This page contains discussions that have been archived from Village pump (technical). Please do not edit the contents of this page. If you wish to revive any of these discussions, either start a new thread or use the talk page associated with that topic.
It would save me a couple of seconds on each move, but generally, there are not very many easy reasons to move a page. Most require a more detailed reasoning. What would you put in the drop down except "Capitalization" "Spelling error" and "per discussion at ___"? =- Mgm|(talk)10:02, 27 April 2009 (UTC)
Hi. In the spanish wikipedia village pump somebody asked about PDF versions in english wikipedia. Registered users can see a PDF version in Toolbox, at left. But IPs can not. In spanish wikipedia users and IPs can do it. Why not do the same thing in english wikipedia?
(17:47:05) *******: XalD: Just because es does something, en doesn't have to do it too
(17:47:11) *******: different languages do different things
(17:47:15) xald: that is what the IP is asking in es.wikipedia village pump
I know that es: and en: are different projects, but if users can use the button, why not the IP? You just only need modify the Mediawiki Toolbox message. Thanks, ~~×α£đ~~es23:15, 26 April 2009 (UTC)
Because it's not yet enabled on enwiki. I know there was talk of doing it soon, after it's confirmed that it will scale ok to that many anonymous users. ^demon[omg plz]15:45, 27 April 2009 (UTC)
Is it possible (with scripts, monobook.css manipulation, etc., I don't know) to temporarily watchlist a page so that it would, for example, only appear in my watchlist for a day or a week? --Conti|✉14:39, 27 April 2009 (UTC)
I was uploading a new version of Image:Air Canada Jazz.svg, and the changes I made only seemed to take effect in the thumbnail at the top (although not in the thumbnail of the file history). When looked at the SVG itself by clicking on it, it was exactly the same as the previous version. This can be confirmed by opening it in a text editor. Only when I reuploaded the file, the change took effect, and then only in the past version, with the latest file the same as the original. Am I missing something, or is this an (un)known bug? Paranomia (talk) 22:36, 25 April 2009 (UTC)
This has happened to me a few times before. After you upload, the file is updated but your browser does not refresh the image on your screen because it is stored on your cache. Bypassing your cache fixes the problem. You can find browser-specific instructions for bypassing your cache at Wikipedia:Bypass your cache. Also another note, Wikipedia also has a cache. You can purge that as well by typing ?&action=purge at the end of the URL. More info here if you're interested.-- penubag (talk) 06:27, 28 April 2009 (UTC)
When did user creation dates begin to be recorded? Typically this may be found at Special:ListUsers. However, the date is not available for my account. I vaguely remember creating the account a long time ago, but have no idea even which year it was. I was inactive for a few years, and I am sure that I made contributions before my first recorded edit, as well. Dates seem to be available for much of 2001, which makes sense as the the project began then. Was there initially a period where dates were not recorded? Or were some records lost for some reason? Just curious and it has been bugging me for a while. Thanks, Bendono (talk) 08:20, 28 April 2009 (UTC)
The use creation logs only go back to September 2005, when the feature was added. The dates on ListUsers are new, and (for users predating the creation log) seem to just be the date of the user's first edit. Algebraist08:29, 28 April 2009 (UTC)
Thanks for the response. If the above is accurate, then under what circumstances would an active editor, such as myself, not have a recorded user creation date? I have almost 12,000 edits, but none of them are being picked up as an artificially generated creation date. Bendono (talk) 10:04, 28 April 2009 (UTC)
I'm not sure, this might be possible if you registered before September 2005, but had not made any edits before that time, or possibly even if you had made a few edits between January 2001 and January 2002, when "Phase II of the software" was deployed. The original phase I software (January 2001-January 2002) was known to not accurately keep records of all the edits, so edits you made at that time might not be listed in your edithistory, and thus not have been picked up by the deployment of the 2005 functionality. I can ask the developers if you really want to know. For more info see History of Wikipedia. --TheDJ (talk • contribs) 10:48, 28 April 2009 (UTC)
< thedj> TimStarling: can it be that users registered before 2005, but with 0 edits until 2006, did not get, and still don't have a registration date for their useraccount ?
< Splarka> hmm, yes, I remember something like that
< TimStarling> yes
< Splarka> if their first edit was after the new user log was implemented they don't get a guessed date
Expand two/three letter project code into full langage?
Is there a template or an otherwise clever way to expand a two/three-letter language code into the full language? I.e. pl = Polish, etc. Thanks, –xenotalk14:04, 28 April 2009 (UTC)
Sexcellent! Were you doing this for a particular reason? I'm currently working up a standard iw-ref template and would hate to duplicate efforts. –xenotalk14:25, 28 April 2009 (UTC)
I remember we used to have a lot of similar systems on meta ... Make sure you don't do any unnecessary duplicate work :D --TheDJ (talk • contribs) 14:34, 28 April 2009 (UTC)
interesting. out of curiosity, is there a template that translates the name of region, nation or people into its adjectival form (which is usually the same as its language form? for instance, Germany->German; Asia->Asian; Arab->Arabic...? —Preceding unsigned comment added by Ludwigs2 (talk • contribs)
Interesting. I can find the adjectival forms 'Saskatchewanian', 'Klingon' and 'Lilliputian', but not the ones for people from Belgium or the Democratic Republic of the Congo. that's gotta say something about wikipedia.
I suppose I could write up a template (just be a huge switch statement, with the obvious maintenance problems). would there be any use for that, do you think, aside from the one use I have for it? --Ludwigs219:23, 28 April 2009 (UTC)
define reasonable. for starters, I'd have to go to each and every country data template {{country data England}}, {{country data Argentina}}, {{country data Australia}}, etc. - all of which are fully protected pages - and request the addition of a new parameter/data combo. well, I suppose I could get an admin to AWB it... It would make sense to do it that way, I'd just want to research it some more, first, so we don't screw up half a gazillion flag templates in the process.--Ludwigs221:18, 28 April 2009 (UTC)
That is certainly a barrier. :p If we do add it, we can at least be pretty sure it won't change or need to be changed very often, but I certainly understand the hesitation. – Luna Santin (talk)06:25, 29 April 2009 (UTC)
Seamlessly collapsing portions of a table
I am attempting to seamlessly collapse portions of a table, found here:
Sorry I didn't make myself clear, I meant maybe some kind of greasy monkies script that would strip the underscores on the fly as I'm saving the page... –xenotalk23:44, 23 April 2009 (UTC)
I don't think it's a good idea to do it automatically, right now on this page there are at least 2 signatures with user_talk, and in articles some links may legitimately contain underscore as well (say, as a link to FILE_ID.DIZ). —AlexSm15:53, 24 April 2009 (UTC)
Yea, you're probably right. I was hoping that there was some script that could intercept when I pasted a link with underscores and strip them. Probably too much effort for too little gain =) –xenotalk15:59, 24 April 2009 (UTC)
It wouldn't realy break anyting though, as far as MediaWiki is concerned spaces and underscores are treated the same way in internal links (or rater all spaces are converted to underscores in the resulting HTML markup) so FILE_ID.DIZ and FILE ID.DIZ are both links to the exact same article. --Sherool(talk)17:46, 24 April 2009 (UTC)
Well, I didn't say "break", just that it's not nice to change it automatically: in articles it might be intentional, and on talk pages it would creates unnecessary changes, resulting in more complex diffs. —AlexSm16:27, 25 April 2009 (UTC)
Xeno, if you're willing to select some text and then press a button, you might want to try my script user:js/urldecoder with extra option urlDecoderIntLinks=true, although it does much more than just replacing _ to spaces. —AlexSm16:27, 25 April 2009 (UTC)
No. A well placed EMP or asteroid strike would involve a degree of unrecoverable losses. A large fraction of the data would eventually be recoverable via various mirrors and dumps and such, but doing that would be a side effect of our distributed information culture and not really part of a data protection plan per se. Dragons flight (talk) 23:34, 23 April 2009 (UTC)
An EMP attack is pretty likely on a 100year timescale. There are many datacentres which offer pretty cheap EMP-resistant storage. At the very least, an antipodeal backup should be provided to ensure that the strike doesn't destroy all of the sites. Databarracks is one such UK provider. Andrewjlockley (talk) 09:09, 24 April 2009 (UTC)
Ha. Yes an eight-year-old organization should plan for a "100 year" event that has never happened in Florida (the main server farm) and pretty much can't be expected to happen save for nuclear war and the end of civilization. Backups are good, distributed backups are better, but let's not get silly about this. Dragons flight (talk) 09:32, 24 April 2009 (UTC)
Don't forget we also have a nice backup on the toolserver in Germany, so the disaster would have to take out Florida and Germany to be successful. MBisanztalk09:46, 24 April 2009 (UTC)
Neither do big chunks of the user table, or any deleted/oversighted revisions, or any media AFAIK. The toolserver is probably not a very useful backup mechanism: we'd have all 280 million revision entries, but no associated text! Happy‑melon10:39, 24 April 2009 (UTC)
There's a full copy of everything at the Toolserver/Amsterdam-data-center except for images; toolserver users just don't have access to everything for privacy reasons. Toolserver users have access to views of the normal database tables, the text is on clematis. Mr.Z-man22:51, 24 April 2009 (UTC)
They're looking to build a complete backup datacentre to run the site (not just store it) somewhere in the northeastern US. I don't know if that will incorporate database slaves that would retain the site content in the event of the unrecoverable loss of the database servers in Florida. But in any case a huge amount of our 'current' content is stored in the memcaches of the Apache servers, which would definitely be duplicated 'up north', and on the squid servers in Florida, Amsterdam and Seoul; those are roughly evenly-spaced around the world. Someone would just have to write an 'unparser' to convert from HTML back to wikitext. And of course, anyone who's downloaded a copy of the database dump is storing a complete copy of the most important part of our data; that could easily be restored if necessary. If an EMP were to hit right now, we'd probably just be set back to August 2008, when the last full enwiki dump was completed, or to March 2009 with the last current-version dump (and a lot of messy licensing issues on who to credit going from one to the other). I'd say that between copies of the dumps, what data is replicated to the toolserver, and whatever's lying around on Brion's USB key (:D) we could probably recover pretty much everything up to a month or so ago; the most likely things to be lost are user preferences/watchlists, user passwords (which could be a bit of a bummer) and deleted pages. The oversight table would almost certainly be lost (bah humbug :D). But as DragonFlight says, thanks to the information age, it would be a tall order to actually permanently lose much 'public' content, even in an end-of-the-world scenario. Happy‑melon10:39, 24 April 2009 (UTC)
The Seoul cluster, which was always the smallest of the three, is being decommissioned (or already was). It was part of a space sharing agreement with Yahoo that is ending / ended. Also, the last full history text dump of enwiki was more than 2 years ago, and not in 2008. We've had aborted dumps, partial dumps, and corrupt dumps since then. Yes we could get off the ground with the current versions dump (from March) and stub-meta-history for author lists and copyright purposes (last completed in October), but if an act of God destroyed Florida there would probably be a lot of article history that we never recover. Dragons flight (talk) 11:02, 24 April 2009 (UTC)
Ah, was it 2006 not 2008? Interesting also about the Seoul cluster; I wasn't aware of that. They're still listed on wikitech, so I guess the curtain's still up for the moment.
Actually, with the current versions dump from John Smith's computer, plus the revision and logging tables from the toolserver, we'd have all the info we needed for copyright purposes: a text to work from, and a complete list of its authors. So it would just set us back to March, and we'd have to delink everything before that magic 'first version' in page histories. It would be wierd, but not end-of-the-world. Happy‑melon11:29, 24 April 2009 (UTC)
I think it's entirely reasonable to protect WP from the consequences of a nuclear/EMP war. I don't think distributing the data would be sufficient. I think that a nuclear-hardened backup should be specifically created. Andrewjlockley (talk) 12:17, 24 April 2009 (UTC)
If you want to put up the time and money to create such a backup, then you are certainly welcome to copy any and all materials from m:Database dumps to any repository you wish. As for Wikimedia working on such a project, I think it would be a waste of resources. Dragons flight (talk) 12:46, 24 April 2009 (UTC)
What's the size of the backup? Bearing in mind that there are already easily accessible nuclear-hardened stores, what's the argument for not using them? Andrewjlockley (talk) 02:14, 25 April 2009 (UTC)
can you download the whole think to DVD's? they shouldn't be affected by an emp. Is anyone doing this already?
Dye-based media such as CDs and DVDs have a limited shelf-life, especially when exposed to light. Not sure whether magneto-optical devices such as minidisks (which have a longer shelf-life) are also susceptible to EMP. I believe Sony is developing/has developed a high-capacity storage medium based on this technology, but intended for institutional needs. SharkD (talk) 04:32, 28 April 2009 (UTC)
Concerns about Wikimedia's disaster recovery infrastructure should be directed to people who actually know what they're talking about, rather than random Wikipedians who read this page. You should also do your homework before asking anywhere. A lot of the information being thrown around here is third-hand "I heard it somewhere" kind of information. — Werdna • talk13:50, 29 April 2009 (UTC)
The problem has been raised (in a very roundabout way, unfortunately) that short pages may be especially vulnerable to defamatory attacks, since material added to them is likely to show up on Google search result pages, and continue to do so (because of the Google cache) even after it has been deleted from Wikipedia. I don't entirely see this myself, but if anyone can add any enlightenment on the problem, please comment at WT:WikiProject Disambiguation/Vulnerability to attacks.--Kotniski (talk) 10:54, 29 April 2009 (UTC)
If I try to go to Rastafari movement I get this:
<link rel="stylesheet" href="/w/index.php?title=MediaWiki:Print.c
at the top of the page with the usual background but if I click Rastafari which redirects to Rastafari movement to works ok...
