head segment of chelicerates and the second head segment of trachates is lost in development. All four start with an acron at the anterior end bearing compound eyes. All have nephridia on some or all head segments, some of which are lost in development in chelicerates. All - other than chelicerates - have antennae on the first head segment, and crustaceans also have antennae on the second head segment. Only chelicerans have chelicera, on the second head segment and first body segment, and pedipalps, on the third body segment. Crustaceans have mandibles on the third head segment and maxillae on each of the fourth and fifth head segments. Trilobites and chelicerates bear legs on all remaining head segments, but crustaceans and tracheates have legs on the anterior body segments.
This totally confuses me, and I created the diagram, read the cited sources thoroughly, and first read about this topic in 1991 (Gould's Wonderful Life ). Dave souza and I had a discussion further up the page that identified 3 possibilites, and there may be more:
It's part of the main article on the subject, e.g. at Arthropod head problem, and the alt text should simply say e.g. "Head structures in main arthropod sub-phyla - see main text for details"
In some other cases provide alt text providing the missing information that is shown only in the diagram. Not easy to do concisely, but necessary for readers dependent on the text.
An additional intermediate possibility - Arthropod does not describe in detail the Arthropod head problem, it aims for only a quick illustration of the Arthropod head problem's complexity, to reinforce the point made by the mai text "Working out the evolutionary stages by which all these different combinations could have appeared is so difficult that ...". Dave suggested "the various sequences of head segments in trilobites, crustaceans, chelicerates and tracheate arthropods present a complex matrix with no evident evolutionary sequence".
For "portraits" that identify individual subjects (people, places, buildings), we can't provide enough details to identify the subject in the limited length of alt text, and it would be difficult to describe the subject unambiguously even in a page or 2 of text. For example can you guess what the subject of this is: "Small river in foreground, with four people punting. A wide empty lawn leads to a large Late Gothic chapel with many pinnacles, seen end on with a large arched window, and flanked on either side by buildings in neo-classical style"? In addition this assumes the blind reader understands the architectural terms "Late Gothic" and "neo-classical". I'd much prefer "View of A from over the River B, showing C on the left and D in the centre". But (small, because it's the answer to the puzzle) "King's College Chapel and the Gibbs building photographed from The Backs, looking east across the River Cam", which IMO is better becuase it tell the blind reader the significance of the pic, was rejected at a recent FAR.
Anyone who wants to draw up rules for alt text needs to study a few hundred pics, of different types and with some appearing in more than 1 article, to begin to provide a useful guide. --Philcha (talk) 10:29, 7 August 2009 (UTC)[reply]
I have studied hundreds of pictures used in Wikipedia articles, and have summarized what I've found in WP:ALT next to helpful advice from others. WP:ALT is a work in progress and obviously can be improved, but its general principles apply to a wide variety of images.
"This totally confuses me". Which part of that alt text is confusing? Every statement in it is clear. There are trivial errors in it, but they can be fixed easily. That alt text is nowhere near as concise and visual as the diagram, but that's common for alt text: we cannot expect miracles. The audience here is expert in reading text and building up mental images as they go, so it's not so bad as one might think.
If the body of an article contains all the information that would normally be in alt text, it's fine for the alt text to be a placeholder that merely says where the description of the visual appearance is. This avoids repeating the info, which is a win. (This point has come up a lot in this thread, so I added it to WP:ALT.)
Dave's suggestion for the alt text would be fine (but see quibble in the next bullet) for an article where the gist is "This stuff is really complicated! Just look at that diagram! But don't bother to try to understand it." At the other extreme, if the diagram is intended to fully explain the problem, and no other text in the article explains it fully, then the long-winded and complete alt text is the better alternative (or at least, whatever part of it is not duplicated elsewhere).
A quibble about Dave's suggestion: "evolutionary sequence" would be better just to say "pattern", as this is both shorter and draws fewer editorial conclusions.
"it would be difficult to describe the subject unambiguously even in a page or 2 of text" There's no requirement, or even suggestion, that alt text must uniquely or unambiguously identify an image. Identification is not alt text's job, and it would be completely impractical to insist that it be alt text's job. If a unique ID or name is important, that info should be in the caption anyway; and alt text should not duplicate the caption. Thanks for bringing up this question; to help make this clearer I added to WP:ALT #What not to specify.
Also, alt text should not undertake to explain the significance of the image, or to explain details such as naming or authorship that are not immediately obvious from the image itself. Any such explanation (a) should be sourced, and (b) should appear in the caption or the article text, so that sighted readers can see it too.
As mentioned above, that example assumed (wrongly as it turned out) that the location was named in the caption, & was therefore not repeated in alt text. Also above, it seems that your preferred version was "rejected" by Slim Virgin, who wrote it. However I agree writing good (not "brilliant", which no one has suggested) alt prose is difficult. As to names, see my comment way above - WP:ALT assumes that the subject will be adequately covered in the caption, which will usually be correct, and then, over-simplistically in my view, concentrates entirely on describing the composition, ignoring other elements such as style and iconography. Johnbod (talk) 11:17, 7 August 2009 (UTC)[reply]
That's a good point: if style is an important visual aspects of the image, then the alt text should describe it (assuming that this does not duplicate the caption or article text). I added a mention of style and composition to WP:ALT #What to specify. Iconography would be more of a problem, though, as it is often so focused on meaning and interpretation that it would require a source and therefore cannot appear in alt text. Perhaps an example of reasonable (and/or inadvisable) use of iconography in alt text would be useful for WP:ALT; can you suggest something along those lines? Eubulides (talk) 05:54, 8 August 2009 (UTC)[reply]
Mentioning the royal diadem on the Greek coin above is a simple example that is sufficiently common-knowledge not to need sourcing. For some readers a mention will be enough; unfortunately there isn't realistically room to describe or explain it in an alt caption for the benefit of others, & you can't use links. I suppose the policeman's uniform and radio in the WP:Alt photo example fall under iconography, which is obviously covered ok in the example as it is now. If describing the Narmer PaletteFile:NarmerPalette-ROM-back.jpg one would want to mention the relative size of the figures, and so on. This image (see Retablo (Latin America)) would be one where both style and iconography have to be dealt with - perhaps one of the "quick and easy" school would like to have a go? Johnbod (talk) 13:16, 8 August 2009 (UTC)[reply]
Just realised that great works of art are another very difficult category, because describing what makes them great is a matter of interpretation. The "physical description" approach currently recommended in WP:ALT would often produce alt text at least as long as in the current version of {{Annotated image/Arthropod head problem}}. Try for example Dali's Christ of Saint John of the Cross, where my first reaction on seeing the actual painting (Kelvingrove Art Gallery and Museum, Glasgow) was "Wow, he's shown the crucifiction from two perspectives at the same time, from above (God) and below (humanity)". I can also remember how a Japanese painting (pen & wash?) of a kingfisher sitting on a bulrush impressed me - this captured the essence of the scene in under 30 brush strokes.
Thanks for bringing up Christ of Saint John of the Cross. In that article, the article text already describes the painting's visual appearance in great detail. For art articles like this, brief alt text would be appropriate, e.g., "See adjacent text." I added words to that effect to WP:ALT #What not to specify.
Music is a different issue. It's not a "nightmare", in the sense that hearing impaired readers have much better access to Wikipedia than the visually impaired do, mostly because audio is much rarer than images in Wikipedia. Also, the MediaWiki technology used for audio is so different that most of this discussion is irrelevant to it. Since this thread is about visual impairment, I'd rather take up audio in a different thread, if there's any interest.
Agree. I think the point that Xover is making that Eubulides brushes aside with the driving lessons example is that a sightless person thinks differently, uses the brain differently, and this cannot be mimicked by putting on a blindfold. The fact that Graham87 was the only person who figured out Putin was on the left and not the right in the photo example above shows that.[5] None of us sighted people noticed that. If we simplify the process into "just give a simple description", we will be missing the boat completely. —Mattisse (Talk) 23:42, 7 August 2009 (UTC)[reply]
I think you mis-state the case somewhat with your Putin example. None of us sighted people really bothered to read the alt text very carefully would be more accurate, as we could already see that Putin was on the left, and so automatically saw that in the text. Regardless, without clear guidelines and a clear definition of who our audience is for this alt text I remain convinced that adding it is mostly just busy-work. I'm all for doing whatever we can to help visually impaired editors, but we need to know first of all what it is that they find useful, not what some half-baked project thinks they'd find useful. --MalleusFatuorum00:12, 8 August 2009 (UTC)[reply]
I think that until the problem that alt text is supposedly fixing is properly defined (people who have never seen, people who lost their sight later in life, ...) then it's just a waste of time to discuss anything, because the result will be as ill-defined as the problem currently is. I'm quite happy to let anyone who feels that they're in some way enhancing wikipedia's accessibility by adding alt text for an audience they don't understand to continuing doing so, but I simply don't feel inclined to do so myself. --MalleusFatuorum00:57, 8 August 2009 (UTC)[reply]
I agree that until the audience is defined, as the two groups you mention would have different needs, that it is a useless endeavor. I have stopped suggesting it in GA reviews. —Mattisse (Talk) 01:04, 8 August 2009 (UTC)[reply]
This doesn't just narrowly apply to visually impaired editors, as I think Philcha said above. My brother became deaf when he was in his twenties, so I can describe pop songs to him by comparing them to the ones he heard, and of course he can "feel" the beat of the music. But if he'd been born deaf? --MalleusFatuorum01:22, 8 August 2009 (UTC)[reply]
Alt text is a simple technology that aims to help visually impaired people, regardless of whether they're congenitally blind (which is relatively rare) or became blind after birth, or have some sight (more common).
Different visually impaired people have different expertise and needs, just as Wikipedia users in general have different expertise and needs. That does not mean that alt text does not help the visually impaired, any more than it means that article text does not help readers at large. Similarly, the existence of differences among visually impaired readers does not mean that alt text is fatally flawed because it cannot address all visually impaired readers equally, any more than the existence of differences among sighted readers means that Wikipedia article text is fatally flawed because it cannot address all sighted readers equally well.
It is a resaonably adequate engineering solution to have sighted editors describe images as best they can, using the guidelines in WP:ALT. There is no need to insist on perfection here, or to refuse to add alt text unless unanimity is achieved on alt text's goals, any more than there is a need to insist on perfection in article text, or to refuse to add article text unless unanimity can be achieved on article text's goals.
"the problem that alt text is supposedly fixing" This phraseology is contradicted by the strong support for alt text from Graham87 as well as by its strong support by reliable sources such as those at the start of #Alt text helps the visually impaired. There is no controversy among reliable sources that alt text helps the visually impaired.
"a clear definition of who our audience is" Thanks for bringing up this problem. For the intended audience of alt text, I added to WP:ALT #Why it is useful a brief summary of some characteristics of visually impaired readers. If this is not clear, or if there's something wrong with this description, can you please specify the problem and propose wording that would fix it?
"clear guidelines" Likewise: if the alt text guideline is not clear then let's fix the problem. What about WP:ALT, exactly, is not clear? Can you propose wording that would make it clearer?
I'm kinda skimming, as always, I fear. It seems Eubulides is the principal shepherd of this proposal. So then, Eubulides, would you be willing to play the role of AltTextShepherd at FAC? Rather than passively Opposing 'cause a nom 'cause it has no AltText, or its AltText sucks or whatever, instead just, you know, {{sofixit}}, with appropriate comments to the nominator. Consider it an example of seizing a teachable moment. Ling.Nut (talk) 10:26, 8 August 2009 (UTC)[reply]
Just to say that in my experience Eubulides did just that, pointing to the guideline and commenting helpfully on proposals. Since implementation has to start somewhere, guiding our proposed best articles to a reasonably acceptable standard of alt text is a good place to begin. Thanks for that assistance. . . dave souza, talk15:11, 8 August 2009 (UTC)[reply]
I think we should try to cater alt text to people who have never seen before and people who aren't familiar with the subject of the article. I have no idea how to approach alt text for complex paintings like the ones cited above. I once went to the Art Gallery of Western Australia when they had an exhibition of Indigenous Australian art designed for the blind; the tactile diagrams made no sense to me at all. Sighted people told me that they couldn't understand some of the paintings either.