The url of the links to each version in the date/time column are out of sync with the displayed
date/time and with the displayed images. e.g. the date 13:33, 29 April 2009 links to a file with 1353 in the url. The 13:05, 29 April 2009 date links to a file with 1333 in the url and so on down the list. Charvest (talk) 14:08, 29 April 2009 (UTC)
Those are actually linking to the correct files. The way we currently name archive file versions is a little odd: it's got the timestamp of when the file was replaced rather than its original upload date. --brion (talk) 16:11, 29 April 2009 (UTC)
Does the English Wikipedia currently have a method for toggling between images? I'd like to add a button that toggles between two images of identical size for an article at FAC, the Euclidean algorithm. The idea is to allow the reader to shut off a GIF animation by replacing it with a still image of the same size. An FAC reviewer suggested that the animation, although helpful for understanding the algorithm, distracted them from reading the article. I do not want to convert the GIF into an OGG movie. The frames of the animation are available as SVGs. Thanks for your suggestions, Proteins (talk) 13:29, 30 April 2009 (UTC)
More than 20 groups please
Hey, an editor has spoken to me of his difficulty in creating a template for the singles of Prince because the template does not allow anymore than 20 groups. Can this please be increased, it seems harmless enough. Thanks. — R216:14, 30 April 2009 (UTC)
Is there a way to sort navbox columns in a way similar to a sortable wikitable? There is a an ongoing debate regarding the table at Template:2009 US swine flu outbreak table, some people feel that it should be sorted alpha by state, while others argue to sort by number of cases. If this could be made sortable it would solve the problem.
Tool for editing and viewing wiki markup, offline?
I use Linux (Ubuntu 8.04, Hardy Heron). I would like to be able to copy and save the plain text versions of Wikipedia articles (with the wiki markup) and edit and view versions of the articles at home. I want to practice editing tables, and so forth, without having to use the Sandbox. How can I do this? Thanks very much! Whatever404 (talk)
That sounds like fun- but can I spoil the party and ask, how. And if that question is too naive, how closely does the Turnkey Ubuntu version emulate our classes and settings, which are needed for table work? It does sound like a useful exercise. --ClemRutter (talk) 13:32, 20 April 2009 (UTC)
Thanks, everyone. Charlotte: if I install the items you mentioned, does that mean that the wiki environment on my computer will function in the same way as Wikipedia? Whatever404 (talk) 00:39, 24 April 2009 (UTC)
Whatever404: I see that you have no subpages in your user space yet. The usual way we test things is to create test pages under our own user page. Here is a link to your first personal test page: /Test1. Numbering your test pages usually is a good idea, since they tend to become more over time. I created my /Test57 some day ago... And I hope you have discovered the Show preview button when you edit pages?
Thank you, David, but I was interested to edit tables in ways that would probably only be useful to me, and would probably not be used in articles. Some of these edits are related to medical information that I would rather not make public. I guess it would be a more appropriate thing to ask the Wikimedia developers but I didn't really know where to go, to do that. Thanks. Whatever404 (talk) 16:53, 24 April 2009 (UTC)
I'll add that there are times I'd love a lightweight, offline app that I can use to test out tables, templates, and other wikitext structures without editing directly into wikipedia. it never actually occurred to me to set up a local wiki, and it sounds a bit like overkill. but... --Ludwigs217:22, 24 April 2009 (UTC)
You could maintain the text & markup in whatever editor you like, copy and paste it into an edit window, and use the "Show preview" button to view the results. Ntsimp (talk) 14:39, 27 April 2009 (UTC)
Recently, Mammamia9905 has been blocked as a sockpuppet of Petergriffin9901. BaltoPat9902 was blocked earlier. Naturally, I'm interested in finding usernames that end in 9903 and 9904 so I can see what they've been up to. Can anyone think of a reasonable search method?—Kww(talk) 12:10, 1 May 2009 (UTC)
Sure, easy enough with grep and a list of usernames. The problem of course is that there are about 9 million usernames (9,577,482 in fact) on en.wiki, so any search of them not that's not optimized is a bit nasty. For your specific query, the raw data (all usernames on en.wiki) is available here: tools:~mzmcbride/all-users-2009-05-01.txt.gz (37 MB). I grepped the file for ".*990[0-9]$":
An embarrassment of riches, but not as bad as you might think. Relatively few are actually Name9902 and Name9903. It'll take a few hours, but I'll manage it some afternoon when I'm bored. Or maybe just submit a massive checkuser request to keep those guys on their toes.—Kww(talk) 22:05, 1 May 2009 (UTC)
History tab at the top
This discussion has been closed. Please do not modify it.
The following discussion has been closed. Please do not modify it.
Precedent: Articles at the top have tabs with labels such as "discussion" and "history". The discussion tab leads to the talk page. The use of the word "discussion" is good because it is easy to understand and not jargon. Once someone clicks it, it shows the talk page (not named "discussion").
Suggestion: There is a "history" tab. Like the above, a history page is fine. However, the tab should not say "history" because this is jargon. Instead, it should say "authors".
Use of "authors" would give credit to editors, much as we give editors credit in the form of barnstars. It's just a nice thing to do to encourage writing, give recognition, and reduce jargon. User F203 (talk) 18:23, 1 May 2009 (UTC)
But that would be incorrect, implying that the page is a list of editors who have contributed to the current page text. But the history page also includes vandals, edits that were reverted, edits that have been totally overwritten, random drivebys by bots, protecting admins who have made no actual contribution to the content, etc etc. All authors are listed on the history page, but not everyone on the history page is an author. Happy‑melon18:37, 1 May 2009 (UTC)
Every editor listed was an author is some form, even if what they wrote was reverted. But this is the wrong forum per instructions at the top. It also doesn't have to say "authors". It could say "editors". Even admins who protect it get some credit because they shaped the article by protecting it. Look at this suggestion as positive! User F203 (talk) 19:28, 1 May 2009 (UTC)
It's been called history for the longest time, I doubt we will change it now. Too ingrained in the gestalt. –xenotalk19:31, 1 May 2009 (UTC)
I remember seeing another tab change about a year ago so change is possible. This is a positive change! Anyway, according to the instructions, this is the wrong forum so please end the discussion here. User F203 (talk) 19:37, 1 May 2009 (UTC)
Some JavaScript could do it pretty easily (a modified version of User:MZMcBride/hideentries.js, I imagine). To do it in MediaWiki core would be a bit more tricky. There's no bot flag in the logging table. You'd have to create an option to exclude a specific user using a URL parameter or something. Not really clean. --MZMcBride (talk) 19:42, 1 May 2009 (UTC)
Hmmm... It's not too important to be honest, I just noticed that ProcseeBot can sometimes monopolize the log. =) –xenotalk19:45, 1 May 2009 (UTC)
maybe I'm missing something obvious here, but wouldn't it be possible to have the software automatically add your signature? I mean literally - when you hit the 'Save Page' button, the software compares the old version to the new, checks to see if the last characters of the newly added bit (barring whitespace) are ~~~~ and if not, adds them. you'd have to exclude changes marked as minor, and changes that occur in the middle of a paragraph, but with a little regexp it's certainly doable.
of course, I don't know what kind of impact that would have on server load, or how that would compare to the current load induced by sigbots... --Ludwigs216:27, 28 April 2009 (UTC)
Technically it would be trivial. Socially, it would be annoying as hell, when you get signatures popping up in places you don't want them. How can the software tell between an edit that's unsigned because you forgot, and an edit that's unsigned because you don't want it to be signed?? Happy‑melon17:19, 28 April 2009 (UTC)
well, I was just assuming that the only times you don't want an edit to be signed are easy to identify programatically (minor edits, fixes, section headers, etc.). As a rubric, anytime you add one or more full paragraphs to the body of a section (in the sense of text that begins on a fresh line and ends on a fresh line, and excludes headers) you want a signature; anytime you modify an extant paragraph (or add something that's only a section header) you don't. maybe there's odd cases (like adding a list item to someone else's example list) that need to be accounted for, but I bet we could account for 99% of talk and administrative edits using that rule. plus, we could always add a checkbox to suppress signatures for abnormal cases. I mean, the sigbots have rules for when they do and don't add signatures, and they seem to work ok. this just does the same thing with the advantage of immediacy (the software has access to both the current and and revised versions in a way that sigbots coming in after the fact don't). --Ludwigs218:00, 28 April 2009 (UTC)
This seems like an awful lot of effort for something that, ultimately, isn't the end of the world; the occasional unsigned comment is caught pretty quickly, and there are plenty of times where signing isn't needed. Typing "~" four times just isn't that hard. EVula// talk // ☯ //03:49, 29 April 2009 (UTC)
I'm a fan of automating what can be automated. four characters may not seem like a lot, but multiply it by a couple hundred thousand edits per day and it adds up to a lot of collective wasted time. but yeah, it's not a huge deal. --Ludwigs204:18, 29 April 2009 (UTC)
Well, there's a rumor that we might shift the discussion pages to something more akin to LiquidThreads, which would negate the entire need for an automated signing system. But personally, I think automating it would cause far, far more trouble than its worth (and would take more collective time to fix false-positives than collective time spent typing four tildes). EVula// talk // ☯ //05:47, 29 April 2009 (UTC)
yeah, that's a distinct possibility. good enough for me, unless someone else thinks it's a good idea and wants to chime in about it. --Ludwigs206:08, 29 April 2009 (UTC)
I've been assigned LiquidThreads as my next task, so you can probably expect it over the next few months, probably towards the end of the year. — Werdna • talk15:46, 1 May 2009 (UTC)
Just now Wikipedian2 (talk·contribs) asked why Twinkle tells him that his account were "too new". His account however is far older and has far more edits than are normally required. His account is also far past the the $wgTorAutoConfirmAge and $wgTorAutoConfirmCount setting from TorBlock. There also is no entry for the user in the Special:AbuseLog, so he wasn't deautoconfirmed through that (would that be a permanent deautoconfirmation, BTW?). So I'm at a loss. Is there another way to lose autoconfirmation status that I'm missing? The user was apparently recently caught in an autoblock, and was ip-block exempted. Can that have anything to do with it? Amalthea11:46, 29 April 2009 (UTC)
Hmm. I'd be very surprised if there was a bug in that trivial logic, but I'll try to work it from there. Sorry for the noise. Amalthea10:26, 2 May 2009 (UTC)
Interface redesign mock-ups
Trevor Parscal has been creating mock-ups of a redesigned, more efficient interface. Some of the preliminary sketches are viewable here. Comments, suggestions, etc. welcome here or on the talk page there. Cheers. --MZMcBride (talk) 19:39, 1 May 2009 (UTC)
I'm a fan of those mock-ups. I personally think monobook is overused and outdated, and I think it would do good to give it a little face-lift that will improve usability as well as aesthetics. cmelbye (t/c) 04:53, 2 May 2009 (UTC)
It looks like an improvement for newer users: it's simple. I like the aesthetic, but I'd probably customize it for myself: first, I hate drop-down menus (which, by the way, often fail on touch-only devices like the iPod Touch or iPhone, which don't have "hover" options), and second, I'm used to the general scheme of our current interface. It might be improved by a little tweaking using things we already have: for example, there should probably be a way to tell which tabs are active, like the orange border/bolding currently used through the "active" CSS class. Aside from that, my concern would be mainly in the implementation side of things: what technical features would it use to achieve the design features? Would these technical features degrade gracefully? Would they allow the sort of customization I'd desire? Would old customizations (user scripts) be broken? {{Nihiltres|talk|log}} 06:45, 2 May 2009 (UTC)
I don't think there is any getting around using dropdown menu's in the end. We just have too many options per page. —TheDJ (talk • contribs) 11:14, 2 May 2009 (UTC)
The way to get around using drop-down menus is not to use them. Pretend they don't exist (which may be the case for some users). The only place they are used now in the edit page is for inserting characters. The mockups are nice, but they need to work for the vast majority of users if they're going to be the default. Delicious carbuncle (talk) 16:43, 2 May 2009 (UTC)
My point is rather that we need to make sure that those menus are effective. If a redesign means, for example, that everyone browsing on an iPhone or tablet (which can't hover on elements) can't access the menus, we have a problem; if a redesign means that a pile of user scripts break really nastily, we have a problem; if a redesign prevents users from customizing the interface at all … you see what I mean? My concern is that, although the redesign might help quite a number of viewers, we don't want to break things for the rest of them, myself included. I'd love to see a live mock-up. {{Nihiltres|talk|log}} 20:16, 2 May 2009 (UTC)
Change tagging activated
Hi all,
After working out some minor performance issues, we've activated change tagging on Wikimedia. You can see the available tags, and filter by them at Special:Tags. Individual tags can be styled using the mw-tag-$tag class applied on changes lists (tag names are sanitised for this purpose).