Helen Keller is an interesting case because she had normal hearing and sight until she was struck with scarlet fever. See a letter her teacher, Anne Sullivan, wrote on September 18, 1887; use the find function in your browser to get to it and search for "September 18". In the letter, Sullivan speculates that she has a basic knowledge of colour and other concepts from when she could see before the age of 19 months. In my case, I had retinopathy of prematurity, so I was born with normal sight but lost it shortly after birth. Graham8706:47, 8 August 2009 (UTC)[reply]
What's the practice when existing FAs are discovered to have not a jot of alt text for their images? Is it a requirement? Tony(talk)13:45, 8 March 2010 (UTC)[reply]
I can't speak for others, but when that requirement rolled out before my last FAC nomination, I went back to my previous FAs and added the ALT text. At the same time, I pinged Eubulides for help to make sure that I added was good on one article before I went through the two. I would hope that other editors who are somewhat prolific at FAs have a similar plan in mind. Beyond that, I'm sure that they will be FAR-ed and updated eventually, but I would hope that they are kept up after FACs. Imzadi1979 (talk) 14:01, 8 March 2010 (UTC)[reply]
Well, I posted a note about it at the article talk page. I guess I'll return in a few days to see if anyone is going to add the alt text. There are many images. Tony(talk)14:38, 8 March 2010 (UTC)[reply]
Given that we have about 2,000 FAs that predate the alt text requirement, obviously this will take some time, especially since the nominators may no longer be active in the article (or be active at all). Plainly articles submitted for TFA consideration should have alt text.--Wehwalt (talk) 15:12, 8 March 2010 (UTC)[reply]
I am gradually adding alt text to my FA noms that preceded the alt text rule. I've done most, but as Wehwalt says, it will take time. Brianboulton (talk) 22:02, 8 March 2010 (UTC)[reply]
Since alt text, while useful, is not vital for readers (even those who cannot view the images), I think we should allow a grace period for older FAs; many users are still not even aware that such a featured exists, let alone know how to write it or are aware that it is required in FAs now. Dabomb87 (talk) 22:41, 8 March 2010 (UTC)[reply]
Letting people know on the talk page in question is the best bet here: alt text is rather trivial to add and only takes a few minutes to read the guidelines and figure out what you're supposed to be doing anyhow. FARs for alt text seems... excessive. Der Wohltemperierte Fuchs(talk)23:22, 8 March 2010 (UTC)[reply]
I wouldn't FAR solely on lack of ALT text, but any FA that's left alone from the day it was promoted will eventually be taken to FARC from some reason, and if ALT text hasn't been added before then, it should be added then. Imzadi1979 (talk) 01:47, 9 March 2010 (UTC)[reply]
Yes, that's common practice. I don't recall seeing any FAR solely for lack of alt text, and I've read all the FARs since the alt text requirement was added. Eubulides (talk) 02:53, 9 March 2010 (UTC)[reply]
I think including the use of alt texts for images as an FA requirement is silly and can be at odds with other criteria. For example, one of the examples listed at WP:ALT for Mary Bartelme lists "Head and shoulders of a serious and dignified woman in her forties, with dark hair up and in a dress with high lace collar and a cameo at her throat, Edwardian style" as a model description. The words "serious and dignified" are totally subjective and POV, and indicating her age is not really helpful given people age at different rates. Then there's the diagram example: "Carbonated hydroxyapatite enamel crystal is demineralized by acid in plaque and becomes partly dissolved crystal. This in turn is remineralized by fluoride in plaque to become fluoroapatite-like coating on remineralized crystal." That's basically the exact same thing you would read in prose, and thus is no more helpful to visually-impaired editors than whatever could be expressed in the article text. Frankly, nt every visual image can be adequately described in words. I think it's a nice idea to have alt text for images, but to make it a requirement to pass FAC? I don't see the sense in that. WesleyDodds (talk) 05:59, 9 March 2010 (UTC)[reply]
I agree with WesleyDodds's "it's a nice idea to have alt text for images, but to make it a requirement to pass FAC?"
The ALT attribute is suitable for only brief identification, as anything more could often duplicate text in the main content. The LONGDESCR attribute could avoid duplication, by using an internal anchor (e.g. "#LD_img_topic"), but WWC says LONGDESCR is poorly supported by browsers, and AFAIK Mediawiki does not support LONGDESCR.
The current guideline says alt desc should be a physical description. This has multiple issues, e.g.:
It's no use to readers who are blind from birth.
Sometimes a physical description is irrelevant as well as a waste of editors' time. E.g. "Lord Nelson, supported by Hardie, is dying on HMS Victory while the English fleet devastated the French" explain the significance of the scene. OTOH a physical description would be e.g. "A man of X age, in the (long description of 1805 UK admiral's uniform) lies dying of a bullet wound (when?) on the deck of a (long description of UK first-rate builder in late 18-cent). A second man, (long description of 1805 UK naval uniform of Hardie's rank), sits and supports the dying man ..."
A physical description of a concept looks unlikely to be helpful. E.g. the "crown and stem groups" dgm at Opabinia#Theoretical_significance.
If there are problems with the Alt-text guideline or you feel it shouldn't be a guideline, please discuss that at WT:ALT. This page is for discussing FAC. Colin°Talk09:20, 9 March 2010 (UTC)[reply]
We're discussing how it applies to the FA process. Even WP:ALT says alt text is "recommended" but nothing firmer than that, so why should it be mandatory for articles coming through FAC? WesleyDodds (talk) 10:36, 9 March 2010 (UTC)[reply]
It started that way (and that discussion belongs here) but the last two lengthy posts deviated. Please keep this relevant. WT:FAC is not the place for guidelines and policy to be worked out. I don't know what version of WP:ALT you are reading but mine says "Every image should have alt text, unless the image is purely decorative and does nothing when you click it." It is absolutely firm on that, in as much as any guideline can be. Colin°Talk11:01, 9 March 2010 (UTC)[reply]
And the discussion of whether the WP:ALT guideline should be enforced at FAC at all belongs at WT:FA?. This discussion is about what to do about existing FAs that were promoted prior to the criteria change. Colin°Talk11:18, 9 March 2010 (UTC)[reply]
In the "Goal" section, it says "For all these reasons, the Wikipedia image use policy recommends alt text for Wikipedia images". Also guidelines are simply that: guidelines. Why is it now essential to have alt-text as part of the Featured Article Criteria? I understand that this section started discussing what to do about prior FAs (which itself could also be a discussion carried out at WT:FA?, if you want to get technical), but I've just heard about the alt text requirement and wanted to bring up some points I feel are necessary to consider. as this is the plce where the criteria is put into effect, I feel it's important to examine the points I brought up here, given alt text captions seem prone to OR/POV descriptions or can be pretty redundant to existing prose. WesleyDodds (talk) 11:28, 9 March 2010 (UTC)[reply]
Policy recommends guideline. That's all the quote says. The guideline itself mandates alt text. I understand you have points to bring up and I'm in no way trying to suppress those. I'm just saying that this discussion shouldn't be derailed by a discussion that belongs elsewhere. Why should a discussion on whether "alt text captions seem prone to OR/POV descriptions or can be pretty redundant to existing prose" be conducted here and limited to the audience gathered here? Discuss the guideline on the guideline talk page. Advertise that discussion here if you want. Colin°Talk13:20, 9 March 2010 (UTC)[reply]
Colin, the guideline doesn't mandate ALT test. Guidelines don't mandate anything; they are advisory. But even in its own terms, that guideline only encourages ALT text, as does the MoS (or did when I last looked). SlimVirginTALKcontribs01:51, 10 March 2010 (UTC)[reply]
My concern isn't necessarily with the guideline itself (although it has its faults). After all, that why we have "Ignore all rules". My concern is that this guideline is one of the few explicitly singled out as necessary for meeting the FA criteria. WesleyDodds (talk) 13:41, 9 March 2010 (UTC)[reply]
Yeah, I get you, but the required vs. recommended thing doesn't work well at FA. Some reviewers intepret "recommended" as required, anyway, and on the other side of the spectrum, nominators translate "recommended" as "not required, so I don't need to follow it". Either way, I can see a lot of unproductive discussion springing from the introduction of a little leeway into the criteria. Dabomb87 (talk) 13:52, 9 March 2010 (UTC)[reply]
I'm not saying there should be leeway in the criteria. What I'm saying is that alt text for images is a rather odd thing to make a requirement of the featured article critieria. WesleyDodds (talk) 03:07, 11 March 2010 (UTC)[reply]
The addition of the alt text requirement at FAC can be viewed in the context of the 2005 change that required inline citation after the Seigenthaler incident. At the time the change was added to WIAFA, more than half of FAs were out of compliance; it took three years to bring those articles to compliance, and it was a slow, deliberate process. FARing articles for alt text-- a guideline-- would be silly, and writing alt text is a form that not all editors have mastered (moi, for example). We can gently ask for it on talk, but making a big to-do over it isn't likely to be productive. An older FA is not going to be defeatured simply because it's lacking alt text. SandyGeorgia (Talk) 13:26, 9 March 2010 (UTC)[reply]
Alt text became a stated requirement in the Featured article criteria since July 2009. I don't see consensus for it being applied retrospectively however and agree with SandyGeorgia in that FARing articles for the alt text --guideline-- would be silly. Regards, SunCreator (talk) 19:12, 9 March 2010 (UTC)[reply]
Purpose of alt text I don't know that I have much perspective to add on this, but I would like to point out a few things that may be going overlooked:
Alt text is mandatory in HTML. There are bugs at bugzilla about HTML validity and any solution there would override any policy on Wikipedia to the extent that they will encourage/mandate/generate alt text on any Wikimedia project. It would be better to encourage users to come up with intelligible and useful alt text rather than generate dummy "*"s (as is sometimes done simply to pass validation, but obviously lacking any semantic purpose.) As an added bonus, by demanding alt text on Wikipedia, it may serve to popularize alt text elsewhere, making the Web more accessible for everyone.
Alt text is not just for the blind, but for devices that cannot render images (e.g. Lynx users.) Alt text is used to replace an image when it is not available, irrespective of why it is not available. If—God forbid—there is another loss of files (as has happened before) or if images get corrupted, alt text can be useful in replacing those destroyed files or finding new ones that compliment the article in the same manner.
Even if the sole purpose of alt text was for the blind, the argument "it doesn't help users born blind" is irrelevant, since it would (ostensibly) help those who were not born blind. The simple fact that it isn't a perfect and complete solution shouldn't stop us from implementing a partial and provisional solution.
Even if a majority of featured content lacks alt text, that is hardly a reason to remove it as a criterion. For those featured articles which currently lack it, the text can be added and if anyone nominates one of those articles for former featured article status, it can easily be added at that time. For that matter, if someone really wants to solve this problem, it can be done in a week with a handful of editors.
From the perspective of usability and proper HTML, if featured articles are to be the finest that Wikipedia has to offer, then it seems to me that alt text should be included as a criterion. For that matter, it is very simple to add, so it is not going to seriously derail any FA/FL/FT nomination. —Justin (koavf)❤T☮C☺M☯ 01:43, 10 March 2010 (UTC)[reply]
Addendum Wikipedia currently does generate dummy alt text for the purpose of passing validation. For instance, the current Main Pagepasses validation by including totally useless alt text like "Kathryn Bigelow" for a photograph of Kathryn Bigelow and "French Foreign Legion emblem" for that figure. These are identical to the title of the images and defeats the entire purpose of alt text. —Justin (koavf)❤T☮C☺M☯ 01:46, 10 March 2010 (UTC)[reply]
Back to the basic questions:
Is alt text mandatory for new FAs. In other words, will nominations be promoted without alt text.
If alt text is mandatory for new FAs, must it comply with WP:ALT. In other words, will nominations be promoted if their alt text does not comply with WP:ALT.
If an older FA is reassessed at FAR, is alt text mandatory for that FA. In other words, will the FA will be delisted if alt text is not added.
If an older FA is reassessed at FAR and alt text is added, must its alt text comply with WP:ALT. In other words, will FAs at FAR be delisted if their alt text does not comply with WP:ALT. --Philcha (talk) 03:49, 10 March 2010 (UTC)[reply]
It's difficult to discuss any aspect of ALT text without noticing the inherent problems with it. I know people mean well by wanting it but I fear it's just not a good way to spend our time. An example from today: I'm not allowed in ALT text to tell readers who don't have access to images that the image they're missing is of Big Ben and the Houses of Parliament. Instead, I have to talk about a clock tower next to a four-story building. In the ALT text guideline itself, we have an example of describing a woman as in her 40s when she looks considerably older, and there are similar examples everywhere of very subjective judgments being made. You could argue that these things don't matter, but if we believe ALT text is important, then they do. The whole thing is really very confusing and I fear not helpful to anyone, especially not at FAC where nominators already have a lot to think about. So to answer Phil's questions, it shouldn't in my view be mandatory for old FAs; the text shouldn't have to comply with WP:ALT; and in my view it shouldn't be mandatory for new FAs either. SlimVirginTALKcontribs05:58, 10 March 2010 (UTC)[reply]
It's a noble objective and I'll add it if asked but whether it improves the encyclopedia overall is open to debate. Adding alt text uses up time and levies an opportunity cost. Instead of writing alt text describing a picture, the editor could be writing the body of an article and may be inefficient use of time. It takes up article space which may be a problem for large articles. It could be seen as an additional unnecessary time consuming hoop to jump through that dissuades contribution of volunteers. As noble a goal as it is, it also smacks of political correctness foisted upon editors to compel certain behavior. It simply looks bad for volunteer editors to be judged by a criteria that is not intrinsic to the article writing process by those who can rectify the situation if its that important to them. Finally, it's not clear whether the practice is effective at achieving its aims. Too often the recommended alt text refers back to the text so is redundant. It is reasonable to ask why the caption shouldn't be sufficient? Alt text is currently recommended and encouraging it is great, but if it is cited as a requirement, that is instruction creep. Lambanog (talk) 08:04, 10 March 2010 (UTC)[reply]
Captions Captions supplement pictures, giving context to what you are seeing. Alt text takes the place of the picture in the event that it is not displayed or seen. —Justin (koavf)❤T☮C☺M☯ 08:23, 10 March 2010 (UTC)[reply]
Replying to several comments above, from different editors:
"The whole thing is really very confusing" It's a lot less confusing than it's been made out to be. Don't tie yourself into knots. Just use a simple heuristic: pretend you're describing the image over the telephone to someone. Write that down. Then stop. There's no need to make it perfect, any more than a listener would expect you to make it perfect over the phone.