Looks interesting but... how do you tag pages? Is there documentation for the extension/addition? –Drilnoth (T • C • L) 14:43, 1 May 2009 (UTC)
Tagging is done by the software. Presently, this is done mostly by the Abuse Filter – the actions taken by a filter can include applying a tag to the edit in question. — Werdna • talk15:40, 1 May 2009 (UTC)
Great feature indeed, though I'm a tad worried for the younger editors. The interface is getting more cluttered by the day.... —TheDJ (talk • contribs) 10:06, 2 May 2009 (UTC)
Any chance of wrapping tags in an identifiable class? I know that the whole li entry is given a class that identifies the tag (what happens if more than one tag is applied, BTW? Multiple classes?), but being able to style the "(references removed)" phrase itself would be useful. The same class should apply to all tags, as well, otherwise you have an open-ended set of classes to look for. Shall I open a bug? Happy‑melon11:03, 2 May 2009 (UTC)
Ah! Thanks. I'll repost my comment from WP:AN - An obvious measure to reduce confusion would be to link the "tag" back to the filter which appends it. Was this change announced anywhere (other than the extremely vague note at WP:VPT)? It doesn't seem to be covered on WP:AF, unless I missed it. Expect more confusion.Delicious carbuncle (talk) 18:38, 3 May 2009 (UTC)
Agreed. I keep a pretty close eye on software changes and I was still "WTF?" when I saw it the first time. It needs a link or something. Or a tooltip or a CSS mark of some sort. --MZMcBride (talk) 18:40, 3 May 2009 (UTC)
Might be helpful if we knew which page you were looking at. Sorry to hear about your troubles; while I do strongly support the ability of anons to edit, I also generally think it's in the best interest of any regular contributor to register an account to avoid various issues, including this sort of thing. – Luna Santin (talk)21:39, 29 April 2009 (UTC)
I think the removal of large amount of text from a page triggers it for ip users. You can create an account to avoid this. Chillum23:46, 29 April 2009 (UTC)
Yeah it is, I hit it almost every time I'm writing up a manga chapter list from home (because of the sources I add). 「ダイノガイ千？！」(Dinoguy1000)22:01, 4 May 2009 (UTC)
question about parserfunctions, nesting, and processor loads
I made a template - {{do list}} - to standardize text lists and keep line breaks from occurring in the middle of phrases, but another editor (User:Edokter) suggested it might be too heavy on the preprocessor. The issue boils down to whether it's more efficient to use nested conditionals (so that unused IFs never get evaluated at all) or to use sequential conditionals (in which all the IFs have to be evaluated, but nothing gets stacked in memory). eg, which of the following is better?
or (alternately) should I just scrap this template as being too expensive either way you look at it? --Ludwigs215:02, 3 May 2009 (UTC)
Neither option would be dangerously heavy on the preprocessor in the quantities used in that template, IMO. Nested conditionals get very difficult to read in line-wrapping editors, and it becomes easier to get brace counts wrong. In principle, yes, nested conditionals are fractionally more efficient (the preprocessor isn't building a stack, it's recursively expanding objects inline, so the memory usage is probably comparable), but not overwhelmingly so. Happy‑melon17:07, 3 May 2009 (UTC)
I raised the question because the template was advertised on {{Navbox}}, so it would see it's primary use there. With Navbox so widely used, each page edit would IMV have a potential negative impact. — Edokter • Talk • 17:18, 4 May 2009 (UTC)
Problem with the infobox
Hi, I'm contributing to the article, Saidabad. There is a problem with the infobox formatting. I'm not able to fix it. Any help is much appreciated. Randhir 18:48, 3 May 2009 (UTC)
The Zone and Ward variables take the value entered for Metro and wrap it like "List of <metro> corporation wards" for the template. Thus you can't wikilink the metro entry as it will break the template. Nanonic (talk) 19:06, 3 May 2009 (UTC)
If you mean the white box that tells you what tools are what, yeah. I'm having that problem too. Please; anyone; its not a minor issue!--Pr1nce0fDarkn3ss (talk) 00:21, 5 May 2009 (UTC)
as Nanonic said, you can't use links for that parameter - it needs to be plain text because it's used to create a different wikilink below. I've removed the link for the time being. if for some reason 'Hyderabad' isn't sufficient, and the 'Andhra Pradesh' part needs to be specified, the template's going to have to be modified. --Ludwigs204:51, 5 May 2009 (UTC)
Valid change tags
If you go to "Special:SpecialPages", there is a cryptic link to "Valid change tags", which can also be reached from the Special:Contributions page. One of the more unusual "valid change tags" is "nonsense movies". Not "nonesense books", "nonsense article", "nonsense Pokemon", but only "nonsense movies". Should that be instead, "nonsense article", or "nonsense moves"? Excuse me, it is "Nonsense movies?". What's up? Or should I say, "What's up, Doc?" 199.125.109.77 (talk) 23:23, 3 May 2009 (UTC)
This was a bit of a mistake, and I've gone ahead and changed a bunch of the visible ones (i.e., those with hits) for now, and have asked for the help of others (here and here) to do more. The tags, themselves, should probably in some form be renamed to their actual abusefilter numbers (or something else easily parseable), so that they can be easily traced back to their generator; or, at the very least, should be made neutral in all high-visibility areas. --slakr\ talk /01:31, 4 May 2009 (UTC)
I'm not sure what you mean by a mistake, but I now see that one sock has allegedly been using the edit summary "Nonsense movies?". However, in making this as a tag, I would encourage simply shortening it to "Nonsense?". 199.125.109.77 (talk) 16:25, 4 May 2009 (UTC)
Currently they only come from the abuse filter, as its a new feature and the abuse filter is the only thing that adds them now, but there's no restriction on where future tags may come from. Mr.Z-man18:26, 4 May 2009 (UTC)
I would like to suggest a very useful feature to wikipedia, the search bar front and center (top) so people can find what they need quickly, which is the sole purpose of wikipedia. The articles can afford to move down a little bit for a more prominent search bar. Being a web developer, its important to create ui for the user that's the ultimate goal. Its kind of annoying to look for the tiny left column search bar and type your keyword that way. I hope you can fix this problem to better the best web resource ever created. —Preceding unsigned comment added by 173.73.168.77 (talk) 02:58, 4 May 2009 (UTC)
Is there any way to exclude user and talk pages from Google, yet still allow them to be searched from within Wikipedia? 199.125.109.77 (talk) 16:36, 4 May 2009 (UTC)
If that is added, can text be found from within Wikipedia? And I was looking for a global fix, not one that had to be added in a million places. 199.125.109.77 (talk) 17:11, 4 May 2009 (UTC)
That does not affect Wikipedia's internal search function. As for global fixes, this can easily be done technically, but you'll have to get consensus first, which has proved difficult in the past. See Wikipedia:Search engine indexing. Algebraist17:22, 4 May 2009 (UTC)
As I read it the main objection before was that the internal search engine did not works so good, but it has now been fixed, thus removing that objection. If anyone wishes to pursue it, the place to bring it up is at WP:VPR, though. 199.125.109.77 (talk) 17:54, 4 May 2009 (UTC)
This is showing in Firefox but not in IE. I am also told it does not work in Safari. Why might this be? Stifle (talk) 13:43, 30 April 2009 (UTC)
Yeah, broken JPEG image (inverted b/w image). I've seen stuff like that before. I have no idea how you can fix it though. —TheDJ (talk • contribs) 15:19, 30 April 2009 (UTC)
Image loads fine for me. Consider that B/W color inversion is often a matter of data corruption, which can occur if something happens to cause your computer to not download the image correctly. -moritheilTalk00:47, 4 May 2009 (UTC)
Either percent-encode the '[' and ']' characters, using '%5b' and '%5d' respectively, or use the magic word {{urlencode:[blah]}} . The latter might be a problem if "blah" contains certain special characters, such as "/", so it doesn't seem to be suitable for full urls. I've changed Russell Ward for you.-gadfium06:18, 2 May 2009 (UTC)
Is it technically possible for a bot to take information off of wikipedia (or the below two webpages) and then contact editors with x amount of edits about an Article for deletion?:
Does anyone know how to make the "v/d/e" links disappear from the navbox here? I don't see what's up with the template code at the moment. — Carl (CBM · talk) 19:36, 2 May 2009 (UTC)
can you point to a page where it throws the error? my first guess would be that if (table.className == "navbox") might throw an error if the className for the table is undefined, but that's just a guess.
Well, the error is gone, but it does not do what I expected it to do (find the first table with classname=navbox, then assign id=first-navbox) and after the cache caught up, does exactly what is is intended to do. — Edokter • Talk • 18:14, 5 May 2009 (UTC)
Links missing on userpage?
I spend most of my Wikipedia day looking at my Watchlist, and rarely look at my main Userpage, except upon browser start-up. However, today, after restarting my normal browser mid-day, I returned to find all the normal links on my Userpage were, uh, blank.
For example, the text looks like:
Joined Wikipedia on , .
The source (as rendered in View Source from my browser) looks like:
<ul>
<li>Joined Wikipedia on <span class="mw-formatted-date" title="2007-08-16"><span class="mw-formatted-date" title="08-16"><a href="/wiki/August_16" title="August 16" oaugust=""></a></span>, <a href="/wiki/2007" title="2007" o2007=""></a></span>.</li>
</ul>
Obviously, there's no text in between the <a ...></a> tags, so nothing is displayed. But, to the best of my knowledge, I didn't change anything to cause this issue.
What's going on here? I've just discovered that expressions like RFC 333, ISBN 9999-456-456 (I don't know what else) get automatically linked to various places. Is this something recent, or has it always been like that and I've only just realized it? What is the idea behind this functionality? Where is it documented? Surely this is the sort of thing we ought to be using templates for, not have links created when they might be totally inappropriate. (See WT:MOS#RFC 1924 links in headers for a recent discussion which shows the undesirability of this feature.)--Kotniski (talk) 06:35, 5 May 2009 (UTC)
Hmm, so it looks like another example of a not-too-great idea that got fossilized in. And of how some feature documentation is hidden away on almost impossible-to-find pages. (I've added the information to Help:Link to make it perhaps a bit more accessible.)--Kotniski (talk) 09:40, 5 May 2009 (UTC)
I used google advanced search for only en.wikipedia.org to find the page and I had to try with 2 or 3 combinations to find it. --Stefantalk09:46, 5 May 2009 (UTC)
Remember, "a not-too-great idea" is your opinion, and is not necessarily true for all people in the world. I quite like it, and find it very handy, frequently.
Yes, true. But not if you start from "RFC" (and that was what raised the problem - it was a complete accident that I noticed the same thing was happening with ISBN on my way over).--Kotniski (talk) 08:13, 6 May 2009 (UTC)
I think this feature was created before templates existed. Generally speaking, our approach in the past was to automatically format things, rather than adding reams of template code to do it as we do now. I think a lot of people still consider the old approach far superior. — Werdna • talk17:35, 6 May 2009 (UTC)
It's a bit of a mess. ISBN 123456789X is picked up by the wiki-magic, but ISBN-10 123456789X isn't and a phrase like ISBN 123456789X, 987654321X is only half done. There's a bug report out there somewhere. Mr Stephen (talk) 21:09, 6 May 2009 (UTC)
That isn't a bug, it's a feature. You have to take out the -10 or -13 and just use ISBN followed by the 10 or 13 digit number, with or without spaces or hyphens. 199.125.109.75 (talk) 04:38, 7 May 2009 (UTC)
I've recently come back to my Wikipedia page (User:Xnux) after a while to discover that the template I use to organize my userboxes isn't displaying correctly in certain browsers. When I click one of the first four "[show]" buttons on that template, the other "[show]" buttons are not moved down; instead, they overlap the userboxes. This happens in Mozilla Firefox 3.0.10 and Google Chrome 1.0.154.64, but not in Internet Explorer 8.0.6001.18762. Does anyone have advice or suggestions? Thanks, XnuxtheEchidna02:27, 6 May 2009 (UTC)
I have ran a test. The cause of this problem is the "float:left" that all the userboxes have. The HTML of UBX'es boggles my mind btw. I have no idea how that came to be, but it's truly beyond saving. I don't think there is any way you can quickly solve this problem. I'll ask AzaToth about some of the original decisions in userbox layout, perhaps he has a clue. —TheDJ (talk • contribs) 09:31, 6 May 2009 (UTC)
I suspect the problem has to do with using collapsible divs inside collapsible tables. let me see if I can fix it by converting the inner sections to {{userboxtop}} format. --Ludwigs213:08, 6 May 2009 (UTC)
yeah, that was it. I've fixed it on your page. I'm debating whether I should add a note about it somewhere... --Ludwigs213:35, 6 May 2009 (UTC)
Watchlisting single pages
Can Special:Watchlist not be used to watchlist only the main article? It can be useful sometimes if e.g. a talk page has more traffic than its corresponding main page or vice versa and ideally only the latter is watched. -- Mentifisto03:36, 6 May 2009 (UTC)
You can create a user subpage containing links to only the precise pages you want to watch, and then look at the related changes of that page. Functionally, it's pretty much the same as a watchlist. You can link to the changes with [[Special:RelatedChanges/Pagename]] for quick access. Someguy1221 (talk) 07:22, 6 May 2009 (UTC)
A decent workaround, but why can't this just be sorted in the code? I have so many javascripts rewriting the look of pages it starts to get distracting. –xenotalk13:40, 6 May 2009 (UTC)
Article & talk pages are always treated as a unit for the watchlist -- this is vital, since otherwise you would never see discussion updates that are relevant to your article, or article updates that reflect the discussion. It's a deliberate architectural choice and we have no intention of changing it. --brion (talk) 15:04, 6 May 2009 (UTC)
But sometimes a talk page generates so much discussion (trolling and the like) that you don't want your watchlist flooded with, although you do want to see if anything actually changes on the article itself (or it could be a policy page, for example). I don't think this is the kind of choice that the devs should be making for us - if it's easy enough to change, then please give us the option and we'll decide for ourselves whether we want to use it.--Kotniski (talk) 15:19, 6 May 2009 (UTC)
I often want to watch, say, the FA of the day simply for any talk page stuff that comes up. There's also the most obvious example -- user pages, usually you wouldn't need to worry about the 'article' page in that case. ♫ Melodia Chaconne ♫ (talk) 17:08, 6 May 2009 (UTC)
Code corrupts my talk page
Could someone take a look at the last few sections of my talk page as some code which I'm unable to debug makes all subsequent entries indent. __meco (talk) 08:56, 6 May 2009 (UTC)
It's because the football squad template is a full width template and shouldn't be indented. Removing the indent fixes the page. Nanonic (talk) 11:39, 6 May 2009 (UTC)
I rely heavily on Wikipedia but since recently updating to Windows Internet Explorer 8 I have been unable to open Wikipedia pages (in English). I am given: “Do you want to save this file, or find a program online to open it” followed by “Name: Viticulture, Type Unknown File Type, 18.1KB, From en.wikipedia.org”. A click on Find tells me: “Windows does not recognize this MIMI type.” !!! Note: I am a Wikipedia reader, not an editor. How can I return to accessing your pages? (I first tried to find a solution on the Help Desk and was told to try the Village Pump.) —Preceding unsigned comment added by 93.41.186.175 (talk) 10:28, 6 May 2009 (UTC)
18.1 KB is roughly the size of the first image in the Viticulture article. It sounds like IE8 is unable to recognize the .jpg file type. I have no idea why that would be. Going back to IE7 might be your best bet until IE8 is a little more stable. Franamax (talk) 11:46, 6 May 2009 (UTC)
Make sure the box in the editing tab of your preferences entitled "Use external editor by default (for experts only, needs special settings on your computer)" is unchecked. Otherwise, I don't know. Graham8701:24, 7 May 2009 (UTC)
That would only be a problem if you were logged in. If you're logged out, you should be able to view all pages properly; if not, it's something to do with your computer or browser. IE 8 has been around for a while, and this issue hasn't been reportedt, so it probably doesn't occur in the default installation of that browser unless something else is wrong. Graham8701:30, 7 May 2009 (UTC)
tables and divs
has anyone else noticed that tables and divs don't play well together? at least not in Safari, and sometimes in Firefox and Google Chrome. see the changes I made from here to here on User:Xnux, and a similar problem I had on Template:WikiProject_Footer between this and this. the first case, the collapsible divs inside the {{userboxtop}} table were overwriting each other for no particularly good reason; in the second, {{WPCouncilRec}} (which is a table) was getting sucked into the outer div of the old version of the footer template on Safari. I think it might have something to do with the way 'float' elements work inside of containers, but I can't tell whether this is a rendering bug in Safari and/or Firefox, or an inheritance issue in wikipedia's CSS (or maybe just my bad luck to see similar problems in quick succession?). if others have noticed it, I'll set up a test case page so we can figure it out. --Ludwigs216:40, 6 May 2009 (UTC)
It's pretty old news. Tables and (floating) divs have never played well toghether, and not just on Wikipedia. Divs were heraulded as a replacement for tables when used for page layout, but have serious shortcoming. Whenever possible, whenever (a combinations of) boxes require more control over placement, we tend te deprecate divs in favor of tables. That is why most newer templates (ie. Navbox and Ambox) are built with tables using native HTML. — Edokter • Talk • 16:51, 6 May 2009 (UTC)
Yup, inherent to the widely differing boxmodels used for tables and div's. (Note that in some modern browsers, you can give a div "table-like" behaviour, by specifying "display:table" instead of "display:block".) All in all, the div model basically does not work as many people expect, and even more so in Wikipedia. This quickly leads to errors. Combined with the many bugs that used to exist in div rendering, tables simply behave more consistent among many browsers, and it is the reason why both *mbox and navbox code has been designed to use tables. But to reiterate the UBX issue: "Lack of design". —TheDJ (talk • contribs) 17:16, 6 May 2009 (UTC)
This may have been asked before many times, so apologies if it has. Are there any plans to properly unify the login, so your edits are moved over & also, so if you want to, you can use the same userpage? Thanks. dottydotdot (talk) 19:59, 30 April 2009 (UTC)
But why would they need to be mirrored? The content that was edited is not duplicated to that other site, unless it is imported. Happy‑melon07:51, 1 May 2009 (UTC)
I do believe that some work has been done on creating global preferences. So that you would only need to set your signature and timezone once. —TheDJ (talk • contribs) 09:48, 1 May 2009 (UTC)
So that somehow, it would show that you had 500 edits on Wikipedia & you're not completely new. I was just wondering if there were any plans to make it a bit more "unified" but I can see the want to keep them separate. Cheers dottydotdot (talk) 10:16, 1 May 2009 (UTC)
Showing that you have made 500 edits when you haven't would be wrong. There are several tools people can use to see how new someone is more reliably than that anyways. — Mike.lifeguard | @en.wb19:18, 2 May 2009 (UTC)
The proposal to mirror edits across projects is consistent with the notion of a "unified identity" which is quite different from the notion of "unified login". "Unified login" is a user interface convenience, but "unified identity" changes in basic ways how the editor role is implemented. --User:Ceyockey (talk to me) 11:30, 3 May 2009 (UTC)
I've been having problems with edit preview pages expiring under Internet Explorer 7.0.6001.18000. If I navigate away to check a detail and return IE says that the page is expired. Sometimes even it claims not to bable to find the page when I finish an edit. This problem, has only appeared in the last couple of days.--Peter cohen (talk) 10:40, 1 May 2009 (UTC)
This is a general problem IE has with forms (it also tends to lose any changed values in the form). Your best bet would be to get into the habit of opening new tabs/windows for looking stuff up while you have an edit window open (or you could upgrade toarealbrowser ;) ). 「ダイノガイ千？！」(Dinoguy1000)22:06, 4 May 2009 (UTC)
Thanks for the reply. I thought my question might have been missed for good. So last night I uninstalled three addons to try and fix some performance problems which have also appeared recently. Those improved straight away and now experimenting while typing this reply I do appear to have fixed the problem.
I wasn't paying proper attention to what I was being asked when I grabbed some trial software until I had already said yes to a couple of "do you want to install this?" questions - I had assumed they would be about the shareware and not other crap. Anyway uninstalling the lot seems to have fixed things. The culprit was one of Relevant Knowledge, Smartshopper or a toolbar that wasn't Yahoo or Google. Is it worth noting these candidates for future reference should someone else experience similar difficulties?--Peter cohen (talk) 22:39, 4 May 2009 (UTC)
You can check the FAQ at IE7.com. If you have this problem you need to avoid navigating away from the edit window. I believe this script will prevent loss of form data by forcing all links to open in a new window when accessed from the edit screen. It's been changed quite a bit since the original version (see previous thread) so I can't vouch for it offhand. — CharlotteWebb02:56, 8 May 2009 (UTC)
Please write me a script for "subpages of this page"
This is probably quite simple:
A button in "toolbox" called "Subpages" that leads to
addOnloadHook(function(){addPortletLink("p-tb",wgServer+wgArticlePath.replace("$1","Special:PrefixIndex/"+wgPageName+"/"),"Subpages","t-subpages","See all subpages of this page");});
I've been using something just like this except that it adds a tab at the top instead, and it uses something like:
wgPageName.replace(/_\(.*\)$/, "")
Instead of just wgPageName because I like to cut out disambiguation terms e.g. "_(song)" and "_(dance)" to make it useful in article space. Might cause problems if you're looking for sub-pages of the village pump so it's up to you. — CharlotteWebb03:00, 8 May 2009 (UTC)
Has the navbox class lost the ability to use the collapsible argument? That is, are constructs like:
{| class="navbox collapsible"
|-
<yada yada>
|}
now pointless with respect to the collapsible part? When I use them, the resulting boxes look right other than there is no collapsibility there at all. Has something changed recently with the underlying code? Thanks, Baccyak4H (Yak!) 18:23, 6 May 2009 (UTC)
Yes. The show/hide button is added to the first header cell encountered, so if there's no header cell, there's no collapse button. Plus even if you could add a button, clicking it would hide the entire table... Happy‑melon07:42, 7 May 2009 (UTC)
Thanks again. I'll keep that in mind although I suspect a wikisoftware hiccup, as it seemed the titles were always there. Resolved Baccyak4H (Yak!) 20:22, 7 May 2009 (UTC)
Interwiki link problem
I was trying to link to my edit counter on my Userpage using an interwiki link like this. It seems that the interwiki markup converts &, ?, and =. Why is interwiki link rendered broken but the external link is fine? Is this behavior necessary for the interwiki links? —Ost (talk) 21:55, 6 May 2009 (UTC)
Thank you for the explanation. I see that bugzilla:15274 deals with the underscores in google links. Would it make sense to add the other characters into this bug or has the community decided that this is acceptable behavior for interwiki links? —Ost (talk) 19:42, 7 May 2009 (UTC)
Yeah, it sanitizes the title just as it would that of a wiki page. You'd be better off just using a template for this. One probably already exists. That way you can update the url as needed, and by that I mean at least a dozen formerly popular edit-counter tools have for some reason quit working or become obsolete/next-to-useless. — CharlotteWebb15:43, 7 May 2009 (UTC)
I'm trying to improve the appearance of this article: 2009 NCAA Women's Division I Basketball Tournament. The site had a list of the automatic bids (in Qualifying teams section.) I created a map of the automatic bids to correspond to the teams. I also created a list of the At-large teams, and a map corresponding to them. However, placing the map as "float right" leaves the maps poorly placed relative to the lists. The first list is a two column list I tried changing it to one column, but that makes it worse, as the map floats further down the page. My current thinking is to convert the list to a table, so I have control over the width, and make it narrow enough so that the map is placed to the right of the list, then do the same for the second list. Is there a better way to ensure that the map of the automatic bids is tot he right of the list, and the map of the at-large teams is to the right of the list? (I wrote this question before noticing the discussion about tables and divs just above - it sounds like I should use tables, but please let me know if there are more elegant solutions.)Sphilbrick (talk) 22:19, 6 May 2009 (UTC)
To prevent the maps from being pushed down, you need to place them at the top of the section. That way, the rest of the section wraps around the map. I've moved them up, hope that helps a bit. — Edokter • Talk • 12:29, 7 May 2009 (UTC)
Text in watchlist but not in edit summary
Four articles on my watchlist (all Iowa counties) have showed up with unusual text. Here's one example of the format as it is on the watchlist, except with no links:
Sorry if this seems like an insulting comment, but I don't understand how anyone thought it would be helpful for these "tags" to suddenly appear on everyone's watchlist without explanation. If I understand the situation correctly, any admin can give themselves rights to edit the abuse filters. Perhaps one would be kind enough to edit the "tags" so that the link back to the filters that added the tag or at the very least to an explanatory page (i.e. [[WP:ABUSEFILTER|no soup for you]])? Delicious carbuncle (talk) 03:58, 7 May 2009 (UTC)
I'll be perfectly honest: yes, that is an insulting comment. It essentially discards the hard work of the numerous developers, many of them volunteers, who were responsible for developing a useful feature of MediaWiki, purely because the changes was not rolled out in a manner you thought was appropriate. I'm sure many people would agree, myself included, that this improvement should have been better documented for maximum usability. I don't think many people would take your view that you would rather avoid a few days confusion by not having new features at all. The developers are writing software for 800 WMF wikis and thousands of independent ones; MediaWiki is not Wikipedia's plaything. Werdna's comment above was more than enwiki deserved. Happy‑melon07:52, 7 May 2009 (UTC)
Hang on a minute there! This isn't about Mediawiki development and please don't try to turn it into that. My comment in no way relates to the potential usefulness of this feature, nor to the fine work done by developers. The confusion this has generated here on enwiki was entirely avoidable. Merely having the capability to tag edits doesn't mean that using that capability should precede informing the thousands of enwiki users who will be seeing them in their watchlists. Unless I'm missing something here -- and I might be -- interested editors can look at the filter hits (eg filter 3) or the tagged edits (eg "blanking") and see the same information, so not using tags at enwiki wouldn't actually delay anything. Did I miss a sitenotice saying that "you may notice new things in your watchlist"? Why can't the tags be linked to an explanatory page? These are each simple steps that seem to have been missed. Why is communication so lacking on a change which affects all users? This is meant to a collaborative project. Delicious carbuncle (talk) 14:58, 7 May 2009 (UTC)
Well lucky thing there's some protection against this. Tell ya what does make my head hurt, the way self-transclusion is okay on, err… self-documenting templates—you know (that I know that you know), the ones where the template transcludes the doc-page which transcludes the template which… never mind! — CharlotteWebb15:18, 7 May 2009 (UTC)
Not seeing system messages
Resolved
A lot of system messages do not appear for me. For example, MediaWiki:Tags-intro was edited (two threads above) but I still see the default message.
This is not a cache issue before anyone asks. I can see these messages on my alternate account, but not on my main account. I've cleared all the gadgets and scripts in case they were interfering. Is there another option somewhere that I've not noticed? Thanks, — Martin (MSGJ · talk) 08:02, 7 May 2009 (UTC)
If we go to any file page that is linked to a Commons image, like File:Stellar aberration.JPG for example, one of the menus on the top of the page shows "Create this page". I understand that Commons images are not editable on en.wikipedia, hence we don't have the "edit this page" menu when we view it here. But "Create this page" is totally confusing when we're already seeing the page. Is there any meta article which has got a clarification? Jay (talk) 08:25, 7 May 2009 (UTC)
"edit" on file/image pages means editing the file or image description page, not editing the file/image. For commons images, one could still create a local image description page. -- User:Docu
I was looking at the usability aspect of it. For this specific case I would consider the "Create this page" menu a usability bug, since the user is already "seeing" the page. Further, on clicking the menu, it takes to a page which says "Please do not manually create this page." and I don't know what that could mean. Also the page title now says "Editing File:xxx". It's all confusing. Jay (talk) 11:20, 7 May 2009 (UTC)
You could edit the mediawiki message that says "create this page" to blank it for that namespace. -- User:Docu
Exactly, this is what I was attempting to do! I added categories to Wikipedia images without a problem, but hit a block with the Commons images. "Create this page" is not intuitive when you want to add an already existing page (that's what a user sees) to a category. Moreover there is this warning of "Please do not manually create this page." The warning brings up the doubt - as opposed to creating the page in what other way ? Jay (talk) 13:19, 7 May 2009 (UTC)
I think for a commons image, the "edit" button should probably be hacked to use a url which edits the commons image (and the other tabs should behave similarly). Of course anything short of disabling page creation in namespace six without upload will be inherently sloppy. Supposing there is some compelling reason to create a local non-image image page corresponding to a commons image, there is no local notification when the latter is deleted, no way to get a list of local pages orphaned in this fashion, no red link in local talk pages where the now-deleted image was once discussed, etc. Also if I wanted to figure out what percentage of locally uploaded images are "fair use" or what percentage are free and should be moved to commons, discrepancies like this will royally scupper up my numbers.
Of course I've often wondered why so many people are hesitant about making commons-related edits on that site rather than here (well I do have a few theories I won't get into right now).