"I'm not allowed in ALT text to tell readers who don't have access to images that the image they're missing is of Big Ben" No, if you were describing the image over the phone, you could expect the typical listener to know what Big Ben looks like, because it's an iconic image. So it'd be OK to say "Big Ben". As it happens, WP:ALT #Proper names mentions Big Ben as an example of an image whose name is OK to use in alt text, for exactly this reason.
"You could argue that these things don't matter, but if we believe ALT text is important, then they do." It's important to have alt text. But it's not important that it be perfect. All it has to do is convey the gist of the image.
"very subjective judgments" The case you mention is not subjective: that photo was published when the woman was 46, and clearly she was in her 40s. It's true that sometimes alt text can rely on subjective judgments. But that's OK: most of Wikipedia relies on subjective judgments. Just describe the image and move on. Again, there's no need to be perfect.
"I fear it's just not a good way to spend our time.""it's not clear whether the practice is effective at achieving its aims." We've had direct feedback from a blind Wikipedia editor that alt text is quite useful. And reliable sources say that alt text is important for visually impaired people: for example, a 2008 study of blind access to Wikipedia listed lack of alt text as the first obstacle. See Buzzi M, Leporini B. Is Wikipedia usable for the blind? In: Proc 2008 W4A. Beijing: 2008. (ACM International Conference Proceeding Series; vol. 317). p. 15–22. doi:10.1145/1368044.1368049.
"It's no use to readers who are blind from birth." Not true: alt text can be quite useful to readers who are blind from birth. Besides, only a small minority of blind people are blind from birth. Most blindness comes when people are older, due to age-related macular degeneration.
"It takes up article space which may be a problem for large articles" No, the overhead for large articles is quite small. For example, for Brad Pitt, a large article in FAC that I happen to have reviewed recently, only 1559 out of its 230,769 bytes of HTML are alt text: so the overhead is roughly 0.7%. And this is not counting the image bytes: the alt text for an image is typically a small fraction of the size of the image.
"Too often the recommended alt text refers back to the text so is redundant." Can you give examples of that? WP:ALT#Repetition says that alt text should not repeat article text. There are a few exceptions where placeholders are the best one can do (see WP:ALT#Placeholders), but they ought to be fairly rare, and are rare in the FACs and FARs that I've observed.
"why the caption shouldn't be sufficient?" Typically the caption is not sufficient by itself, because the image conveys useful information that's not in the caption; in this case the useful information in the image needs to be conveyed to the visually impaired reader. In the rare case where the caption is sufficient by itself, then neither alt text nor image is needed.
"it also smacks of political correctness" Sorry, but I'm afraid I don't understand that comment. I looked at Political correctness and the only mention of this topic is a criticism of 'Terms relating to disability, such as "visually challenged" in place of "blind"'. Is it the use of the term "visually impaired" that is objectionable? But this is a technical term (see Visual impairment), not a political one: "visually impaired" is not a euphemism for "blind". Another question: is it just alt text in particular, or WP:ACCESSIBILITY in general, that "smacks of political correctness"?
"Back to the basic questions" I've reviewed the alt text of every FAC and FAR for months. Most FACs now have alt text that is just fine, with no comments or work needed. Most FARs are about articles that predate the alt text requirement and therefore lack alt text, but the alt text is the least of these articles' problems. Any FACed or FARred article that has its other problems fixed has its alt text fixed almost as a matter of course, with little extra work needed for the alt text. I don't recall any of those "basic questions" actually coming up in practice: editors are pretty good about adding alt text, and it isn't that much work.
My big problem with the requirement of image alt-text to meet the FA criteria is the inherent subjectivity it introduces to the articles, when FAs are supposed to maintain a neutral POV. Alt text is intended to describe an image for someone who can't see it, whatever that reason may be. The problem is that seeing an image and describing an image are two very different things. The former is a sensory experience; you can perceive it as it is if you have the faculties. The latter requires communication of something that isn't words into words, a process that introduces a third party that distorts that information in a profound way. It's very OR. Note many of the examples offered by the alt text guidelines are very subjective interpretations. Furthermore they can be pretty useless; Eubulides says one description is suitably because "when the woman was 46, and clearly she was in her 40s", but it isn't because we still don't know what a woman in her 40s looks like, and furthermore there's no set definition of what a person in their 40s is supposed to look like because all people age differently. On the opposite spectrum, you can end up with bland text that is redundant to the caption. Also, I find the comment "alt text can be quite useful to readers who are blind from birth. Besides, only a small minority of blind people are blind from birth" ill-thought-out. If a person is blind from birth, what good is describing something they cannot perceive in a manner they have no reference for? Alss, you say it's useful to them, but then you dismiss them. Either the concerns of users blind-from-birth are worth addressing or they aren't worth addressing at all. WesleyDodds (talk) 03:07, 11 March 2010 (UTC)[reply]
Eubulides, I agree with Wesley about the subjective element. The other problem is that how we're told to do it keeps changing. I've had details removed from ALT text quite inexplicably e.g. a man on a horse walking down a certain street—I wasn't supposed to say what the street sign said. Sometimes it's okay to say Big Ben, but with other well-known things I've been told not to name them.
Anyway, I don't want to give you a hard time, because you do great work. When this last came up, it seemed to me there was a consensus not to mandate it at FAC. I was about to post a more formal poll about it, but I first of all looked at your contribs and saw what a great contributor you are, so I decided to let sleeping dogs lie. Every time I add ALT text to an image now, I am literally doing it only for you, and that's how I justify it to myself. :) But I do hope one day that you'll reconsider. SlimVirginTALKcontribs03:23, 11 March 2010 (UTC)[reply]
Thanks for that compliment, SlimVirgin. I appreciate your making it here; it's better than a barnstar! Your street-sign example doesn't sound right: if the image names the street that can be in the alt text. More generally, I'm sure that mistakes are made when writing or reviewing alt text, just as they're made in other parts of Wikipedia, but the presence of natural human errors doesn't mean we should abandon alt text. Eubulides (talk) 20:50, 12 March 2010 (UTC)[reply]
Subjectivity How is it inherently subjective to describe an image but not a boat or an author? For that matter, couldn't there be a featured article on (e.g.) the Mona Lisa? —Justin (koavf)❤T☮C☺M☯ 03:53, 11 March 2010 (UTC)[reply]
As it stands, Mona Lisa is a perfect example of subjectivity. The alt-text just says "see adjacent text"; the adjacent text is "The painting is a half-length portrait and depicts a woman whose facial expression is often described as enigmatic. The ambiguity of the sitter's expression, the monumentality of the half-figure composition, and the subtle modeling of forms and atmospheric illusionism were novel qualities that have contributed to the painting's continuing fascination." I've highlighted every part of that wording that's a clear violation of NPOV. – iridescent10:44, 11 March 2010 (UTC)[reply]
Great! After all this time, the enigma is solved: it's just another portrait of a woman with a facial expression. (Does the "iride" part of your signature violate NPOV too? Poor thing, you'll have to be reduced to a mere scent in the cause of policy enforcement) Yomanganitalk11:54, 11 March 2010 (UTC)[reply]
Thats a little unfair, Yomangani. That text is clearly a violation of WP:TONE, WP:NPOV unless it is attributed to someone, and as it is there could fall under WP:OR. (Then again that article is not featured either) I agree that alt texts should not be a requirement to achieving featured article status. Our encyclopedia should be open to all, but that does not mean every image, tiddle, and taddle, has to be made available. There is a logical limit to things. A blind person, regardless of the alt text, will never be able to appreciate an image unless they can actually view it. The same for users using lynx or browsers settings that block images. I know I am probably going to get yelled at for saying that, but lets be realistic here. (Or do you want to prove the political correctness charge above?) Does an alt text saying, "portrait of a man with glasses looking forward and smiling" or "a sail boat on the horizon with the sun setting" truly convey to a person the details of an image to allow them to appreciate the hues of colour, the attention to detail of a painter, or the sentiment and emotional feeling an image can instill? No. And in most cases a summation of the alt text is in the caption and is easily deducible when paired with the text of the article. Images shouldn't be be going into articles unless their subject is talked about in the article to begin with (WP:IMAGE), and the caption points what is already in the article and should not introduce anything new. Aside from that WP:ALT has clear problems and conflicts with WP:OR, among other policies that need to be reconciled. Wikipedia:WikiProject Spoken Wikipedia provides audio readings of each article, and that is something that could actually be of significant value to a handicapped reader, but we don't require that. Larger font would also be of use to handicapped readers, perhaps we should insist on a larger default font size on our featured articles? Some readers could be colour blind, perhaps we should insist on point out the colour of things? Some of our disabled readers may not have access to a Screen reader, perhaps we should find an open source version to provide to our readers? Seems to me, Alt text is quite low on the list of useful things we could do to make the wikipedia more accessible to handicapped persons, and to expect it to be included on a featured article, but not other more useful things, is fairly silly. —Charles Edward(Talk | Contribs)13:34, 11 March 2010 (UTC)[reply]
"A blind person, regardless of the alt text, will never be able to appreciate an image unless they can actually view it." This comment seems to be based on a misunderstanding of alt text's role. There is no requirement that alt text be an exact representation of its image. All that's needed is to give visually impaired readers the gist. The idea is to be much better than nothing, not to be as good as the image. (Please see WP:ALT#Brevity.)
"In most cases a summation of the alt text is in the caption" If that were true, we would not need alt text. But that's not alt text's role. On the contrary, WP:ALT#Repetition recommends against the caption summarizing the alt text.
"Wikipedia:WikiProject Spoken Wikipedia provides audio readings of each article, and that is something that could actually be of significant value to a handicapped reader, but we don't require that." First, that would be a lot more work. Second, it wouldn't address the problem, as these readings skip over the images.
"Larger font would also be of use to handicapped readers, perhaps we should insist on a larger default font size on our featured articles?" No, because readers with milder visual impairments can easily use larger font sizes themselves, or screen magnifiers. This is standard practice, and it means that Wikipedia editors need not worry about font size issues and WP:ACCESSIBILITY. If there were a similar technology for describing images aloud automatically, then we wouldn't need alt text at all. Unfortunately, no such technology exists.
"Some readers could be colour blind, perhaps we should insist on point out the colour of things?" We do have guidelines about how to make articles accessible to color-blind readers, yes; see WP:COLOR. When I notice a problem like that in an article I point it out and it's quickly corrected. It's typically less of a problem in practice.
"Seems to me, Alt text is quite low on the list of useful things we could do to make the wikipedia more accessible to handicapped persons," But as seen in the previous bullets, we're already doing the other things on that list, at least, the ones that would actually help. Also, we have a reliable source that lists alt text first on the list of things to do for visually impaired readers of Wikipedia (Buzzi & Leporini 2008, doi:10.1145/1368044.1368049).
I do see the violation of the content policies as quite significant. The NOR policy says that everything in an article must be attributable to a reliable published source, even if not attributed. That is, if someone requests a source for "this woman looks as though she's in her 40s," you have to supply one or remove the words. Material is also supposed to be neutral, and neutrally worded, or at least balanced. ALT text violates both these policies, and it necessarily violates NOR, because the descriptions are always those of a Wikipedian. To mandate something at FAC that necessarily violates a core content policy is a little odd. And when you're dealing with ALT text in contentious articles, it becomes a real problem, not just a theoretical one. SlimVirginTALKcontribs13:42, 11 March 2010 (UTC)[reply]
Not significant, only teething problems. The WP:ALT is just a bit messy and requires to be cleaned up. I would recommend anyone concerned enough to get engaged in discussion on WT:ALT. One solution to WP:NOR, which admittedly is like a sedgehammer to crack a nut, would be to simply put the caption text in the alt text - that ends WP:NOR and WP:NPOV issues completely, as it would be to the same established standard acceptable for caption text today. Regards, SunCreator (talk) 14:17, 11 March 2010 (UTC)[reply]
The caption text doesn't describe the image the way we're told ALT text must (a man looking to his left wearing a pink pullover etc). And even if they are only teething problems, the point remains that it violates the core content policies. We don't allow anything else to do that, teething or otherwise. SlimVirginTALKcontribs14:28, 11 March 2010 (UTC)[reply]
The core content policies take priority over the specifics or implied 'must' of alt text because it is a guideline. I am confident that editors will firstly attempt to keep to core content policies when adding alt text and those with a keen eye, such as yourself, will pick up on anything overlooked during the WP:FAC/WP:FAR processes or before. The alt text being required is still a relatively new thing and as such may take a while to mature. That doesn't mean we should reject or ignore it because it is not perfect, but instead work towards making it a good polished guideline. Regards, SunCreator (talk) 16:03, 11 March 2010 (UTC)[reply]
No, I'm afraid the ALT text has taken priority over the content policies. We now mandate original research in ALT text in featured articles, plus NPOV violations. The only reason this isn't an issue is that very few people, if any, read ALT text, but that's a good argument for not bothering to add it. On the other hand, if people do read it, and it's therefore worth adding, we ought not to be adding it because it's rarely if ever going to be policy-compliant. And so we have ourselves a bit of a conundrum. SlimVirginTALKcontribs16:08, 11 March 2010 (UTC)[reply]
I agree with SV. The alt text is not a guideline for FA. It is a requirement, because proponents made it very clear they wanted alt text included, the price for not doing so would be a failed nomination. A college student doesn't HAVE to do what a professor asks, he can easily drop out of school. It is a requirement. Whether it is a good one or a bad one is not for me to say. I had no notice of the WP:ALT discussions and so took no part in the process.--Wehwalt (talk) 17:09, 11 March 2010 (UTC)[reply]
Finally, a serious transparent debate about the absurdities of ALT being a requirement, and a thorough look into how its application as recommended by Eubulides violates several guidelines and policies. As one to have pointed out some of these problems months ago, I fully endorse the scrutiny: every time this issue was more timidly brought up, it got cut off with a sentimentalist lecture about how one user's POV and OR would actually benefit our visually challenged colleagues and readers. Dahn (talk) 17:37, 11 March 2010 (UTC)[reply]
I'm not following this WP:NOR/WP:NPOV violation claim at all. I accept the Mona Lisa body text fails as as substitute for Alt text. It has some problems as body text too so let's move on from that example. If one regards the image itself as a primary source, in what way would a faithful literal description of that image become OR? Is this any different from when one uses a book/film/video as the primary source for a plot summary? Or different from when I rephrase my source text to avoid plagiarism -- the way OR is being described here, I'd have to cite a thesaurus for every word change. And the alt-text = caption argument completely misses the point. Colin°Talk18:01, 11 March 2010 (UTC)[reply]
I've worked on a couple of FAs where the images and their interpretation were the key issues, yet I was required to add my own opinion regarding what those images showed. I avoided the issues by not describing the images in full. Example from Muhammad al-Durrah incident, a highly contentious article about a Palestinian boy who was shot and, it appeared, killed in front of a France 2 cameraman who filmed it. One of the images after his death shows him moving. Some say it is the boy's death throes. People who argue the incident was staged say he is peeking at the camera because he wasn't actually shot. (I had to delete this image because of fair-use problems):
[[:File:Muhammad al-Durrah final scene.JPG|right|250px|thumb|This scene from the France 2 footage, in which the boy appears to lift his hand away from his face, occurs after the material that was broadcast. Enderlin said he cut it because it showed the boy's death throes. Three French journalists who viewed it said it did not (see below).|alt=The same scene as above with the man and the boy, but more blurred. The man is crouching, leaning toward his left. The boy is lying across the man's knees. The boy's right arm is slightly raised.]]