I know categorization of images is a widely neglected task, and I can't fault anyone for being interested in doing it but I'd offer the following rules of thumb based on the license status of the images in the category you are creating: (a) If they are all "fair use", just ignore my advice. (b) If some of them are "fair use" and some of them are free images on the commons, create a category here and a category there (both with the same name) and link between them with something like {{commonscat}}. (c) If some of them are free images despite being uploaded locally, move them to commons (so that they can be shared with other projects) and go back to "b". (d) If they are all on commons but you want a category to exist locally (just in case "fair use" images might be added in the future) I personally think a visibly empty category with a link to commons (using {{commonscat}} or even {{softredirect}}) would be more useful than a category chock-full of empty local image pages, ones which contain nothing but a link back to the category (and which as far as the software is concerned are not associated with any particular image except by coincidence). — CharlotteWebb15:09, 7 May 2009 (UTC)
I understand the problem you mentioned about creating local categories for the virtual Commons pages. But is it allowed today to create the local categories? If so what is happening to the local (description?) page when the Commons image is deleted?
In your solution (d), what do you mean by a visibly empty category? Can you give an example of such a category. Jay (talk) 18:17, 7 May 2009 (UTC)
A1: I'm sure it's allowed but I think it's a bad idea for the predescribed reasons (and I consider the software flawed for smiling and nodding when an description page has no image, or vice versa in a few stranger cases). A2: Probably nothing until somebody navigates through the category by chance, clicks on the file-name and discovers that no image exists—this could take years. A3: In this context I meant a category where the lack of local images belonging to it is obvious. A4: (If you mean one that is kept perpetually empty with a link to commons) no, none offhand but it might be worth considering in certain cases. — CharlotteWebb19:20, 7 May 2009 (UTC)
How about adding a link to edit the commons description at MediaWiki:Sharedupload? It should make clear in one way or the other that one starts editing there. -- User:Docu
Hi I made this revert using Twinle it appears to have removed what I wanted, but when I look at the article page the text is still there, but if I go to edit it the text is not there. I have refreshed the page a couple of times but it still appears. Any ideas what is happening. BigDuncTalk21:48, 7 May 2009 (UTC)
It is sorted now but I had pressed F5, I use Opera as my browser and have it set to always check if cached paged is updated on the server, maybe going a bit slow thanks for the reply. BigDuncTalk22:05, 7 May 2009 (UTC)
Review of user script
Would anyone like to give feedback on User:M/monobook.js? You can import the script in the usual ways. It uses jQuery and a couple of other libraries. Part of it simply cleans up the main edit page. The more interesting part is a 'shortcut editor'. When you highlight a phrase of text (sometimes works with links), an alert box will pop up asking you to rephrase what you selected. Cancel abandons, ok saves your new wording to the page. An edit summary is automatically generated, you can look at some of my contribs for examples. (e.g. ', in total about' → ' totalling aproximately '). Invalid highlighted text will pop up a debug alert, as will success upon hitting ok. These can be commented out. The script reloads the page when done, but this can be commented out also. Suggestions and patches are appreciated. If something doesn't make sense, I can comment it. –MT23:31, 7 May 2009 (UTC)
Banner
You know the banner displayed in the Wikipedia pages for users? The one with the greenish-blue lettering and a "[help us with translations!]" on the side? I accidently dismised it, and can't find how to brng it back...ResMar21:45, 5 May 2009 (UTC)
Try clearing your cookies. If you're not sure how to do that, please say which browser you're using. Tra(Talk)22:12, 5 May 2009 (UTC)
If you get this problem in future, you can clear your cookies by going to Tools > Delete Browsing History > Delete cookies > Yes > Close. Of course, this won't help you now as the vote has ended. Tra(Talk)19:46, 8 May 2009 (UTC)
Hi. Iam having difficulties changing the Military history WikiProject's template on the following page: Talk:Angolan War of Independence. I did change the class variable to C, but it keep saying the class is Start. Iam experiencing this problem on other pages also. Please answer me on my talk page. THANK YOU!!! Slapsnot (talk) 15:37, 6 May 2009 (UTC)
Gadget850 is correct: {{WPMILHIST}} does not accept the C-Class rating. The broader reason is that the WikiProject Military history did not adopt the C-Class rating when it was introduced, and the outcome of a !vote on its adoption in March 2009 was "no consensus". — Bellhalla (talk) 16:08, 6 May 2009 (UTC)
I have trouble reading Wikipedia's very small font size and my browser does not make it any bigger when I zoom in. Is there a way to use my custom css page to make the font larger? Thank you! --Andrew Kelly (talk) 22:28, 7 May 2009 (UTC)
That doesn't seem to be true as I happen to have mine set to #bodyContent { font-size:125%; } and it does enlarge the text. I haven't measured it but I know it's noticeably smaller when I log out so I'm guessing what I see is 125% × 127% = 158.75% of the default size. — CharlotteWebb14:03, 8 May 2009 (UTC)
It is true, the default on #globalWrapper is already 127%, so changing that font-size to anything smaller than 127% will not increase in growth. But, by changing #bodyContent, you're increasing just the font size of that element rather than overwriting the size declaration of the global div, so it does multiply, but only inside that div (and not the whole page, as the original suggestion will do). CSS cascades, but not for multiple percent declarations to the same element. In that case it simply overwrites. --Splarka (rant) 16:22, 8 May 2009 (UTC)
I see what you mean. I must have thought you meant the globalWrapper was inheriting the 127% from somewhere else, I don't know, the <body> tag maybe. — CharlotteWebb17:20, 8 May 2009 (UTC)
Filter certain users from Special:Contributions/newbies?
Is there a way to filter certain users out of the display for [5]? It's a great tool for detecting vandalism and spam from new users but the amount of newbie vandal-fighters' contributions is making it less useful than it could be. Kimchi.sg (talk) 02:37, 8 May 2009 (UTC)
Well, one might argue that the latter group should also be viewed with suspicion. There's no easy way to filter them out. I guess you could create a javascript that hides/removes lines from the page based on whether they match a certain white-list of usernames and/or edit summaries but you'd have to create and maintain this list. And of course if you are viewing 50 edits and hide 12 of them you'll only see 38 rather than an extra 12 from the next page, etc. I'm sure anything like this could be done more cleanly outside the "human" interface sent to normal web browsers. Your best bet might be to see if Huggle, Twinkle, AWB, or similar video-game software has a feature like this, or otherwise request it. — CharlotteWebb14:23, 8 May 2009 (UTC)
etc. but there is no sign of these codes in the wikicode. Can other people see the mess as well, and what is wrong? Cheers, — Martin (MSGJ · talk) 15:21, 8 May 2009 (UTC)
I was writing a short application for a couple Wikipedian IRC users (it will be an RC notice script to keep users quickly informed of edits to their watched pages), but I have a problem in the test bot; it can't write data to the socket. If anyone would like to, I'd be very grateful if you could give us some hints as to what's wrong with the code. I can't apparently I'm "calling a property of a non-object" with the command fwrite( $YamaBot->manager->bot[$yid]['socket'] , $msg . "\r\n" ); I even tried a global variable. I feel guilty for making a few users wait so long as well, so if you could help me. --Yamakiri01:16, 9 May 2009 (UTC)
Hisense Kelon
Is the visible text of Hisense Kelon manipulated on purpose by sysops to prevent legal difficulties or was it manipulated by somebody else? Please compare the visible text and the text you see when you edit. Obviously the change was made in the interest of Gu Chujun. --Cethegus (talk) 05:33, 9 May 2009 (UTC)
Cosmetic changes (Wikitext cleanup options of pywikipediabot)
On my talk page someone disagrees about some specific issues involved in the cosmetic changes-script (which is applied when my bot adds interwiki's). Because I am mostly active on nl.wikipedia, I am not sure if there is any policy on these topics on en.wikipedia. Can anyone help me out on this issue? Thanks in advance. Kind regards, --Maurits (talk) 21:59, 2 May 2009 (UTC)
Just speaking personally, I'm not a fan of extending "WP:" to "Wikipedia:", editors usually choose one or the other for a reason, even in the piped portion. I'm not sure what our current mind on changing Image: to File: when making other edits is. –xenotalk22:07, 2 May 2009 (UTC)
In recent days, I examined several changes the script can do, I think the following settings are more or less safe to use on en.wikipedia and could probably be recommended for interwiki bots as default settings for English wikipedia:
text=self.fixSelfInterwiki(text)text=self.standardizeInterwiki(text)#text = self.standardizeCategories(text)text=self.cleanUpLinks(text)text=self.standardizeheadernames(text)# not in scripttext=self.removeunicodecontrols(text)# not in script#text = self.cleanUpSectionHeaders(text)# #### text = self.putSpacesInLists(text) # already de-activated by defaulttext=self.translateAndCapitalizeNamespaces(text)#text = self.removeDeprecatedTemplates(text)text=self.resolveHtmlEntities(text)#text = self.validXhtml(text)#text = self.removeUselessSpaces(text)#text = self.removeNonBreakingSpaceBeforePercent(text)#try:# text = isbn.hyphenateIsbnNumbers(text)#except isbn.InvalidIsbnException, error:# pass
(updated 01:01, 3 May 2009 (UTC))
It would be nice if there was way to let it apply the general fixes from AWB (see Wikipedia talk:AutoWikiBrowser#AWB general fixes module for python). I don't think "translateAndCapitalizeNamespaces" should be a problem in article namespace, as the only one it's likely to be applied is "Category:". -- User:Docu
Given that sample, the option is probably more useful for languages with localized namespace names. It should probably be turned off too. -- User:Docu
The changes made are minor and, I agree, for some part quite irrelevant. However, I only apply them to pages when interwiki's are changes. The main problem, I think, was on the file: vs. image: question. From the discussion on Wikipedia:ANI#File: vs. Image: I conclude that Image: has indeed been deprecated by File: and therefore, the changes were correct.
For some part, I like the "safe version" proposed bij Docu. However, the hyphenation of IBSN-numbers -for example- are useful in my opinion, and that goes fore most options. Because the bot works on various languages, some changes might be less relevant on the en.wikipedia indeed. Switching them off would however switch them off for all languages. Note that the qualification is "more useful" against "less useful" and not "not useful" or something worse.
Thanks for your reactions. Kind regards, --Maurits (talk) 15:16, 3 May 2009 (UTC)
For the namespace part, I think when it was written, shortcuts like "WP:" were redirects and not a $wgNamespaceAliases. As for the file/image question, I rather not enter it. Ideally one could just use just the "Capitalize" part. As for the ISBN settings, I'd have to look into it in detail.
It shouldn't be too complicated to modify the script that it doesn't apply all settings to all languages (e.g. there is already some code for nl and de). -- User:Docu
I'll check out the possibilities of "nationalization" (although we should be careful in applying it), indeed I noticed some nl-code at the bottom of cosmetic_changes.py. I excluded 'wp' and 'wikipedia' from translateAndCapitalizeNamespaces to solve the actual issue. Kind regards, --Maurits (talk) 17:14, 3 May 2009 (UTC)
My solution didn't work, perhaps someone with better Python skills can help me out. There is already a line that excludes main namespace:
# skip main (article) namespace
if thisNs and namespaces:
text = wikipedia.replaceExcept(text, r'\[\[\s*(' + '|'.join(namespaces) + ') *:(?P<nameAndLabel>.*?)\]\]', r'[[' + thisNs + ':\g<nameAndLabel>]]', exceptions)
(the whole is like this):
def translateAndCapitalizeNamespaces(self, text):
"""
Makes sure that localized namespace names are used.
"""
family = self.site.family
# wiki links aren't parsed here.
exceptions = ['nowiki', 'comment', 'math', 'pre']
for nsNumber in family.namespaces:
if not family.isDefinedNSLanguage(nsNumber, self.site.lang):
# Skip undefined namespaces
continue
namespaces = list(family.namespace(self.site.lang, nsNumber, all = True))
thisNs = namespaces.pop(0)
# skip main (article) namespace
if thisNs and namespaces:
text = wikipedia.replaceExcept(text, r'\[\[\s*(' + '|'.join(namespaces) + ') *:(?P<nameAndLabel>.*?)\]\]', r'[[' + thisNs + ':\g<nameAndLabel>]]', exceptions)
return text
Probably a variant of the main namespace exclusion is needed to exclude wikipedia namespace. I tried to list it as an exception, but the script didn't accept that. Thanks in advance, --Maurits (talk) 17:31, 3 May 2009 (UTC)
You could change the line above to if thisNs and namespaces and thisNS != "Wikipedia": in the sample above, but then you still have to deal with the "WT:" pseudo-namespace. At "Wikipedia:Namespace#Case_insensitivity", I listed a series of titles for the same pages. You can use it to test which ones get normalized to "Wikipedia" ('WP', 'Project', 'wikipedia', 'wP', 'project') and which ones don't. As this isn't ideal, I'd skip the function. -- User:Docu
Thanks for your advice. As there shouldn't be any links to talk pages in articles, I think that this solution will do. If it doesn't, I will skip it when appropriate. Kind regards, --Maurits (talk) 21:39, 3 May 2009 (UTC)
In the same edit,[6] cosmetic changes.py also changed Image: to File: in three locations. Using Image: vs. using File: for images is totally personal preference, although I would strongly recommend it, as it emphasizes that it is an image. There are very few situations where File: needs to be used instead of Image:. One is the stats tool http://stats.grok.se/ According to the release notes for MediaWiki 1.14.0, "Image namespace and accompanying talk namespace renamed to File. For backward compatibility purposes, Image still works. External tools may need to be updated." According to Wikipedia:Images, "The "File:" prefix may be used interchangeably with "Image:":", so as I see it is is totally personal preference, and should not be changed by a bot, or just to change it. 199.125.109.77 (talk) 01:41, 4 May 2009 (UTC)
Problematic edits
Without trying to read python, these are the edits which in my opinion are problematic. There may be many more. Please add to or comment on each. In general, en:w is not a dictatorship, and unless it is explicitly in a guideline, I do not see that a bot should be making "cosmetic changes". 199.125.109.77 (talk) 16:19, 4 May 2009 (UTC)
Agree, the current function (written before we had WP: and WT:) doesn't function as it should. -- User:Docu
Adding a space next to the header = sign
"Example: ==Section title== becomes == Section title ==
NOTE: This space is recommended in the syntax help on the English and
German Wikipedia. It might be that it is not wanted on other wikis.