I didn't get into how his leg was also raised, and wasn't raised in the scene before it, and how his hand had moved away from his face, and how he seemed to be looking straight ahead i.e. at the camera. That is, I avoided the OR and NPOV issues by not writing the ALT text properly. SlimVirginTALKcontribs18:24, 11 March 2010 (UTC)[reply]
Guessing at a person's age, commenting about their subjective expression, making statements about things in the image that are unverifiable are all problems. Take this as an example. "A sail boat on a sea with the sun setting at the horizon", seemingly straightforward. But what if it a ship at a distance and not a boat? What if it is a lake and not a sea? What if the sun is rising and not setting? Yet how could colorfully describe that setting without including those basic objects? And how could you know with certainty what those basic objects are without a source? If you boiled it down to "a boat on water with the sun in the background" past experience would tell me that it would be shot down as not descriptive enough in a FAC review. Any fanciful or subjective wording should not be included in alt texts as is now being required unless a source is available for it. —Charles Edward(Talk | Contribs)18:30, 11 March 2010 (UTC)[reply]
Here are the relevant sections of the policies. Our sourcing policy, WP:V, says: "This policy requires that a reliable source in the form of an inline citation be supplied for any material that is challenged or likely to be challenged, and for all quotations, or the material may be removed. This is strictly applied to all material in the mainspace—articles, lists, and sections of articles—without exception ..." (my bold).
Our WP:NOR policy says: "'All material added to articles on Wikipedia must be attributable to a reliable published source, even if not actually attributed in the text. This means that Wikipedia is not the place to publish your own opinions, experiences, arguments, or conclusions" (my bold).
Captions are subject to the NOR policy just as everything else is. But ALT text has no sourcing. It's intended to be a description of the image as seen by Wikipedians, which is OR by definition. SlimVirginTALKcontribs03:01, 13 March 2010 (UTC)[reply]
The image is the primary source. Citing bits of policy doesn't cut it, like lifing "contradictory" pieces out of religious texts. Please can someone tell me how alt-text (done properly) is different from the plot summary in Cartman Gets an Anal Probe, Casino Royale (2006 film), To Kill a Mockingbird or any other similar FA. All of these are sourced (presumably) to the primary text/film/cartoon. It is ridiculous to say I can describe a moving picture but not a stationary one. Colin°Talk19:50, 11 March 2010 (UTC)[reply]
Can't do other than to agree; it's an absurd inconsistency. The point of sourcing is that anyone who doubts what's being said can go check for themselves, by watching the movie or reading the book or whatever. It's not a goal in and of itself. --MalleusFatuorum19:55, 11 March 2010 (UTC)[reply]
How about this from the lead of Thomas R. Marshall: "Head and shoulders of a sixty year old man with a serious expression and many deep wrinkles in his skin. He has a bushy mustache and his hair is parted. He is wearing a high collared shirt and a neck tie." Is he 60, how can we tell, maybe he is 70? Is his expression serious, seems ambiguous? And this one in the body for the image of the statehouse: "A view looking down on a large building made of limestone. It is three stories high with two wings sweeping out from a central atrium with a domed stained glass roof", is the building really made of limestone how do we know? Are we sure it is three stories high, I count four. Are we sure it has a stained glass roof, looks like copper to me? In any case, none of this is referenced in the article and is not made clear by the photo, yet was insisted upon in the FA review, and passed after being entered. —Charles Edward(Talk | Contribs)20:05, 11 March 2010 (UTC)[reply]
Point taken, but are such articles featured? Considering an image to be a primary source in the way a book is is a bit of a stretch here anyway. The closest thing in that policy is "a work of art". I really don't feel that WP:PRIMARY has any application in to this issue. This is blatant Original Research using a non-literary source because what is often being said about an image is subjective. Paraphrasing a document is one thing - guessing about things in an image is entirely another. My primary reason for disagreeing the use of ALT texts is not policy though, and I wrote above, it is that we don't do other more useful things to make our articles handicapped accessible, so why is this required? —Charles Edward(Talk | Contribs)20:42, 11 March 2010 (UTC)[reply]
Yes, have you looked at the plot sections of any of our featured articles on TV, film or literature? There's no stretch wrt regarding the image as a primary source. Why should it matter if the source is embedded in our web page or linked at the bottom? Indeed, if we couldn't display the image (because of copyright), we could still describe it in our body text could we not? We describe moving images (films). I agree that some folk here have found difficult image alt-text examples but I could list loads of difficult articles that are jolly hard to do well while following policy. That doesn't stop us trying and surprising ourselves. As for the "other more useful things" argument, that is a logical fallacy. Colin°Talk21:04, 11 March 2010 (UTC)[reply]
I must also agree with Colin here. There is a fundamental misunderstanding of what constitutes "original research", a misunderstanding whose logic, consistently applied, would prohibit us from articulating any expression at all aside from direct quotations and mathematical/scientific formulae.—DCGeist (talk) 20:00, 11 March 2010 (UTC)[reply]
okay, so we can describe what is in a source, what an expert says about xyz, but we should not describe what is in a picture that is in front of us? While, yes, it is a subjective description, there is also an element of subjectivity in the narrative of what various people have written about xyz subject. While I find alt text a drag to write, isn't there something more important to do than to write pages about why we don't want to do it? Auntieruth55 (talk) 20:04, 11 March 2010 (UTC) <---off to review an article[reply]
The point here, as far I can see, is that the ALT approach is not sustainable, because it is entirely evolving around users' subjectivity - even in the most benign cases, subjectivity is there, which is not the case with all other analogies made above. A plot summary that is subjective can be rephrased to simply say what the plot is about, without guessing; there are works of literary criticism out there that one can use to back further assertions with quotes, there is the book itself to quote verbatim in cases where the event described is subject to interpretation. And, once this is done properly, the resulting plot summary will be the equivalent of a caption, not the equivalent of an alt text. Both caption and summary will say what the image/text has for its subject, not "how" that subject is represented. And even in those cases where it would say how the subject is represented, this can be backed with quotes from secondary sources, preferably attributed.
The case with alt text is that it adds an entire extra layer of subjectivity, that we otherwise strive to avoid by keeping things within reasonable limits (why we cite sources, why we don't do editorial opinion, and so on). In most cases, it does so for absolutely no reason. To quote one example above: simply introducing an image with the caption "SS Whatever in Whichever harbor" does the job of describing the subject; insisting that we *need* to also say, in alt, "the sun is setting, the waves are splashing, the captain is on deck smoking his pipe" - that adds an entire layer of guesses. Guesses of no encyclopedic value value (us seers, we don't understand "SS Whatever" because there are waves splashing), parasitical to the caption, and evidently contrary to explicit wikipedia policies.
This is not in any way comparable to summarizing a source and what it says about x subject - doing that is immediately useful for all readers who approach the subject, and the fact that a description/judgment is attributable to a reliable source reduces the amount of subjective detail from infinite to a manageable number. It is probably always gonna be there in some amount, but other cases don't open the door and say: "Come in. You may be of no actual use to anyone, you may prove yourself impossible to correct, but we will make you a norm."
In the long run, and if left alone, the alt text "solution", at least as it is formulated now, would prove itself unfeasible and undesirable for anyone, even for those who recommend it. The very reason why we "write pages about why we don't want to do it" is that this was imposed on all of us as the definition of quality, when in fact it is quite obviously not, and as a requirement, when it most often solves no actual problem of accessibility (it actually describes things of no actual importance to anyone). We could perhaps leave this thing take its actual toll on the project by letting it prove its obvious contradictions to almost everyone, which it will do sooner or later, but my guess it wouldn't want to let it run amok because it we care about wikipedia. I for one care about tyrannical, impractical guidelines being fabricated, and about editors being supposed to relinquish their understanding of wikipedia guidelines and their common sense for he sake of some purely sentimentalist claim. Dahn (talk) 07:28, 12 March 2010 (UTC)[reply]
For crying out loud folks. TLDR. It is quite apparent that few here have actually looked at any of our plot summaries in all our FAs. They aren't sourced to published literary criticism; they are nearly 100% sourced to the primary material being described. And saying "done properly, the resulting plot summary will be the equivalent of a caption" just shows a gross misunderstanding of the purpose of each. The caption is "Casino Royale", or whatever. I'll repeat: alt-text == plot summary where one deals with a static image and one with a moving one. Exactly the same policy-based flaws can be found in practically every sentence of a plot summary if one is anal enough to apply the arguments used here. I'm not going to bore you all with an example cause I'm sure you are clever enough to do it yourselves. Please, go read some of our FA plot summaries, wait for the little bell to go ding, and stop trying to find excuses to do something tedious but completely harmless to this project and its values. Colin°Talk08:39, 12 March 2010 (UTC)[reply]
Colin, let me summarize it for you: the info in alt text is in most cases irrelevant to any user, blind or non-blind, and does not provide the reader with any objective info (as opposed to leaving some room for objectivity); the analogy with plot summaries for either texts or films is deeply flawed and misleading. Why "deeply flawed", why "misleading"? The answer is in the TLDR part, and I'm not going to bore you with it. Dahn (talk) 08:53, 12 March 2010 (UTC)[reply]
No, the answer wasn't there, just confusion. I can't understand how you say alt text "does not provide the reader with any objective info". I'm not following that at all. I'm not understanding how because the image moves, we can describe it but once it stays still any attempt at description undermines Wikipedia. The hyperbole used here makes me think someone has just suggested we drop WP:V because all those citations are a bit bothersome. Please find me an FA plot summary that you feel would pass the scrutiny being applied here to alt text. Colin°Talk09:21, 12 March 2010 (UTC)[reply]
To be honest, what do you suggest I do now? if I answer in detail it's TLDR, if I don't you'll keep repeating stuff I already answered to. For "moving images": false analogy. For one, nobody here is claiming that describing images is impossible: for pictures, we do that with captions, that do a good job of describing the what and where, and maybe the how, in a manner that has a long tradition. The goal there is that, provided the caption is properly written, anybody with a modicum of intelligence will in the end understand what the picture is of, not what impression that thing makes on me and you. Without going into the various irregularities that some plot summaries may still feature, even at an FA level, I picked out at random The Beginning of the End (Lost) (never watched it, btw). Do you see, in the plot summary, any subjective adjective other than "twitchy" (which one could consider removing, as it adds nothing). I see assessments of symbolism, narrative intentions, critics' speculations et al. in the sections after the plot section, where they are clearly sourced and attributed. What I do see is in the plot section is the equivalent of a caption adapted to the full length of a TV episode.