If there are any complaints, please file a bug report."[7]
And just how does one file a "bug report"? I would amend the above to "This space is recommended on the German Wikipedia only." There is nothing that I am aware of that recommends it on the English WP. WP:MOSHEAD States "Spaces between the == and the heading text are optional (==H2== versus == H2 ==).", meaning that a bot should not be changing this, although I would have no objection to changing unmatched spacings, such as ==H2 ==. 199.125.109.77 (talk) 15:11, 4 May 2009 (UTC)
I think it's an advantage adding this space, I just wouldn't add it to the recommended settings. -- User:Docu
Agree that if an editor likes doing them this way it is fine, but a bot? No thanks. 199.125.109.77 (talk)
It appears this script will no longer run on en.wikipedia (see bot talk). However, while spacing etc is trivial, it is an issue that some editors care about. Therefore, it is totally inappropriate for a script to impose changes unless supported by policy. For example, as I understand policy, no script should insert or remove spaces in ==Headings== because it's bound to irritate some careful editors. Johnuniq (talk) 09:07, 9 May 2009 (UTC)
Changing <br> to <br />
This is wiki markup, not rendered HTML, and if the software wants to pedantically render <br> as <br /> have at it, but <br> is easier to write and more readable, and does not get split up onto two lines, so I see it simply as personal preference and not something that needs to be corrected by a bot. 199.125.109.77 (talk) 16:19, 4 May 2009 (UTC)
I agree. -- User:Docu
The Wikipedia display is rendered as XHTML (look at the source for any page). <br /> is the proper output. Wikipedia runs all of the output through HTML Tidy, which will, among other things, convert HTML breaks to XHTML breaks. There is no real difference in which you use, as they are all tidied. ---— Gadget850 (Ed)talk17:32, 5 May 2009 (UTC)
Outside the scope of this question, but Wikipedia output is HTML rather than XHTML (look at the HTTP headers for any page). The fact that it sends rather oddly formed HTML with an XHTML doctype declaration and PTAGCs that look like ETAGOs is of course our old friend Appendix C rearing its familiar, but still ugly, head again. 8-) It's documented under the W3C's XHTML specification, but if you send it out under text/html, then that's HTML not XHTML and you're a brave chap to try and parse it with an assumption of XML. Why MediaWiki uses Appendix C rather than HTML 4.01 is a (long-standing) mystery to me.
As to 199.125.109.77's original question though, this is wikitext markup, not rendered HTML. Either is equally correct (they have to be, we'd break very badly otherwise) and so it doesn't, shouldn't and mustn't matter. In that case, why force <br />? Andy Dingley (talk) 17:57, 5 May 2009 (UTC)
Purposely or accidentally using broken code because Tidy fixes it just makes it harder and adds more dependencies for anyone who isn't us to reuse our content. One of the more frequent complaints in the #mediawiki IRC channel is from people trying to use Wikipedia content in their own wikis (primarily infoboxes and other templates) and having them be horribly broken because they don't have Tidy installed. Using <br> over <br /> saves 2 keystrokes on something that rarely needs to be used in normal wikitext, so I really don't buy the "easier to type" argument (and the "/" is generally right next to the ">" on a qwerty keyboard; its not like you have to move your fingers all over the place). Mr.Z-man04:16, 6 May 2009 (UTC)
Also, for some of the reasons we don't use an XHTML Content-type, see mw:XHTML, there's probably some more extensive discussions in the wikitech mailing list archives. Mr.Z-man04:25, 6 May 2009 (UTC)
Why we can't use XHTML as a content-type is obvious to any clueful web developer (and it's IE again). The question is why the impossibility of generally publishing XHTML on the web still causes developers to fake it with Appendix C rather than staying with HTML. In particular, why the general clueful MW people fall for this. Andy Dingley (talk) 14:17, 7 May 2009 (UTC)
Changing Image: to File:
If this is an image it should stay as Image, if it is an audio file or video file, feel free to change it, but how would a bot know which it is? 199.125.109.77 (talk) 16:19, 4 May 2009 (UTC)
There isn't an advantage in making this change. It shouldn't be a recommended setting. -- User:Docu
Expanding ISBN numbers with a guessed spacing, such as 0356047113 to 0-356-04711-3
These are mainly clicked on, and it seems completely pointless to change how an editor entered them. In fact, running them together makes cut and pasting them easier, but I would not object to any editor who preferred using them with the dashes. 199.125.109.77 (talk) 16:19, 4 May 2009 (UTC)
The ISBN manual suggests to use hyphens or spaces to make it human readable. As spaces would be a problem, hyphens are clearly the option to choose. -- User:Docu
I think that refers to the barcoding itself, or at least how the ISBN should appear in the book itself. I'm not sure that it is a directive to the rest of the uses of ISBNs. I also note that it states that there are always 5 groups, 3 of which are of variable length and I suspect that refers to the newer 13 digit coding. I'm a little curious how the bot thinks it knows that it should be 1-871082-13-7 or 0-02-871380-X. You will note that book sources mechanically takes out all the dashes (or spaces), taking you to[8] if you click on ISBN 0 02 871380 X, or ISBN 0-02-871380-X. 199.125.109.77 (talk) 17:30, 4 May 2009 (UTC)
Docu is quite right, the ISBN manual recommends hyphens or spaces, with a readability preference for hyphens. The positioning of the hyphens/spacing is not 'guessed', but based on simple rules. Mr Stephen (talk) 20:55, 6 May 2009 (UTC)
If a bot sees a hyphenated ISBN, it should leave it alone. Unhyphenated ISBNs should be left alone as well. Conceivably it might be justified to add proper hyphenation to an ISBN that did not yet have it, but it would be a low-priority task (a cosmetic change) and surely not worth a separate edit. EdJohnston (talk) 03:03, 8 May 2009 (UTC)
pages whose proper rendering depends upon html-tidy
Said Mr. Z-man several sections above:
Purposely or accidentally using broken code because Tidy fixes it just makes it harder and adds more dependencies for anyone who isn't us to reuse our content. One of the more frequent complaints in the #mediawiki IRC channel is from people trying to use Wikipedia content in their own wikis (primarily infoboxes and other templates) and having them be horribly broken because they don't have Tidy installed.
I agree 100%. Would it be easy enough to configure the html-tidy to put wiki-pages with invalid html into a hidden maintenance category (similar to the one which alerts us of ref-tag errors)? Surely it would be better to correct the source (once) rather than correct the final output (at each page-load/edit—of every article using the template and on every wiki they are copied to). — CharlotteWebb17:06, 8 May 2009 (UTC)
An interesting idea. Of course that would be one hell of a cleanup directory. Also, HTMLtidy does a lot of semi-cosmetic stuff, that we wouldn't want to be marked as "broken". (moving pre and postfix whitespace in span elements outside of the span elements for instance.) —TheDJ (talk • contribs) 00:54, 9 May 2009 (UTC)
By that I guess you mean that some of the features of the html-tidy extension are now handled directly by the MediaWiki software, and that you don't mean wiki users have gotten any better at writing html . Still I think the principle of writing good code for its own sake would apply regardless of the exact mechanism by which bad code is repaired. So when I say html-tidy I'm referring more broadly to any part of the software that makes broken code less broken. — CharlotteWebb22:48, 9 May 2009 (UTC)
Yes, the non-cosmetic changes are what I was referring to—cases where it has to add, remove, escape, or change the order of tags on the basis that the previous code was clearly no-good. I don't know how well html-tidy differentiates between these steps and ones which only change white-space, but it should. — CharlotteWebb12:28, 9 May 2009 (UTC)
What links here Page
For the What links here pages, you get filters, but one thing is missing: Show only redirects, Show only transclusions, and Show only links. Please post this to bugzilla; I don't have an account, thanks.68.148.149.184 (talk) 09:43, 9 May 2009 (UTC)
Looks like htmlTidy assumes you can't possibly want to put list elements inside a span, which seems reasonable enough. You'll want to use a div tag instead. What's this for? — CharlotteWebb02:54, 8 May 2009 (UTC)
abc
def
123
456
Yes, this is tidy being tidy. Try <span class="foo"><pre>bar</pre></span> and you'll get <pre><span class="foo">bar</span></pre>. --Splarka (rant) 07:51, 8 May 2009 (UTC)
A question: Is it possible to filter the recent changes list to display changes by a particular user, or more specifically, IP addresses beginning in a certain number? I have a dynamic IP-vandal who keeps inserting false information into football-related articles, so I'd like to be able to bring up the recent changes made by IPs beginning 91.109. I has a little play with related changes, but couldn't make it do what I wanted. Any ideas? – Toon(talk)15:40, 9 May 2009 (UTC)
There isn't a way to do that on recent changes, but there is a gadget that allows you to search for contributions based on an IP range. Enable the "Allow /16 and /24 – /32 CIDR ranges on Special:Contributions forms" under the User Interface Gadgets section of your preferences. Nakon16:05, 9 May 2009 (UTC)
OK, this brigns up another issue: is there a way to set a date to display contibs after? E.g. display contributions for May 2009? The field at the moment is limited to year (and earlier) and month (and earlier), which brings up a lot of contribs for 91.109.*. – Toon(talk)16:32, 9 May 2009 (UTC)
I would just view the page without a date filter and then ignore anything after X date. Nakon16:34, 9 May 2009 (UTC)
Yeah, but the problem here is that I get a list with hundreds of IPs, most of which wouldn't appear if there was a filtered date as they haven't edited in May 2009. Also the IPs aren't sorted by the date of their contribs, but numerically by IP. This means that not only does my browser wheeze for a while, but I have to go through each IP looking for ones which have edited recently. – Toon(talk)16:40, 9 May 2009 (UTC)
The limitation is in what the API is able to output. It is horribly inefficient to query multiple usernames or a prefixindex of users sorted by date, so it returns the data sorted by user and then sorted by date. I could have had the gadget sort by date after in-browser, but this would make your browser quite a helluvalot more prone to wheeze.
What could be done is to give a very narrow date range, eg: &ucend=20090506235959&ucstart=20090509235959 (3 days). This is beyond the scope of the gadget as it (for simplicity) just uses the existing User Contribs UI, but you could fork the gadget to support this as a personal userscript. It would be easy to just have two input boxes for timestamps as those above, though a more complex one could have interactive date selectors (ew). --Splarka (rant) 07:43, 10 May 2009 (UTC)
That sounds useful... but I have no knowledge of writing scripts or code or such sorcery. I am limited to simply appreciating and wishing for such things, unfortunately! – Toon(talk)22:39, 10 May 2009 (UTC)
Drop-down menus
I know that there is the Add LI menu script, but I'd like to take it a step further (which I've tried but failed to do): Add tree-like maneuverability, where one can hover over an item on the list and another list of items appears out to the side, similar in function to this. Plenty of other Web sites use this configuration, but I can never disentangle the code and make something of my own. So can a friendly JavaScript coder offer me a hand? —Animum (talk) 01:44, 10 May 2009 (UTC)
Have you seen the "Add page and user options to drop-down menus on the toolbar" gadget? It takes some of the interface options and moves them to drop-down menus on the tab bar, including sub-menus as you've described. The code may help with what you want to do. (The creator's page is at User:Haza-w/Drop-down menus. --Ckatzchatspy09:51, 10 May 2009 (UTC)
This is also being worked on. See blogpost and MediaWiki_talk:Common.js#Big_changes_coming_to_JS_core. Recently JS/CSS compression was enabled, bringing the size of wikibits.js from 30KB to 9 KB. The upcoming scriptloader system will minify the CSS and JS, further reducing the size of the downloads.
Also, lets not forget that normally, all JS and CSS is cached in the browser, and is in shared use for all our 16.7 million pages that we have on en.wikipedia, thus significantly reducing total traffic and the size of our pages in general. The idea is definitely, to keep our HTML as simple and as it can be, with as much accessibility as possible, and concentrating any CSS and JS into specific centralized pages that can be cached and well maintained. Also scripting is now starting to become more and more conditional, with editpage and watchlist JS being loaded only on those specific pages. —TheDJ (talk • contribs) 12:21, 10 May 2009 (UTC)
169:1 is a little misleading. That's for a page with no content. Except for a few special pages, the size of the JS/CSS is going to be a constant for all pages, so the ratio will not stay the same. Mr.Z-man19:46, 10 May 2009 (UTC)
"The upcoming scriptloader system will minify the CSS and JS" -> The developers have repeatedly stated that skin/site/user JS will not be minified: bugzilla:12250#c2bugzilla:12250#c6. In that techblog post, Brion refers to the jQuery being minified as it is foreign code and is supplied that way already. This was clarified in irc. Note that minifcation removing all linefeeds affects usability and even drastically slows js parsing in some browsers. --Splarka (rant) 19:57, 10 May 2009 (UTC)
I was trying to set a template to suppress a redirect by automatically adding "redirect=no" after the URL. I know it doesn't redirect with the RfD tag on it, but after it is closed as "keep", the redirect creates confusion. See this edit. It was reverted because I apparently didn't code it correctly. I am requesting that someone remake the edit, however, in a way where it works correctly. Hopefully, what I was intending to to is obvious from looking at the code. -- IRP☎ 18:02, 10 May 2009 (UTC), modified -- IRP☎18:07, 10 May 2009 (UTC)
Personally I think I putting a mess of templates and parser functions inside the section heading creates more confusion. I suggest you just use something like {{noredirect}} and put it on the next line, outside the section heading. — CharlotteWebb18:35, 10 May 2009 (UTC)
This may be the first time I had any success with the bugzilla search engine, but I was able to find this bug 4281 from several years ago. It has to do with being able to start an ordered list from a number other than "1" using pure wiki-text.
This seems to be stalemated on the basis that there are two ways to do this, but the start/value attributes are "deprecated by W3C" and there is some arcane way to do it with css but it isn't accepted by most browsers (as of 2005). Not sure whether either of these things have changed since then, but assuming they haven't, which of them would we rather ignore? — CharlotteWebb18:23, 10 May 2009 (UTC)
Okay, so I guess we'd want to decide what the best syntax would be. Perhaps we can make the value tag work like this:
# value="5" | Five
# Six
# Seven
# value="20" | Twenty
That would at least be similar to what's used for wiki-tables. Not sure what would be best for the "ol" element because there is nothing in the wiki-text directly corresponding to it. Maybe we can make it optional to explicitly use it as a wrapper for the wikitext list (?!):
<ol start="5">
#Five
#Six
</ol>
Of course if we're going clearly define where the list begins and ends, we no longer have to rely on the lack of blank lines. For example this should be ok:
<ol start="5">
#Five
#Six
#Seven
#Eight (not 1 or 5 mind you)
#Nine
</ol>
This usually happens more with unordered "bullet" lists and I fix them because having several lists with one item each is bad html, but people usually revert it because they like having more space and they don't understand the difference. If we could let something like this render as one continuous list I wouldn't have to fight them so much:
== References ==
<ul>
*{{cite book | last = Doe | etc. etc. }}
*{{cite web | url = http://etc.etc/etc.etc?etc=etc | etc.}}
*{{cite news | work = USA Today | etc. }}
</ul>
Increasing the density of HTML in wikimarkup is a Bad Idea in general; it shouldn't be made the norm. If the first example you suggest was implemented, there wouldn't be any need to use the explicit ol tag, would there? Happy‑melon19:16, 10 May 2009 (UTC)
No, only if somebody is looking for a place to put the "start" attribute (I realize that might not be needed if you can just put a "value" attribute on the first item, however that might not always be feasible if all this is assembled from separate pieces i.e. complex templates). Can you think of a better way? — CharlotteWebb19:27, 10 May 2009 (UTC)
I'm not saying the extra tag should be the norm, only that it might be a good idea to make it optional as something would need to exist in order to apply attributes to the list as a whole rather than to single lines. — CharlotteWebb19:29, 10 May 2009 (UTC)
Firefox crashes while searching for a specific word
I'm using Ubuntu 9.04, Firefox 3.0.10. If I enter 'trans' into the search bar to the left and click "Search" Firefox crashes, but doesn't if I click "Go". Can anyone else replicate this? ~ Ameliorate!03:02, 11 May 2009 (UTC)
Thanks, but those two pages don't seem to explain why and where the green type is used. Or maybe I missed it. Yours in puzzlement, GeorgeLouis (talk) 19:51, 9 May 2009 (UTC)
It seems someone decided that green was a good indicator of quotations. Feel free to make it more normal, or indeed visible. OrangeDog (talk • edits) 17:03, 11 May 2009 (UTC)
Changing map in location template
Gotland
The article Romakloster, on a small settlement on Gotland, looks terrible. It started out with only the national map, which takes up a lot of space on the page and still does not show the location of this small place with any precision. The red "pinhead" covers a territory equivalent to a medium-sized metropolitan area to mark the position of a place with a few hundred people. I changed it to the Gotland map, and another user restored the Sweden map along with the Gotland map. So this short article on a tiny place now has two huge maps to show where it is.
I asked Erik Frohne over at the German Wikipedia, who has uploaded the location maps for Swedish counties, if one could make a modified location map that has a small map of Sweden in the corner, thus avoiding the perceived need for a large map of the country in the infobox. He made one (seen here to the right).
The problem is that still I can't figure out how to make this work with the Template:Location map Sweden Gotland and the infobox in the article. Changing the name of the image file in the template code to point to the modified image (the one seen here) seems to make no difference; it still shows the other image. Why? Where is one supposed to make the needed changes to get it into the article page? --Hegvald (talk) 20:39, 10 May 2009 (UTC)
You should see the change once saved (and purged). Note you will not see the change in the preview as it uses {{Location map/Info}} to display the template itself. I also have to say that this image uses way too much detail for the corner image, which are subsequently lost in the infobox; the SVG spans 2.8 MB against the original 63 KB, I wouldn't recommend using it. — Edokter • Talk • 21:23, 10 May 2009 (UTC)
OK, thanks. I only previewed and did not try to save it. I assume the size can be reduced in some way? --Hegvald (talk) 21:48, 10 May 2009 (UTC)
The SVG is 2.8 MB, true, but the rendered PNG image, which is what is actually downloaded, is only 20 KB. The file size of the SVG itself isn't important—that only matters to people who are going to download the SVG source. —Bkell (talk) 05:11, 11 May 2009 (UTC)
I also don't know, what went wrong with your changes. After some testing with your modified template, I made a copy of Romakloster here using the new template. Seems to work fine ... --Erik Frohne (talk) 09:46, 11 May 2009 (UTC)
Funny... when I now "walk" backwards in the history of Template:Location map Sweden Gotland modified, even the earliest version looks good. Apparently, some patience is all that is needed. I purged the page before, but have shut down and restarted my browser completely since then; that may have had some effect. Thanks for all the help! --Hegvald (talk) 10:24, 11 May 2009 (UTC)
Attention: The history of the template looks good, because the current template version is evaluated and inserted by {{Location map/Info}}. Thus any historic version of a location map contains the actual location map. As I told before, I don't know what went wrong and you may discover the same effect on the next template you are creating. Yours, --- Erik Frohne (talk) 12:38, 11 May 2009 (UTC)
This is one that I have been trying, without luck, to figure out for a while. How can one transclude (in this case their User page) to other wiki sites? I have a global account, and I am active on several of the other sites off and on. I would love to just have my Wikipedia User page ported to the other sites so that when someone clicks my name, they can see some info about me. At the moment, I just have redirects that don't auto-redirect on there. Any ideas? Pages in question are User:Tigey and v:User:Tigey. Tigey (talk) 14:53, 12 May 2009 (UTC)
Very well. Not a real problem. I would imagine that it could be a security/bandwidth problem if it were implemented. Thanks, guys. Tigey (talk) 16:50, 12 May 2009 (UTC)
Secure server
I use the secure server (https://secure.wikimedia.org/wikipedia/en/wiki) to access Wikipedia. It is annoying that interwiki links and user-created diff links go to the normal server, effectively logging me out and making SUL useless. Are there any scripts/other such technical solutions that will do URL rewriting to fix this?
I've got around 1400 seconds of lag going. I noticed a considerable slowdown two days ago, with lag notices popping up every other time I checked my watchlist. Maintenance maybe? -- Commdor {Talk}21:49, 11 May 2009 (UTC)
I have over 1500 seconds of server lag. There's been lag for the last several hours but it's got bigger over the course of the day. I suspect it's something like maintenance as well. Acalamari21:57, 11 May 2009 (UTC)
The database server that houses the en.wp watchlists, is being dumped. Nothing to worry about, when the datadump is done, performance will be normal. —TheDJ (talk • contribs) 23:00, 11 May 2009 (UTC)
That's weird. I must be missing something, because download.wikimedia.org says that an en.wiki dump is in progress, but look at the feed:
2009-05-11 22:31:35 enwiki: Dump in progress
2009-05-06 11:14:55 in-progress Extracted page abstracts for Yahoo
2009-05-11 22:31:35: enwiki 14360998 pages (30.387/sec), 14361000 revs (30.387/sec), ETA 2009-05-15 02:46:38 [max 22702415]
abstract.xml
So we have three more days of completely disruptive lags that are frankly bad looking for the project for what exactly? - NeutralHomer • Talk • 14:26, 13 May 2009 (UTC)
At the worst period yesterday it could hardly load the articles (I got several Wikimedia time-out messages). Some kind of general announcement before and during this would have been nice though. I even searched Wikimedia but couldnt find anything. --Saddhiyama (talk) 14:52, 13 May 2009 (UTC)
Lol, whose overnight hours? My overnight hours? Your overnight hours? Australia's overnight hours? The nature of the project makes it unfeasable, and restricting it to a few hours a day would just inconvenience a certain portion of editors for much longer. – Toon(talk)15:08, 13 May 2009 (UTC)
This should be done during overnight hours so it does not disturb editing.
OK, time of least disruption. That would be around 2 to 5 am EST. I have found that few edits are made during those periods. Either way, something needs to be done because at the present point some people are having much trouble editing, edits aren't going through (personal experience included there) and other problems are arising from this server lag....plus it just looks bad on the project. - NeutralHomer • Talk • 15:16, 13 May 2009 (UTC)
That would be "do it at noon in in Europe and Africa, and mid evening in Asia and Australasia, because people outside America don't matter", I take it? – iridescent15:25, 13 May 2009 (UTC)
Never said that. From what I have seen on Recent Changes, the least amount of edits happen during that time period. That's the best I can come up with so that this problem doesn't upset anyone. Someone is going to be upset, but if we can upset just a few, it would better than half shutting down the site with lags. - NeutralHomer • Talk • 15:44, 13 May 2009 (UTC)
I thought so. Maybe we should sacrifice the connectivity of one country so the rest of the world can connect without lag? If you're a RC patroller, you'll know that most of the vandalism takes place during the day (EST). AvN16:31, 13 May 2009 (UTC)
Less editing happens on this project from something like 500-800 UTC. That's a fact. I'm sorry you see it as nationalism. Cool HandLuke18:12, 13 May 2009 (UTC)
The lag level measures the extent to which the slave database servers have fallen behind the master. The s1 cluster, which houses only enwiki, has one database master and five slaves. Most read operations (reading and loading pages, compiling special pages, etc) are done on one of the slaves, which are read-only, while all write actions are done on the master. The slaves are continuously synchronised with the master but, if the rate of write actions to the master is high, they may not be able to keep up. Since your watchlist is loaded from one of the slaves, it may not reflect changes that were made to the master more recently than the slave was synced; hence the need for the message to inform you. But the extent of the database lag shouldn't affect, for instance, the speed of loading pages (although slow loading and high lag may share a common cause).
The dump process is always a high-load operation, which has consistently failed in the past: dumping the 'big three' (enwiki, dewiki, frwiki), enwiki in particular, is a huge operation. There's not a whole lot of point in identifying the least active part of enwiki's day when the dump process routinely runs for a fortnight, when it runs at all.
Brion has said that the dump load on the watchlist server has been removed, so any further lag is probably unrelated. Happy‑melon20:03, 13 May 2009 (UTC)
Number of pages in a category...
...is supposed to be shown in a standard message on the category page ("The following abc pages are in this category, out of xyz total.") However, this "xyz" figure is not always correct, i.e. if you browse the category in question page by page, the total number of entries shown (i.e. 200+200+...+number of entries on the last page) is not equal to xyz. E.g. Category:Mid-importance Croatia articles is stuck with one article missing, and it's been that way for at least a couple of weeks. Is there a way for it to get "unstuck"? Of course, dummy edit to a particular page would do the trick, but how to find that exact page? GregorB (talk) 19:31, 12 May 2009 (UTC)
That's it, thanks! The pages were probably deleted at some point, and while this is mildly annoying, it seems that there is nothing one can do about it at the moment. Hopefully it will get fixed one day. GregorB (talk) 21:34, 12 May 2009 (UTC)
Adjust your template to populate a different category, and then change it back again. This should add the contents to the job queue which will flush them out. That's what I did and it seemed to work. — Martin (MSGJ · talk) 12:00, 13 May 2009 (UTC)
double escaped pipe?
I am working on a template, and I need to know how to do a double-escaped |
The problem is, the second ! was already escaped the first time around, so now I need to double-escape it. Possible? -Zeus-u|c01:07, 13 May 2009 (UTC)
I don't fully understand; that code has unbalanced braces that will screw up most things. Can you give a larger code snip (or just the template you're working on)?? Happy‑melon14:56, 13 May 2009 (UTC)
As you can see, I have no idea what to with that {{!}} present in the first case. I have to convert all of the |s to {{!}} to make the if work, but I don't know what to do with the one that is already {{!}}. -Zeus-u|c18:26, 13 May 2009 (UTC)
If I understand correctly, you want to place an image inside a table, and the image should link somewhere, and the whole thing conditionally. You don't need an imagemap for that, and can do with a fair bit less escaping. The following will work just as well:
When I go to my Watchlist there is a request made to http://toolserver.org/~para/geoip.fcgi that seems to be timing out. It take something like 20 seconds for this to happen. It looks like it might be something for detecting the geographic location of my computer's IP address. Perhaps someone could investigate or turn it off until the problem is fixed? Barrylb (talk) 06:53, 13 May 2009 (UTC)
Works for me. It is not uncommon that some of our services are offline at times. The internet breaks :D (And it indeed determines your location based on IP. Then if you are in a certain zone, you will get invitations for local Wikipedia/MediaWiki events for instance. —TheDJ (talk • contribs) 11:04, 13 May 2009 (UTC)
It does seem to be working now. If there is anything we could do to avoid the timeouts when services are offline it would be nice... frustratingly slow otherwise. Barrylb (talk) 11:44, 13 May 2009 (UTC)
Looking for a value colorizarion template
Hi,
I'm looking for a template that takes a value and outputs a color (green, yellow, red) depending on that value (and other parameters). The idea is to color a table background depending on the number of pages in a category (the kind of stuff User:Dragons flight/Category tracker/Summary did). I'm pretty sure someone created something like that before, but I can't find it. Any clue? -- Luktalk12:31, 13 May 2009 (UTC)
This would be somewhat tricky, but it's possible. The idea would be to use RGB colour values formatted in CSS using the rgb() notation, e.g. "background-colour:rgb(255, 255, 255);" for white. Since you then have decimal values instead of the usual hexadecimal notation, you could use the #expr: ParserFunction to compute the weighted average of a starting and ending colour given a range. I'm a little busy at the moment, so I won't code up an example, but it's certainly possible. Whether it would look pretty is another issue that would probably require significant tweaking with exceptions and multiple ranges. {{Nihiltres|talk|log}} 17:42, 13 May 2009 (UTC)
Piped link propblem
In version of Lorem Ipsum, there was a piped external link to la.wikisource. However, the first word after the pipe was taken to be part of the link, and taken away from the displayed text, resulting in a "Titulus Malus" (bad title). I sorted this out by putting a space after the pipe, which didn't put an extra space in the displayed text. Why did this happen. Should I report this to Bugzilla? Thanks, God Emperor (talk) 15:57, 13 May 2009 (UTC).