Sure, I can some use of alt text in cases such as List of non-ecclesiastical and non-residential works by John Douglas (as much as I still don't understand what those spatial relations may mean to those users who will still never see them), borderline for List of counties in Florida and such (where the alt text for the leading picture makes observations that border on the inane), I can clearly see it for a diagram or some images comparing text attributes or such. But for the love of me: Lebaudy République (where caption would have served to note that this is the airship in question and affixed, we have an absurd alt text telling us about the orange stamp and some faceless people in the general area; where we already read in the caption "The République's gondola and keel lie in a field at Jussy-le-Chaudrier, 3 September 1909" - the relevant info, we get alt telling us "On the right, an airship's gondoloa and keel lie on the ground on the side; on the left several people stand looking toward the camera." - the one informative bit is merely repeated), St. Peter's Basilica ("A very detailed engraved image of a vast interior. The high roof is arched. The walls and piers which support the roof are richly decorated with moulded cherubim and other sculpture interspersed with floral motifs. Many people are walking in the church. They look tiny compared to the building" - priceless; where caption gives us "St. Peter's Basilica from the River Tiber. The iconic dome dominates the skyline of Rome", which is already saying too much and veering into the inane, we get alt: "A view of Rome on a sunny afternoon looking along the river. A bridge crosses the river and beyond it is a hill on which the grey dome of St Peter's rises above ancient buildings and dark pine trees" - none of which is relevant to the picture, almost all of which is subject to interpretation). And so on. Alt text appears to have been designed to give blind people the "feel" of the picture, based on some optimistic (and ultimately condescending) belief that this can be rendered in universal terms and helps the people in question. The result is people having found a way to evade the rigid requirements of wikipedia by creating some loophole for writing down their essays. Nothing in that is comparable to a simple depiction of what happens in a text or a film, whose very purpose is most often explicit (communicable to a mass audience) and, quite often, its own summary.
(Incidentally, the futility of such exercises is made even more obvious by the fact that editors assume blind people would be helped by a stream of their impressions, but also that they would regard words such as "keel", "gondola", "cherubim", "floral motifs", "pine tree" as irreducible concepts, that when mentioned would appear in the mind of any user...) Dahn (talk) 10:07, 12 March 2010 (UTC)[reply]
If you are going to insist an image is an acceptable primary source, then you do need to apply the primary source policy to it, which many alt texts are a clear violation of, as has been pointed out here multiple times. To quote WP:OR
"Reliable primary sources may be used in Wikipedia, but only with care, because it is easy to misuse them. Any interpretation of primary source material requires a reliable secondary source for that interpretation. A primary source can be used only to make descriptive statements that can be verified by any educated person without specialist knowledge. For example, an article about a novel may cite passages to describe the plot, but any interpretation needs a secondary source. Do not make analytic, synthetic, interpretive, explanatory, or evaluative claims about material found in a primary source. Do not base articles entirely on primary sources. Do not add unsourced material from your personal experience, as that would make Wikipedia a primary source of that material."
You CANNOTguess at someone's age, you cannot guess at the meaning of their expression, you cannot guess why an action is occurring, you cannot guess at the material or makeup of items in the image in any way. Unless it is plainly obvious, you can't say it. You can't say "Fifty year old man in tweed suit", you can only say "Man in suit". And when you have cut it down to the basic level that is acceptable by policy, it is nearly worthless for the purpose it is intended - to convey a mood and feeling of an image to the extent that someone who cannot view the image can appreciate it. I would also point out, it is fine to use a primary source to establish a plot, as mentioned above. But if you are using a primary source to make subjective statements about the plot, then you are violating policy. It is ok to use an image to write a very basic description of an image, but the extent to which we are going in putting in details is over the line of what is acceptable and conflicts with other policies. WP:ALT is not unworkable, but it is troublesome in its current form and should not be a requirement here until it is workable. I hate to beat a dead horse, but one last time: Unless what the alt text says is 100% obvious and verifiable from the image itself (or a reliable source) it needs to not be there. No guessing. —Charles Edward(Talk | Contribs)13:47, 12 March 2010 (UTC)[reply]
We don't describe images with captions, we name them with captions. The caption for a pop star photo tells you nothing about what they look like. The caption for a church tells you what it is called, not what it looks like.
As for the example plot, it might have helped to pick one we had both seen as I'm at a disadvantage compared to the alt-text criticisms. From what little I know about the series and the description here, I'm guessing there are some surreal images with time and place illusions that make it hard to trust your eyes or the sequence of events. The text "but finds that the gun is not loaded because Locke had no intention of killing Jack earlier that day" is I suspect the writer putting 2+2 together which is WP:OR but pretty harmless. The text "is keeping quiet about his time there" probably makes an assumption based on limited data. The text "he lies that he has no knowledge of Ana Lucia" assumes the falsehood (if one can be sure it is false) is a deliberate and malicious act: perhaps he's just forgetful. The text "Jack, who is thinking of growing a beard" probably assumes that because Jack has said he is thinking of growing a beard that he actually is -- how can we know is inner thoughts? The text "which shows us that these flashforwards occur before Jack's flashforwards" makes my head spin but I'm guessing the actual story presentation isn't straightforward, meaning the the plot summary can only be made by guessing/assuming and the show's writers might be playing with us a bit.
Re: lengthly criticism of more alt-text examples, please stop. It isn't helping the discussion. I could do the same for article body text but what would that tell us. Nobody is denying there are pathalogical examples but that shouldn't stop us. Colin°Talk13:43, 12 March 2010 (UTC)[reply]
On the pop star issue: and? How is it helpful that you now know what you're presumably never going to see, in terms another person finds important? So that you what? What is the logical situation in which this becomes relevant information for a blind person? Just so that the blind person would not feel excluded? Then we might as well start describing music for the deaf.
With the little knowledge I have of that series (I don't recall having watched more than 20 min. of a random episode), it appears to me that the statements you place in quotes refer to info which is made explicit in other episodes - it doesn't rely on guesses, just uses info that is explicitly dealt with elsewhere. It is perhaps one more reason why not to have articles on each episode of a series, but mutatis mutandis it's still the same basic idea, and the analogy with alt is still false.
The pathological examples are out there, in FAs, and are perfectly compliant with the pathological examples (your wording) presented as good practice in the alt guideline, top to bottom. They are absurd because the guideline is absurd. One either cuts down the alt to the basic useful info that is/should be already in the caption, and uses alt only in cases where it makes sense (where the blind user will miss out on text or numerical values within the picture, or anywhere else it can be reasonably discerned that he is deprived of concrete info explicitly transmitted by the image); or continues and ends up with this nonsense ("there are people on the ground", "the woman is in her forties", "that baby is way pretty"). And the line between useful and idiotic in this case, for any scenario, is so thin, that it's simply not reasonable to even seriously contemplate this being an applicable guideline or policy. If someone wants to apply it and thinks s/he can apply it, let them at least try on their own and not impose this on the rest of wikipedians who realize the implications of going down that path. Dahn (talk) 15:44, 12 March 2010 (UTC)[reply]
I'm puzzled at Auntieruth55's reasoning in "While I find alt text a drag to write, isn't there something more important to do than to write pages about why we don't want to do it?" Perhaps there's something more important to do than to write alt text in to FAC nominators. And perhaps alt text should not be a FAC requirement. --Philcha (talk) 08:43, 12 March 2010 (UTC)[reply]
Just pointing this out as well. The lead of WP:ALT states: "The alt text should be concise and should emphasize the image's most important visual aspects: it should summarize the essence of the image rather than describing every detail." We are expecting an editor to make Original Research and using their own opinion to decide what the most important visual aspect it. We are also asking them to capture the "essence" of the image, which is also subjective. Essence also indicates that it is expected that the significance of the image be conveyed. Without a reliable secondary source, all of that is a total violation of WP:OR. You cannot make any of those calls from primary sources alone. This is the essence of the problem with the guideline. —Charles Edward(Talk | Contribs)13:57, 12 March 2010 (UTC)[reply]
Plot summary. Plot summary. Plot summary. To rephrase what you've just said but for plot summaries (which we allow, sourced 100% to the primary material): We are asking editors to make original research and use their own opinion to decide what the most important plot aspects to cover. We are asking them to capture the "essence" of the story in a fraction of the words/time that the original took, which is subjective. Essence also indicates that it is expected that the significance of the plot be conveyed [I disagree with this wrt plot and image]. Without a reliable secondary source, all of that is a total violation of WP:OR. You cannot make any of those calls from primary sources alone. This is the essence of the problem with plot summaries.
When you guys succeed in removing alt-text from our FA criteria, and no doubt demoting the guideline to an essay, I hope you can hold your heads up if the press write a "Wikipedia sticks two fingers up at the blind" article [yes I know the irony of that headline]. I'd welcome someone just came out straight with a "I can't be arsed helping disabled readers" than all this silly wikilawyering over something quite harmless. Shameful. </unwatching> Colin°Talk14:23, 12 March 2010 (UTC)[reply]
"Something must be done to help our disabled readers. This is something, therefore it must be done" just doesn't wash I'm afraid. It has never been made clear who precisely the alt text is meant to target, just for starters. It's a half-baked idea forced through half-cocked. --MalleusFatuorum14:47, 12 March 2010 (UTC)[reply]
Colin, Its not about screwing the blind for me. Its about the reliability and accuracy of what we are presenting our readers. Isn't a blind person equally harmed by unverifiable information? To me, it looks like we are actually hurting the blind by giving them substandard material that wouldn't hold up the scrutiny we employ elsewhere. Its like saying, because you are blind we are not going to not give you the benefit of the enforcement of WP:V on images descriptions. I personally don't want to get rid of WP:ALT, we just need to fix what is broken with it - which is encouraging editors to make guesses about the contents of an image in order to fluff up its alt text. And while WP:ALT remains broken and at odds with our core policies, it should not be an FAC requirement. —Charles Edward(Talk | Contribs)14:51, 12 March 2010 (UTC)[reply]
I agree. I have never seen why the requirement targetted FAs anyway. If alt text is such a good thing then it should be applied systematically across all of wikipedia's article, not just FAs, and there should be an option to provide default alt text for all images. --MalleusFatuorum16:30, 12 March 2010 (UTC)[reply]
As someone whose dayjob is working with people with disabilities, I am very disappointed and discouraged by parts of this conversation. The internet has the potential to level the playing field for so many people with disabilities, including access to information, communication, leisure, work etc etc. It already does so in many ways. Unfortunately, many web designers still do not give much thought for accessibility issues. Silly but very concrete example: the full range of Skype functions cannot be accessed without using a mouse, which plenty of people can't use, or cannot use easily, for one reason or another.
Alt text for images is mentioned over and over again by experts as a key feature of providing web accessibility for people with disabilities.[6][7][8][9][10] If WP is supposed to be an encyclopedia anybody can edit and to provide free access to the sum of human knowledge, then surely it should also be a website we all can access, read and benefit from. Yes, it is annoying and difficult at times to add alt text; yes, alt text may include a degree of inevitable subjectivity at times; but yes, our best articles, which is what Featured Articles strive to be, should be our best, and that includes conforming to the best practices for not only homegrown concepts such as sourcing, neutral pov, references etc etc, but also the best practices for making this encyclopedia (and the web in general) truly open to everyone. Just as Malleus suggests, the goal should be that all images in WP should have alt text, just as all images everywhere on the web should have them. Since FAs have long provided models for what should be done with WP articles in terms of excellence, this FA requirement for alt-text is appropriate in demonstrating what in fact all should be doing, all the time. --Slp1 (talk) 17:20, 12 March 2010 (UTC)[reply]
What I'm noticing in this link you've provided is the recommendation that "alt text should be five words, e.g. dog leaps for a stick". This is certainly not what is being insisted on at the moment (first suggested example at WP:ALT: A Georgian-style terrace house with four floors and an attic. It is red brick, with a slate roof, and the ground floor rendered in imitation of stone and painted white. Each upper floor has four sash windows, divided into small panes. The door, with a canopy over it, occupies the place of the second window from the left on the ground floor.) – iridescent17:28, 12 March 2010 (UTC)[reply]
I'm not knowledgeable about the right length etc of alt-text, though that could and obviously should be discussed, likely elsewhere. (I'll just quickly note however, that the recommendations of the RNIB, the source cited in [11] may have changed from the "5" words thing... this seems to be there latest page, where it talks about "succinct" [12]). I'm more interested in the principle that our best articles should show off web-design best practices requiring alt-text.--Slp1 (talk) 17:43, 12 March 2010 (UTC)[reply]
What's very clear from looking at the RNIB guidelibes for alt text is that the way it's being done right now (the pretend you're describing the image to someone via the telephone approach) is completely misguided, and anyone who disagrees is subjected to moral blackmail of the sort we've just seen above. So our best articles are ending up poor examples of how to make wikipedia more accessible, not good examples to be followed. BTW, the RNIB guidelines talk about a maximum of a short sentence, beginning with a keyword. --MalleusFatuorum17:49, 12 March 2010 (UTC)[reply]
Yes, thanks I now see the info about "a short sentence", though I also note further up that "Complex images" whose "full meaning cannot be adequately described in a short phrase or sentence" may require longer descriptions, handled in a somewhat different way technically. I'm not sure how WP software etc would handle the other options suggested. I would agree that our goal should be for the actual alt text used to be of high quality; though as with the articles themselves, we may not get it right first time. The point is to have the goal. --Slp1 (talk) 18:09, 12 March 2010 (UTC)[reply]
At the moment, according to the "standards" set by Wikipedia:Alternative text for images, the following is a good example of alt: "Head and shoulders of a serious and dignified woman in her forties, with dark hair up and in a dress with high lace collar and a cameo at her throat, Edwardian style". So is: "Three-quarter oval portrait of a slender woman aged about 30, garbed in black. Her deep-set eyes gaze solemnly over the viewer's shoulder. Her dark, straight hair is parted in the centre without a fringe, combed over the ears, and pulled back in a low bun." If this is the type of info the alt is supposed to covey, we might as well scratch out all our content policies - someone knows better. (Interestingly, I distinctly recall having pointed out the problems with these standards on the associated talk page many months ago, but my comments either ignored or answered with a circular argument about how it's "useful" and how I wouldn't like being blind myself, etc.). Almost nothing in this info is useful to anyone, and cannot possibly be said to impart knowledge - unless it is knowledge of how some editors still set their watches to Romantic era standards. The rest is or should already be covered by the caption. So, no, the point is not to have the goal: in the presence of a caption, the goal is absurd for most conceivable cases. Dahn (talk) 18:22, 12 March 2010 (UTC)[reply]
It is perfectly reasonable, in my view, to discuss what constitutes appropriate alt-text (length, detail etc): consulting with knowledgeable sources such as the RNIB document will help guide us here, and I am quite prepared to believe that some changes should be made to the guidelines/examples. However, determining that having alt-text for images as a goal is "absurd", flies in the face of all the available evidence from those who are in the know about web content accessibility. In both cases, as with researching articles here, we should be deferring to the experts who spend their time studying, working, and writing in this area. --Slp1 (talk) 18:45, 12 March 2010 (UTC)[reply]
Maybe I'm not being clear about this, as it must be fifth time I raised on this page: the option wikipedia already gives, to all its readers, is a caption, and that most often covers the purposes of alt, if alt is to be kept reasonable ("dog leaps for a stick"). The idea of alt is not absurd, but the application of alt as a default in cases where it adds nothing, and where the caption already summarizes all info that is not inane, is quite clearly absurd. I don't have any dispute with "people in the know". I have a dispute with people who misquote them in a situation where such concerns are most likely already addressed by wikipedia's format and proper editing within the limits set by that format. In an overwhelming majority of cases, it's a situation of "if you write a proper caption, you'd already be describing the image as much as it can and should be described". In those cases, adding alt is merely someone's opportunity to rant. Dahn (talk) 19:28, 12 March 2010 (UTC)[reply]
Dahn, the reason nobody is picking up on your "let use captions instead of alt text" idea is because it is crap. Go read WP:ALT and stop wasting our time. Colin°Talk19:37, 12 March 2010 (UTC)[reply]
Things to do when running out of arguments: ridicule your opponent, marginalize him by claiming you stand for a community and he's all alone, and grossly misrepresent his position. Rest assured, since you first promised you were not going to follow up on this discussion, I was not counting on your attention or your time to spare. And, in fact, I could have saved some of my time by not reading your comments at all, but for some reason you keep posting them as replies to mine...
For all those users not motivated to misquote me: my argument is not that we should "replace" alt with captions; my argument is that we already have caption which state what the image depicts in the one reasonable way this can be done, a way which is perfectly compatible with what specialists say alt text should be (the "dog leaps for a stick" example). Comparing images with such captions with random images on the net, which as a rule give no verbal clue to blind people as to what they represent, is comparing apples and oranges. Dahn (talk) 19:46, 12 March 2010 (UTC)[reply]
One can't help but wonder why this available evidence wasn't taken into consideration before this half-baked idea was foisted onto the FA criteria. The situation now is that we have a significant number of FAs with alt text that is to all intents and purposes largely useless, does not meet the recommendations of expert bodies like the RNIB, and a few zealots pushing more of the same. A bit of a joke really. --MalleusFatuorum18:52, 12 March 2010 (UTC)[reply]
Well, to be fair, the WP:ALT guideline makes extensive reference to web-accessibility documents from the World Wide Web Consortium, written by big names in the field and widely reffed by others writing on the issues. --Slp1 (talk) 19:08, 12 March 2010 (UTC)[reply]
(peeking) Wrt the moral blackmail issue, you will find my disgust is aimed at those who had no intention of helping the blind, wanted to remove the requirement for alt-text from our best articles, and perversely applied their intellect to finding ridiculous supposedly-policy-based excuses for not doing something tedious. Now that Slp1 has steered the discussion towards "how can we improve WP:ALT" might I repeat my initial request that this discussion be moved to WT:ALT where it belongs. The "bit of a joke", is that this is now twice that the FAC talk pages have been filled up with a timesink of a discussion when that energy could have been applied to WP:ALT months ago. Alt-text is a web-accessibility-requirement we have shamefully ignored. Let's find ways of adopting it rather than writing nonsense about OR. Colin°Talk19:22, 12 March 2010 (UTC)[reply]
I haven't seen anyone display a reluctance to help the blind, quite the reverse. I recall wondering why the move to provide alt text wasn't more generally targetted rather than just at FA. The discussion as to whether alt text shoould be part of the FA criteria is quite rightly here. The discussion about what alt text ought to look like obviously ought to be elsewhere, but what's equally obvious is that the alt text now being produced is next to useless and likely helping nobody. --MalleusFatuorum21:39, 12 March 2010 (UTC)[reply]
Criterion two is "It follows the style guidelines, including the provision of—". The "including the provision of—" reads to me as mere elaboration, not a limitation to the provisions thereafter. WP:ALT is a style guideline. Weren't FAs, then, supposed to follow it even before the requirement was "hard coded" into criterion three? I don't particularly care for WP:ALT, but do the criteria, as worded, not bind us to it by virtue of the community's decision to make it a MOS guideline? Is this then a conversation more appropriate for WP:ALT's talk page? Эlcobbolatalk17:33, 12 March 2010 (UTC)[reply]
Everything aside from straight-up quotation is OR in some way or other, because it requires interpretation. Reorganization of material, selection of details, and paraphrasing involves making choices and interpreting meaning. Different editors will make different choices. The question is how much reorganization, selection, and paraphrase you think constitutes a violation of Wikipedia's policy or spirit. There is a spectrum - where are you on it? I met someone the other day who thought every article on Wikipedia was OR because of these very principles - perhaps we can agree that none of us fall there? Awadewit (talk) 16:58, 12 March 2010 (UTC)[reply]
Jeez, I hope not. It's like that guy in college who says everyone's a racist. You, you, and you. None of us can ever get away from it. --Moni3 (talk) 17:11, 12 March 2010 (UTC)[reply]
In addition to Awadewit's comment on WP:OR, WP:NPOV runs into "Reorganization of material, selection of details, and paraphrasing involves making choices and interpreting meaning". WP:OR and WP:NPOV are ideals which we can't fully attain, but must always aspire to. --Philcha (talk) 20:41, 12 March 2010 (UTC)[reply]
I agree with the above. Also:
The idea that alt text "is blatant Original Research using a non-literary source because what is often being said about an image is subjective." may sound like a serious problem in theory, but in practice it's rarely an issue. When one is summarizing a text source in text, one does the best job one can; if another editor spots a problem with the summary and objects to it, then the two editors can work it out using standard dispute resolution mechanisms. Resolving any such dispute requires editorial judgment that is often inherently subjective. The same is true for summarizing an image in alt text. Disputes about alt text are quite rare in my experience, but I've had them, and their subjective elements are easily resolved in a similar way, by discussing things on the talk page (it's never gone further than that).
More generally, I'd like to say again that alt text shouldn't be that hard. Please don't spend a lot of time on any particular piece of alt text, worrying about subjectivity and semiotics and whatnot. Just write something down in a minute or two and then move on. Obviously there are exceptional cases (and SlimVirgin's example of Muhammad al-Durrah incident is a good one of an image where the visual description itself is hotly disputed), but these are exceedingly rare in practice.
I only disagree slightly with that. I agree that problems can be easily fixed on sight, and to a degree everything we do here is in part due to our subjective opinions, but we keep that in check by using reliable sources. I think we really do flirt the line of original research quite a bit with these alt texts. There are lots of acceptable examples out there of how it could be done , but there are lots of unacceptable uses in practice. WP:ALT, the guideline itself, encourages editors to go over the line of what is acceptable. Take the example of the "Gregiorian house", WP:PRIMARY says that you can't make an interpretation that a non-specialist could not determine. If all I have is an image, and I am a non-specialist, I don't know if that house is Gregorian, or Tudor, or Federal, or Edwardian, or Victorian. And the line between some of those is so fine that only a specialist would know. (Look at the example of pharmacophore, there is no way a non-expert would have any idea what that image is about; where's the citation?) In both of those cases, that level of specificity is what the caption is for, where it can also be attributed. Making a specific statement like those two examples without a secondary source is a problem.
The most common problem I see in practice is photos of people where we are guessing at their age, using the image and a secondary source to determine about when a photo was taken, which is the very definition of WP:SYNTH - using two sources to arrive at a conclusion not explicitly made by either source. Look at the more complex examples on the ALT page, like the map of moving Canada, it is citing dates and events that would require in line citations elsewhere. Guess to boil my beef down, I think we are being too specific with the alt texts rather than focusing on the real basic elements of the images. I personally use alt text in my articles now, and will continue to do so whether or not this requirement stays or goes. I am just not going to do so with the specificity called for at WP:ALT unless I am actually sure what I am writing is covered by my source. That is the sum for me. I will let the rest of you figure out if you are going to keep it as a requirement; it really has no impact on me one way or the other. I have just been giving my honest opinion :) The ALT guidelines needs more than just a disclaimer about other policies, it needs to be completely redone to discourage the frivolity that is now encouraging. —Charles Edward(Talk | Contribs)21:33, 12 March 2010 (UTC)[reply]
Thanks for that comment. I adjusted the alt text for two of the examples along the lines suggested. The pharmacophore example is a trickier matter, though, as WP:ALT#Chemistry notes; but we probably should be taking this part of the discussion off to WT:ALT. Eubulides (talk) 21:49, 12 March 2010 (UTC)[reply]
Serious concerns have been raised above about the quality of alt text (OR and so on), and the only FA check seems to be that it's present. To me that's analogous to saying that the article meets 1a, because prose is present. I have so far opposed two FACs because of the mini-essays masquerading as useful alt text, and I'm perfectly prepared to go on doing that unless some sanity prevails here. The present WP:ALT guidelines are misguided, and until they're sorted out I propose that the requirement for alt text, at least in the essay form that seems to have taken root, is dropped from the FA criteria. --MalleusFatuorum23:57, 14 March 2010 (UTC)[reply]
Well go over to WT:ALT and work through the issues there. I created a talk-page-topic to discuss the length issue three days ago and nobody but Eubulides and Slp1 have responded. If editors here feel strongly that WP:ALT at present is "misguided" then propose that it be demoted from guideline status, advertise the proposal at a central location, and work out a consensus. Wikipedia-wide policies and guidelines have no business being discussed in detail here. The RNIB guidelines being currently cited at FAC are probably not 100% appropriate to an encyclopeadia were images tend to be worth describing as opposed to the stock photo and graphic design examples the RNIB used. We currently have the silly situation where external guidelines are being cited in over internal ones. Let's fix the internal guidelines. Colin°Talk13:13, 15 March 2010 (UTC)[reply]
This is an FA issue because the current alt guidelines have been added to the FA criteria, so it is perfectly proper to discuss it here. I see no evidence that those supporting the present guideline understand the issues, or what needs to be done about them, so little point in further discussion. Further, who is checking the quality of alt text? The only check I see is that it's present, which is far from the checks done on every other aspect of an FA nomination, which is also a subject that needs to be aired here. --MalleusFatuorum13:23, 15 March 2010 (UTC)[reply]
I do check the quality of alt text, and I believe Eubulides also does. I am uncertain how you could have missed that, given that I have commented on the quality of the alt text of a number of current FACs. Unfortunately for you, my ideas about what constitutes good alt text are apparently not the same as yours. Ucucha13:38, 15 March 2010 (UTC)[reply]
Regardless of whether the FA criteria explicitly mentions WP:ALT, it is still a content guideline that affects all articles on WP, and by implication affects WP's best articles. Your argument could extend to discussing the finer points of WP:OR and WP:V because policy-requirements are part of the criteria. Perhaps we should start a debate on logical quotation marks here? I see no evidence that the lurkers at WT:FAC understand the issues either, but they are jolly good at moaning. There are other people on WP than those who have WP:FAC in their watchlists. Some of them might actually have something useful to add. This is the wrong forum. Colin°Talk13:51, 15 March 2010 (UTC)[reply]
The addition of a half-baked requirement to the FA criteria is a specific FA issue. I'm not aware that the same problem exists with WP:OR or WP:V, unless you're suggesting they're half-baked as well. --MalleusFatuorum13:55, 15 March 2010 (UTC)[reply]
The problem is not that a guideline was explicitly added to the FA criteria. Your problem is that you don't think it should be a guideline. Or do you think it is OK for the plebs to have to put up with it? Colin°Talk14:06, 15 March 2010 (UTC)[reply]
I'm very much in favour of a guideline as it happens, just very much against the current one. You make a fair point about "the plebs" though; my view is obviously that nobody should have to put up with it. The issue at FA of course is that perhaps for the first time another editor actually takes the trouble to look at the alt text, which is why I suppose the issue has come to a head here. Getting anything changed here at wikipedia is a task that's way beyond my limited stores of patience though, so I guess I'll just have to keep moaning. --MalleusFatuorum14:16, 15 March 2010 (UTC)[reply]
Generally I agree with MF's critisms of WP:ALT and the imposition of WP:ALT on WP:WIAFA. However I feel that the alt text at Wikipedia:Featured article candidates/Turf Moor/archive1 is OKish. The real problem at this article is that the pics are depressingly monotonous, so the same applies to the alt - half the pics are a waste of space.
They're definitely people wearing fluorescent jackets. The one on the extreme right is standing up, with his back to the camera. The "writing" is your imagination. :-) --MalleusFatuorum17:23, 15 March 2010 (UTC)[reply]
The problem with the pictures, in addition to being monotonous, is that they are of a subject that is not easily photographed. To show the stands, the picture has to be taken at a distance, and detail is lost. To show the detail, the scale is lost. Auntieruth55 (talk) 16:07, 15 March 2010 (UTC)[reply]
There are many problems with the alt text, not least of which it tells me that in this picture that I'm being shown a grass pitch. It could be a lake for all I can see from the picture. And what image will a reference to the colour "claret" or "light blue" summon in a blind person's mind? The last image in the article is of exceedingly poor quality, and the associated alt text seems to be a work of unrelated fiction. And that's just for starters. --MalleusFatuorum17:05, 15 March 2010 (UTC)[reply]
While I agree colour may often be irrelevant in alt-text, there are three flaws in assuming colour is always irrelevant to a blind person. Firstly, most "blind" people were not blind from birth, so colour may still summon an image. Secondly, colour may carry information (the cheap seats are claret and the light blue ones are for the guests [just making this up]). Lastly, colour carries cultural meaning (such as pink for girls) that even a blind person knows about. Colin°Talk17:58, 15 March 2010 (UTC)[reply]
There may indeed be occasional reasons to mention colour – blue sky may be one for instance, where that may convey an impression of cloudnessness, even though "clear sky" would probably be preferable – but this isn't one of them. And there are innumerable examples of the alt text providing information that just isn't in the image, as in the example of the map of Chile I gave earlier, or the Tuef Moor picture I linked to above. The point that is repeatedly getting lost is that the alt text is supposed to describe the image succinctly, in a few words, not interpret it, embellish it, or attempt to communicate its every last detail. If the precise detail is significant then it ought to be covered in the body of the article's text anyway. There is no justification for alt text longer than one short sentence, and indeed every reason to avoid it if you consider how it's actually used by screen readers or text-only browsers. The often repeated "pretend you're describing the image to someone at the other end of a telephone" is not at all helpful and has led to the present situation in which alt text has become an unwelcome chore for many but what's worse doesn't even fulfill its stated purpose. Hence in its current crippled form it has no place being part of the FA criteria. --MalleusFatuorum18:36, 15 March 2010 (UTC)[reply]
They say a picture tells a thousand words. I find that to properly describe every aspect of the image rather than a one-line sentence that really doesn't provide any information, the reverse is often true as well. I find alt text incredibly difficult to implement, and I have trouble bringing it into the range of one of two sentences. If we can't properly describe an image to a person who cannot see it, or are prevented from doing so by a constant need to reduce the number of words, what is the point of incorporating it? A short and half-assed description is no better than a lack of alt text. I don't see any point in using it if the help it provides is limited at best. MelicansMatkin (talk, contributions) 19:10, 15 March 2010 (UTC)[reply]
Some of these issues are covered by the guideline and some are currently being discussed on the guideline talk page. What is certain is that all that is being achieved by an alt-text discussion on WT:FAC is that folk get worked up / let off steam and achieve nothing all at the same time. Colin°Talk19:22, 15 March 2010 (UTC)[reply]
It would be rather easy to remove the requirement for alt text from the FA criteria until the various issues surrounding its correct use are resolved, which is what I believe ought to happen now. Then there would be no further need to discuss it here The present alt text offerings are not setting a good example and by no means represent our best work in this area. --MalleusFatuorum19:29, 15 March 2010 (UTC)[reply]
I agree that it should be removed for now, until all of this is fixed at least. I know I am not a regular contributor to FAC reviews or this talk page, but I have been following this for some time to try and work my understanding of the process. It seems to me that if something is confusing or broken, we shouldn't have it as a requirement until it is clarified or repaired. Anything incomplete really shouldn't be considered mandatory in representing our best works. MelicansMatkin (talk, contributions) 19:40, 15 March 2010 (UTC)[reply]
Criterion 3 says "brief and useful alt text when feasible". Under the current guidelines it clearly isn't feasible, and so should be dropped until it is. --MalleusFatuorum19:54, 15 March 2010 (UTC)[reply]
How is it clearly not feasible? Until it is actually established that our current efforts at Alt-Text is actually more damaging than helpful, then the requirement should stay in place. With all due respect, it is unwise to rush to judgment based on semi-informed opinions (and I include myself here) about what "good" alt-text is. Currently efforts are being made to get some external reviews of the guidelines, and if these come back saying that they are totally out to lunch, I will certainly support temporary removal of the requirement until the problems are sorted out. But that's not the case yet, and it may well be that the guideline is in good shape. In the meantime, if significant changes in the alt-text guidelines do take place in the next few weeks, I personally promise to do any required rewriting to the alt-text of FA articles promoted from now until the guideline stabilizes. --Slp1 (talk) 20:15, 15 March 2010 (UTC)[reply]
With all due respect, some things are self-evident and do not need expert input. Like the example of the map of Chile I gave earlier, the alt text for which gave the dimensions of Chile, information that was not available in the image or otherwise available to a reader not seeing the alt text. Nevertheless I'd be prepared not to press for the suspension of the alt text requirement in FACs in view of your offer, at least in the short term, pending expert input. I will though continue to insist that alt text is accurate, does not interpret the image, and does not include detail not available to a reader using a conventional browser. --MalleusFatuorum20:34, 15 March 2010 (UTC)[reply]
The Chile example was a good one, and the alt text for that example obviously needed improvement. However, that relatively-minor problem does not support the claim that alt text "clearly isn't feasible". Alt text clearly is feasible in most cases, including that one. There are some exceptions where alt text isn't feasible in the usual sense, and these exceptional cases are already covered in the guideline (see WP:ALT#Placeholders and WP:ALT#Purely decorative images). No doubt the guideline can be improved, and this is being looked into, but there's a big difference between a reasonably routine improvement to the guideline and a sweeping claim that alt text "clearly isn't feasible". Eubulides (talk) 21:04, 15 March 2010 (UTC)[reply]
That Chilean example I gave was from an FAC it was claimed the alt text in had been checked for quality, and I'd be very confident of finding similar issues with alt text in almost every FAC, There clearly is no quality check, and neither can there be until there's a sensible standard to judge alt text against. Nevertheless, I'm prepared to wait and see what success Slp1 has in recruiting the views and assistance of external experts before pursuing the removal of alt text from the FA criteria, as I said above. --MalleusFatuorum21:18, 15 March 2010 (UTC)[reply]
The previous comment seems to be based on incorrect assumptions. First, no reviewer has said anything like "the alt text is all OK" on that review page. (I was asked to check the alt text for quality, and fixed some obvious problems, but have been busy and have not yet done a full check.) Second, your complaint about the Chilean example was about alt text that was added to the article in response to your stated opposition to the nomination on alt-text grounds. Had you not opposed, the alt text for that image would no doubt have been left alone, and would be in much better shape than it is (at least, it would have addressed your complaint). It is not right to cite, as an example of the failure of the current process, alt text that was inserted in response to your intervention using a different approach. Eubulides (talk) 22:36, 15 March 2010 (UTC)[reply]
The alt text on this image from this FAC certainly had been QA'd, as I explicitly asked about it earlier. There is no way from that picture a reader could tell the house was made of brick. This problem is endemic and is being pushed under the carpet. --MalleusFatuorum23:04, 15 March 2010 (UTC)[reply]
No, the image you're talking about was not QA'd: it was added after the reviewer wrote "alt text fine". So this is another example based on incorrect assumptions. Look, I'm sure that one can find errors in alt text. even (if one looks hard enough) in alt text that's been reviewed. So what? That doesn't mean alt text is worthless, any more than the existence of errors in reviewed article text and citations mean that article text and citations are worthless. Eubulides (talk) 23:24, 15 March 2010 (UTC)[reply]
You'll be running out of carpet soon Eubulides. I asked Ucucha explicity about that FAC earlier today, and I was reassured that it had been checked, and that all FACs are checked, for the quality of their alt text. Don't shoot the messenger. It's perfectly clear that alt text isn't being held to the same standard as every other aspect of an FA candidate, and that needs to be sorted. --MalleusFatuorum23:37, 15 March 2010 (UTC)[reply]
I check when articles arrive at FAC, and don't have time to watch every article that is currently at FAC for people adding new images. If you look at the current FACs, you'll see that there are several where I commented on the quality of the alt text or made fixes myself. Ucucha23:47, 15 March 2010 (UTC)[reply]
I suppose that's fair enough, as you don't actually support the article, but your "tick in the box" inevitably means that other editors probably don't check the alt text for themselves, thinking the job's done, which in this case it clearly wasn't. Adding images during an FAC is a fairly rare occurrence though. Would you care to point me towards an FAC whose alt text you'd stand behind right now? --MalleusFatuorum23:54, 15 March 2010 (UTC)[reply]
Look at any FAC at the time I approved of the alt text. I can see how that is a problem, but the same goes for Ealdgyth's source-checking and other people's image copyright checking. Ucucha00:05, 16 March 2010 (UTC)[reply]
Not quite the same, because the alt text is hidden unless you take special measures to see it, and nobody appears to have the vaguest idea as to how to write it anyway. Image licensing and sourcing are well established guidelines. WP:Alt isn't. --MalleusFatuorum00:21, 16 March 2010 (UTC)[reply]
It is very similar. Most readers don't look at image licensing, because one must take special measures to see it: one must click on the image and look at the file page. Similarly, most readers don't look at the alt text, because they need to take a special measure (click on the image and look at its properties). Sources are even worse than alt text: not only do most readers not read the sources, in many cases most readers don't even have access to a source. Furthermore, most readers (and many editors) don't know how image use policy and sourcing really works; often FAC is the way they learn this stuff. Sources and images are regularly added to articles during the FAC process, just as alt text is. The alt text guideline is newer than many other guidelines, but it is well established here and the problems and concerns mentioned above do resemble those of images and sourcing. Eubulides (talk) 01:45, 16 March 2010 (UTC)[reply]
Almost images and clips are one of self-made, PD or copyright. For copyright media, you copy and adapt a FUR you've seen recently, or an existing FUR in the same File. So almost media are no-brainers.
Sources that are not easily accessible are usually accepted as WP:AGF, so there's no special measure.
Malleus and I have already examples with the current "physical description" guideline is inappropriate. WP:ALT has not been thought out, should not be a guideline in its current change, and should not be a FAC / FAR require until well thought out. Examples include:
The significance (short) is more useful than the "physical description" (long) - e.g. Nelson's death on Victory.
Sounds. What's the "physical description"?
Other non-visual senses, e.g. smell and taste - very significant for e.g. Durian, Chili pepper or Chili powder. What's the "physical description" for these?
Sheet music, with the notes and staves and key signatures and all the other stuff I never understand in my brief piano lessons. --Philcha (talk) 05:21, 16 March 2010 (UTC)[reply]
Given the strong feelings here, but a complete unwillingness to actually do something productive about it, I have placed WP:ALT "under discussion". Please see Wikipedia talk:Alternative text for images#Guideline. This step follows our Wikipedia:Policies and guidelines procedure for discussing and potentially demoting a guideline. If anyone knows how to advertise it more widely (community portal?), please do so. I don't recall anywhere on that procedure that guidelines should be discussed at WT:FAC. Would everyone please move this discussion where it belongs. Thank you. Colin°Talk21:24, 15 March 2010 (UTC)[reply]
This discussion is about whether alt text as currently implemented should remain as one of the FA criteria. As such it is perfectly proper that it is discussed here. --MalleusFatuorum21:28, 15 March 2010 (UTC)[reply]
That's precisely what I think of the current alt text guidelines. Lame. My view is that alt text has been given a free ride so far as the FA criteria is concerned, because "we must do something about accessibility and this is something". Time to get real. --MalleusFatuorum23:09, 15 March 2010 (UTC)[reply]
Which reality would you prefer? The one where Wikipedia removes an accessibility guideline, or the one where it pretends it doesn't have one? I thought Wikipedian's worked together to improve and fix guidelines. I guess I was mistaken. Colin°Talk00:22, 16 March 2010 (UTC)[reply]
I'd prefer the one where wikipedia gets it right, instead of going off half-cocked. The present guideline should never have been exalted with that name; let's hope its replacement will be better. If not, then I will continue to argue that the alt text requirement is removed from the FA criteria. --MalleusFatuorum00:28, 16 March 2010 (UTC)[reply]
Related proposal at VP that would make this criterion irrelevant
Please see here. Since alt text is a bigger issue than FA criteria, I have posted something at VPT that might make the entire discussion moot. I write this with the obvious caveats that the proposal may be rejected and I cannot unilaterally implement any change to Mediawiki software. —Justin (koavf)❤T☮C☺M☯ 20:42, 13 March 2010 (UTC)[reply]
(a) well-written: its prose is engaging, even brilliant, and of a professional standard;
(b) comprehensive: it neglects no major facts or details and places the subject in context;
(c) well-researched: it is a thorough and representative survey of the relevant literature on the topic. Claims are verifiable against high-quality reliable sources and are supported by inline citations where appropriate;
(e) stable: it is not subject to ongoing edit wars and its content does not change significantly from day to day, except in response to the featured article process.
(a) a lead: a concise lead section that summarizes the topic and prepares the reader for the detail in the subsequent sections;
(b) appropriate structure: a system of hierarchical section headings and a substantial but not overwhelming table of contents; and
(c) consistent citations: where required by criterion 1c, consistently formatted inline citations using either footnotes (<ref>Smith 2007, p. 1.</ref>) or Harvard referencing (Smith 2007, p. 1)—see citing sources for suggestions on formatting references; for articles with footnotes, the meta:cite format is recommended. The use of citation templates is not required.
Length. It stays focused on the main topic without going into unnecessary detail (see summary style).
The alt text discussion seems stalled, with issues of policy vs. guideline becoming the core of the debate. I'm hoping that picking apart WP:WIAFA with respect to guideline vs. policy will help refocus efforts. I've started a new section since some of this may veer off of alt text.
All of the items in Part 1 are based on policy, are long-standing, or essential parts of "Wiki's best work. Do we agree on that?
All of Part 2 is MoS, a guideline.
Part 3 is images, containing a mix of policy and guideline.
Part 4 is guideline.
With respect to Parts 2 and 4 (WP:MOS and WP:SIZE), these have always been handled as a guideline at FAC; when consensus is to override a guideline, that has applied, and no FAC to my knowledge has ever been held up strictly on 2. or 4. concerns. (To my dismay, since I personally believe WP:SIZE should be more strictly enforced at FAC because of the accessibility and readability issues caused by extra-long articles, but consensus often overrules me.) My point is that, guideline has always been enforced as guideline, not policy, at FAC. We also recently revamped 2c to remove specific wording that was introduced in 2005 when inline citations became required, so that it now more simply refers back to the guideline.
Can part of the problem be sorted by considering that Part 3 mixes policy and guideline? The parts of it that refer to image policy perhaps could be separated from those that refer to MoS guildeines, with MoS guideline portions moved to Part 2, to get FAC out of the perennial MOS wars. Alt text and captions are guideline (and some FA writers include very long captions and those get by here because consensus allows it). More importantly, I'm hoping we'll focus here on which guidelines we specifically mention in WIAFA, since the policy vs. guideline debate over alt text brings to mind another image issue, which I believe also reflects policy on original research; that is, the use of recolored or altered images in articles, which are generally OR.
Will looking at the alt text issue from this point of view help move the discussion forward? I believe it will, since we don't typically refuse FA status when consensus overrules on guidelines, but we need to sort guideline from policy issues. We don't want to see articles FAR'd or denied promotion over alt text, but we also don't want to see original research in images or alt text.
I'm more concerned when original research images are used-- since most of our readers see those and they may be carried on mirrors-- than I am about the rare instances where original research might creep into alt text on images that are hard to describe, and suggest we could handle those on a consensus basis. SandyGeorgia (Talk) 18:17, 16 March 2010 (UTC)[reply]
I disagree with "All of the items in Part 1 are based on policy, are long-standing, or essential parts of "Wiki's best work". As I've said to you (or maybe it was Tony) before, I disagree that 1a as it stands reflects either policy or good practice; "of a professional standard" is not what we should be aiming for. The prose quality of professionally written academic articles is generally tedious, inaccessable and presupposes a reasonable level of knowledge. Wikipedia articles are written with Giano's hypothetical "bright 14 year old with no previous knowledge of the topic" in mind, and and need a quite different approach than that taken by professional publications; however, the unique advantages given to us by hypertext mean that we don't have to explain everything, so the "Readers' Digest/Foo for Dummies" approach isn't appropriate either. – iridescent18:22, 16 March 2010 (UTC)[reply]
Is this really worth discussing? Alt text is part of our requirements now, we are unlikely to gain consensus to remove it. I find that alt text can be written in about five minutes, usually there are quibbles, but it really doesn't take too long to fix. I'm not a huge fan of it, but it is just one of the hoops you got to jump through here. I say this discussion is not productive. By the way, what is an "original research image"? One taken by the nominator? I hope not. I commonly take images of buildings, graves, etc. of relevance to my articles when I can, because free use can be hard to find.--Wehwalt (talk) 20:13, 16 March 2010 (UTC)[reply]
Yes, it is worth discussing, at least in part because almost all of the alt text being written is unhelpful crap, not to put too fine a point on it. It seems perverse to insist on crap being added to what are supposed to be wikipedia's best articles, and I think you may find that there's a significant majority who would be in favour of removing the requirement for alt text from the FA criteria and demoting WP:ALT from being a guideline if the bull isn't soon taken by the horns. --MalleusFatuorum20:25, 16 March 2010 (UTC)[reply]
Fair enough. If it is constructive rather than griping (and a lot of what I read above seems to be griping), carry on, then.--Wehwalt (talk) 20:30, 16 March 2010 (UTC)[reply]
I interpret "original research image" to mean where the selection of the image, or the image's cropping or other alteration, raises significant original-research issues. For example, today's featured article Ruffed lemur contains Image:Varecia variegata overbite.jpg as an illustration of a typical overbite of a ruffed lemur, but no reliable source supports the claim that the illustrated overbite is typical, or even that it is an overbite of a ruffed lemur. This sort of issue is present in most featured articles. We put up with it because most of the time it would be impossible for an article to contain all its images if we strictly insisted on enforcing WP:OR for every image. Eubulides (talk) 20:41, 16 March 2010 (UTC)[reply]
@ Iri, OK, but for the purposes of this discussion, I wanted to focus on the mix of policy and guideline in Parts 2 thru 4, hopefully as a way through the discussion. By "original research images", I meant old images that are doctored, recolored or otherwise altered; that is OR. But Eubulides raises other issues. My goal was to separate image policy from guidelines, so that in difficult cases, we can let consensus rule rather than insisting that a guideline be followed. We routinely overlook guidelines at FAC when there's a good reason to do so, according to reviewer consensus (although I rarely agree on SIZE, but I go along with consensus). I'm trying to find a way around the rare situations where alt text is a problem. SandyGeorgia (Talk) 20:54, 16 March 2010 (UTC)[reply]
Alt text is a problem in pretty much every FAC, certainly all the ones I've looked at anyway. It's only gone unnoticed that it's a problem until recently because everyone regards it as a chore and hasn't been bothering to check it. --MalleusFatuorum21:28, 16 March 2010 (UTC)[reply]
All the problems MF has cited so far were of alt text that hadn't been reviewed. But even if they had been reviewed, it shouldn't be surprising that a sharp critic of the current guidelines would find problems in articles that try to follow the guidelines. Eubulides (talk) 22:10, 16 March 2010 (UTC)[reply]
I don't believe that's true, and in fact I've offered to take a look at any FAC that's been reviewed for alt text, so pick one. What is certainly true though is that I'm no fan of the current guidelines. --MalleusFatuorum22:28, 16 March 2010 (UTC)[reply]
Getting back to the main point of this thread, it might be helpful to separate discussion of policies. The existing WP:FACR's preamble is technically incorrect, in that it says "In addition to meeting the requirements for all Wikipedia articles, it has the following attributes" and then goes on to list (among other things) several policies that are part of (not "in addition to") Wikipedia requirements. To fix this problem the preamble could be reworded to something like ""In addition to meeting the requirements for all Wikipedia articles, particularly including policies A, B, and C, it has the following attributes" where the "A, B, and C" could be a bullet list of the policies that are relevant here. Eubulides (talk) 22:10, 16 March 2010 (UTC)[reply]
Could that be done without altering 1a thru 1e? Those go way back at FAC, are standard terminology, so I'm reluctant to change them-- that would make reading old FACs a nightmare, since 1c, 1e, 1a etc. are standard FAC terminology. SandyGeorgia (Talk) 22:13, 16 March 2010 (UTC)[reply]
Well, I suppose you could fix the technical inaccuracy in the preamble by rewording it instead. And then you could highlight the policies in red or something. I can't say I'm enthusiastic about the red part. But the technical inaccuracy needs to be fixed. Eubulides (talk) 06:32, 17 March 2010 (UTC)[reply]
In addition to the criticisms of WP:ALT by Malleus and myself, a searches in Google Books and Google Scholar show a majority of sources advocate only short alt text, a minority preferred short or long alt text depending on situation, and none advocate only long alt text:
Insignificance (short) or description (long) depending on situation.[1][2]
WP:ALT contradicts all the sources, as well as having the defects mentioned above, and should be demotion from a guideline to an essay until a thorough analysis is done and there has been time for discussion. As this will take time, the requirement for alt text should be removed from WP:WIAFA as soon as possible. --Philcha (talk) 02:18, 17 March 2010 (UTC)[reply]
^Bidgoli, Hossein (2004). "Access to Text Descriptions". The Internet encyclopedia. Vol. 3. John Wiley and Sons. p. 479. ISBN0471222038. Retrieved 16 Mr 2010. {{cite book}}: Check date values in: |accessdate= (help)
^McEwan, Tom (2007). "ALT Text and Basic Accessibility"(PDF). Proceedings of the 21st BCS HCI Group Conference. 2. British Computer Society. Retrieved 17 Mar 2010. {{cite journal}}: Unknown parameter |coauthors= ignored (|author= suggested) (help)
The alt text guideline is already sensible, as it already advocates short alt text, and has for some time. I chose the most recent source above that is said to recommend short alt text, McEwan & Weerts 2007: what it says is that "ALT text should also be kept as accurate and succinct [as] possible." This closely resembles WP:ALT#Brevity's "Alt text should be concise, so that it does not burden the visually impaired reader with unnecessary text." What's news to me is that some sources (McAlpine 2001, and Chisholm & May 2008) recommend long alt text in some cases. Perhaps this advice should also be placed into the guideline, as a minority opinion. At any rate, this discussion would be better done at WT:ALT, not here. Eubulides (talk) 06:26, 17 March 2010 (UTC)[reply]
I'm unconvinced that WP:ALT "already advocates short alt text". The examples are mostly long. Section Essence says "... this job is not done unless the image's essence is conveyed to readers who cannot see the image ... roughly in the order that a first-time viewer of the image would notice the details ... can be interrupted with the remainder skipped if the reader does not want to hear more ...". AFAIK it gives nothing like the short examples in the sources cited at Wikipedia_talk:Featured_article_candidates.
IMO this discussion would be better done here first and then at WT:ALT - the alt requirement needs to be removed from FAC before more waste of nominators' and reviewers' time and possibly creating in FAs alt text that is dead wrong. Then there's time for WT:ALT as, without the big stick from FAC, WP:ALT is a guideline and subject to WP:IAR - and I see WP:ALT is tagged "on discussion". --Philcha (talk) 07:56, 17 March 2010 (UTC)[reply]
In deciding whether to retain alt text in the FA criteria, it might help to look at the guideline's development. It's currently not clear what it should be advising or whether it really has guideline status.
The page was created in 2003 to say that some images needed alt text, but only where the image's meaning was important: "If an image is an important and unique part of what you are trying to say, provide alternate text..." At that time, image captions defaulted to be the alt text (I believe).
In October 2008, a new function was added (I believe) allowing us to write separate alt text with the "alt=" parameter. The guideline continued to give the same advice as before. In March 2009 it said: "Alternative text may not be necessary if the caption itself suffices to describe the image, or if the particular details of an image are irrelevant. A practical test is to ask whether anything would be lost if the image were blanked for all readers. If the image per se contributes nothing to the article, then no alt text need be provided." It added that original research had to be scrupulously avoided.
On June 23, 2009, Eubulides added to the FA criteria that images need alt text, which implied that all non-decorative images needed it, though even the guideline at that point didn't say that. [13] E first raised the issue for discussion on June 17. Others may read the discussions differently, but I read them as saying there was no consensus to add it. When he was reverted, he restored it on July 1. See the discussions here:
Beginning on June 24, 2009, Eubulides made a series of edits over several months to the alt text page that changed its emphasis. It now required all non-decorative images to have alt text. The section saying alt text isn't necessary for images already described in the text was removed. Examples of elaborate and interpretive alt text were added (naming certain kinds of hairstyle, saying which century furniture belongs to, describing facial expressions as e.g. "haunted"), thereby increasing the risk of NOR and NPOV violations. The ability to default to the caption was also obscured—not removed, but added low on the page, not included in the lead, and made quite hard to understand for non-technical editors. These changes may not have been noticed because the page has low traffic. In May 2009, just before the changes began, it had only 500 hits and the talk page 53. Even with the current extra attention it's only on 62 watchlists.
Eubulides started requiring alt text at FACs and FARs. Editors added it when asked to because it's easier to do it than to argue, especially when you're under FAC or FAR pressure.
In summary, there seems (to me) to be no consensus to require editors to add separate alt text to all non-decorative images, mandate it at FAC, or recommend any kind of elaborate alt text where we go into detail about facial expressions and hairstyles.
Sorry, E. I had intended not to raise this again because I respect your work and I know you feel strongly about this, but given that others are raising it, it isn't fair to keep it going as though most people are on board with it. SlimVirginTALKcontribs10:13, 17 March 2010 (UTC)[reply]
France", which contains a purely decorative image: the image is not intended to provide any info that's not in the text, and nothing happens when you click on it. Although it is true that Wikipedia 
--.jpg){kind=link}