In external links the target and the title are seperated by a space, not by a pipe: [http://en.wikipedia.org/w/index.php?title=Lorem_ipsum&oldid=288807706#History_and_discovery This version of Lorem Ipsum] Amalthea16:02, 13 May 2009 (UTC)
Thanks. I just found that out and then went to correct the page to see you'd already done it before coming here. God Emperor (talk) 16:11, 13 May 2009 (UTC)
Applying one special page to another
When I click on "What links here", the "Related changes" tool vanishes from the toolbox. Is it possible to apply "Related changes" to a "What links here" page? Ntsimp (talk) 17:25, 13 May 2009 (UTC)
Following concerns raised in the Usability Study, and the proliferation of books of questionable utility, there's an ongoing discussion at VPR about temporarily hiding the "create a book" interface from the sidebar, until the system can be redesigned to be more user- and site-friendly. All welcome to the discussion and straw poll. Happy‑melon17:27, 13 May 2009 (UTC)
Wikipedia's images are not hosted on the same domain as the text. If they have a firewall with strict security settings, then this might be forbidden by their security settings. If you visit different websites, do you see advertisements on ANY webpage ? Advertisements are often also hosted on different domains, and would thus also not be visible. 82.75.218.92 (talk) 13:56, 12 May 2009 (UTC)
As I said, it's only Wikipedia (and Excite. I checked my email and things look so different. I also can't reply to any emails, but that's a Reference Desk problem). They said they'd get someone to look at the situation, and it has been fixed. I'll have to report back on what they did when the person gets here.Vchimpanzee· talk·contributions·14:09, 13 May 2009 (UTC)
Per this discussion, it seems it would be useful if there were a magic word or something that returns the page the user was redirected from. That way we could write conditional hatnotes ("X redirects here; for other uses...") that would be displayed only if the user came through that particular redirect. Any idea whether this is feasible, or whether it's ever been suggested before? I wouldn't have thought it would be particularly difficult, since a link to the the redirect source is displayed under the title, so the information is obviously being forwarded. --Kotniski (talk) 17:34, 12 May 2009 (UTC)
A similar idea was recently discussed at Proposals and garnered some degree of support. Personally, I like the idea because it allows for removal of the hatnotes that pertain to a redirect instead of the article's title. —Ost (talk) 18:11, 12 May 2009 (UTC)
Thanks for linking the bug and describing the difficulty. How is the "redirected from" note currently inserted onto pages when redirected and could this method be expanded to longer messages? Could the message be incorporated into the redirect syntax instead of on the destination article? The last post on the bug recommends JavaScript; is that a possiblity? —Ost (talk) 19:23, 12 May 2009 (UTC)
That message is added directly by the php that generates the webpage. The parsercache is basically the "evaluated" (HTML) version of the "raw" wikicode you see in the edit window. When you view a mediawiki page, the webserver generates the HTML page, and inserts the "cached" version of the article content from the parsercache. There is virtually complete disconnect between the HTML of the interface and that of the article, so it is difficult to do what you suggest. With Javascript you can do almost anything, but it also adds to the load of a page and it needs a plain HTML "fallback" method, because not everyone has Javascript enabled. —TheDJ (talk • contribs) 19:43, 12 May 2009 (UTC)
Where/how did you get all your insights into the response process? Is there something we should be reading, or are you familiar with the source code? - Pointillist (talk) 21:07, 12 May 2009 (UTC)
Well, these things are described somewhere, it's just that wikipedia/wikimedia is very complicated and hard to summarize. Perhaps you are interested in this diagram created by User:Mr.Z-man. —TheDJ (talk • contribs) 23:59, 12 May 2009 (UTC)
OK, I thought caching might be a problem (not insurmountable, I would have thought - there are plenty of other magic words which have different values per request). But how about solving it from the other direction: have tags on the redirect page which cause transclusion of wikitext to the target page. Something like <include>some-wikitext</include>, which would cause some-wikitext to be HTML-ed and inserted at the appropriate place in the returned page (under the title and before the HTML from that page's wikitext, I suppose). Has that sort of thing ever been considered?--Kotniski (talk) 08:11, 13 May 2009 (UTC)
This was proposed recently; while it's not possible to implement as you describe, we can change the "Redirected from foo" message (MediaWiki:Redirectedfrom) to potentially include other stuff, even transclude templates etc. It's just a question of where to put the text so it can be easily but safely included. Happy‑melon14:53, 13 May 2009 (UTC)
So we could add to that message something like {{#ifexist:$1/RedirectHeader|{{:$1/RedirectHeader}}}}? (Then we would have to fix the CSS so that the small, under-the-title style was applied only to the fixed message and not to the transcluded bit.)--Kotniski (talk) 05:58, 14 May 2009 (UTC)
That would clutter article space with those pseudo-subpages, so it's probably not a good idea. More likely to work well would be something like what is done for per-page editnotices in namespaces without subpages since domas disabled the storage of those in the MediaWiki namespace: Store them as subpages of some appropriate template. Also, it would be user-friendly to use a "loader" template that would wrap the notice in a div and include something like {{vde}}. Anomie⚔11:52, 14 May 2009 (UTC)
In addition, you might get some heat from a dev for a change like that. They didn't like a very simple parser function in MediaWiki:nstab-main, they probably won't like an expensive one on every redirected page. Amalthea12:08, 14 May 2009 (UTC)
<moan>Maybe it would encourage the devs to actually come up with a solution instead of forcing us to use workarounds all the time...</moan> Can it really be that difficult to program something like the <include...</include solution proposed above?--Kotniski (talk) 12:42, 14 May 2009 (UTC)
Well there's always the chance that us doing something crazy might prompt the devs to actually provide appropriate functionality rather than knee-jerk against the hack. We could do something like this:
Then we could put at the top of a particular article itself:
<onlyinclude>{{#switch: {{{1|}}} | Foo = Foo is something | Bar = Bar is something else }}</onlyinclude>
This transcludes the entire contents of the page into the redirect message, but only if it's less than 200 characters. So You can use the <onlyinclude>...</onlyinclude> tags to ensure that only the hatnote is included in the message; and a #switch to select between possible redirects. No expensive parserfunctions, and reasonably safe. And you can put the onlyinclude section anywhere you like on the page, maybe at the bottom would be better than the top. This would probably fail spectacularly with templates, but this is only really useful for articles, which shouldn't ever be transcluded AFAIK. That particular string function is not horribly expensive like most of the padleft: based hacks. Thoughts? Happy‑melon12:54, 14 May 2009 (UTC)
Hehehe. Smooth. I like it. I don't think that transcluding the page + the string hack is actually cheaper than the #ifexist, but it's good that the hatnotes can be placed in the article itself. Of course, having them in the redirect would be even better, but that needs a MediaWiki change I guess. Amalthea13:17, 14 May 2009 (UTC)
String length function
On a side note, why is there no string length function in MW? Presumably it would take about ten minutes of coding for someone who knows what they're doing. Another case where a tiny improvement to the software would eliminate the need for ugly and costly hacks like the "str ≤ len" one that had to be used in the above.--Kotniski (talk) 16:08, 14 May 2009 (UTC)
I do not know if this is a one-shot error b/c I've gotten 2 different errors in the last 10 minutes. This is what I got the 2nd time:
Wikimedia Foundation
Error
Our servers are currently experiencing a technical problem. This is probably temporary and should be fixed soon. Please try again in a few minutes.
You may be able to get further information in the #wikipedia channel on the Freenode IRC network.
The Wikimedia Foundation is a non-profit organisation which hosts some of the most popular sites on the Internet, including Wikipedia. It has a constant need to purchase new hardware. If you would like to help, please donate.
If you report this error to the Wikimedia System Administrators, please include the details below.
The syntax colorer is getting confused by the line if string.find(y,'"') then. The same also happens with lang="javascript" on regular expressions containing single single- or double-quotes as in foo.replace(/'/g, "%27");. Just fix it by adding a comment containing another unmatched quote, for instance if string.find(y,'"') then -- 'Fix syntax coloring HTH, Lupo19:50, 13 May 2009 (UTC)
FWIW, there are several errors in GeSHi. For example, it doesn't recognize javascript regular expressions:
I have a hack to get pass the javascript problem. I had reported it to Bugzilla but it hasn't been fixed yet.
abcd=efg.replace(/\"/g,"")}//This should be green comment color, but it is colored blue, the color of a stringvarcolorRectifier="\""//returns coloring to normal//coloring is normal here
In this code, I am removing quotation marks from some text. As it is in a regxp, I backslash it. The backslash escape isn't recognized, and I have to put another backslashed quote to rectify the coloring.`ManishEarthTalk • Stalk—Preceding undated comment added 15:31, 15 May 2009 (UTC).
I believe this has already been fixed in recent versions of GeSHi, but Wikimedia is using an older version. Mr.Z-man20:51, 13 May 2009 (UTC)
When the browser size is minimized to half of the browser window size, the "sample mxml application code" was going beyond the outline and the page was getting extended. Its not a problem though, except that the window stretches to the right beyond the page width. —Preceding unsigned comment added by 122.183.242.174 (talk) 08:52, 15 May 2009 (UTC)
This is a common and deliberate formatting although some of the source code lines in that particular article may be a little long. Your browser probably has a horizontal scroll bar where you can scroll to the right. PrimeHunter (talk) 10:39, 15 May 2009 (UTC)
You do not HAVE to use the scroll bar. When the text goes out of the box, just highlight it ad drag your mouse to the right. Try it here:
Hidden so we don't have to have a page with insane scrolling for the next 4+ days
This is particularly useful when you are reading a long script and want to see what is after the cutoff without having to go down to the scrollbar and scroll.
This feature was added so that code looks the way it should. For example, in most programming languages, if the cutoff was put on the next line, the meaning of the code could change. ManishEarthTalk • Stalk15:18, 15 May 2009 (UTC)
If I click on "edit top" here, I am directed to Gavin, rather than the infobox/lead. It's late here, and I can see no obvious reason for this. Anyone care to figure it out? Thanks. Rodhullandemu00:32, 16 May 2009 (UTC)
What do you mean? Lubuntu does not exist. Where do you think it is transcluded? Maybe you are confusing a page that is transcluded with a page which is linked from another transcluded page. Lubuntu is linked from Template:Ubuntu which is transcluded on many pages, so Lubuntu is also linked from those pages. PrimeHunter (talk) 14:03, 15 May 2009 (UTC)
Yes, but when looking at "What links here", if I choose "Hide transclusions", shouldn't the articles that include the template disappear from the list? i.e. it should then only show the pages, like this one, that have a specific link to Lubuntu. However, it doesn't work; it still includes all the pages that only transclude the template. - Fayenatic(talk)21:37, 15 May 2009 (UTC)
That's not the meaning of the filter. Pages can either be linked to on a page (Foo), they can be transcluded onto the page ({{foo}}), or they can be redirects to the page. Lubuntu is not transcluded anywhere: there is nowhere on Wikipedia that contains the code {{:Lubuntu}}. There are a few pages that link to the page, so they show up as links, as expected. How the link comes to be on the page is irrelevant. Happy‑melon21:42, 15 May 2009 (UTC)
OK, thanks - I misunderstood. I think it would be useful to be able to exclude pages which only link through a transcluded template. Can I request that feature somewhere? - Fayenatic(talk)16:26, 16 May 2009 (UTC)
Redirect editing
Anyone have a quick way to click on a redirect link and open it for editing without having to go to the redirected page, back up to the redirect and edit? ---— Gadget850 (Ed)talk01:45, 16 May 2009 (UTC)
Even without that you could use a javascript to add &redirect=no to the url of all links with class="mw-redirect". — CharlotteWebb01:49, 17 May 2009 (UTC)
Heya ... for the 2009 Super 14 page the current sports event icon features a football ... is it possible to change this to a rugby ball ...[11] ... which is more relevant ? Cheers :1 Boomshanka (talk) 11:19, 16 May 2009 (UTC)
Best to ask on the article talk page, especially since the current image is noted as the "official logo". Frankly, I see little difference in the two images of the balls. ---— Gadget850 (Ed)talk11:38, 16 May 2009 (UTC)
Looking at the page source, I see that the access key is defined to be a comma. In Firefox 3 and IE 7, the key combination would be Alt+Shift+comma. -- Tcncv (talk) 04:12, 16 May 2009 (UTC)
A database query syntax error has occurred. This may indicate a bug in the software. The last attempted database query was:
(SQL query hidden)
from within function "Revision::insertOn". MySQL returned error "1205: Lock wait timeout exceeded; Try restarting transaction (10.0.6.22)".
I'm trying to add info to a blocked sickpuppet page to assist the unblocking admin here. This doesn't help. Any ideas? Rodhullandemu03:12, 17 May 2009 (UTC)
The message is useful to those who can actually resolve the issue: the MediaWiki developers. I agree it's not particularly useful to normal editors. Happy‑melon10:26, 17 May 2009 (UTC)
Actually the message is more clear than some. A lock in a database prevents other processes from writing to it while a particular process that holds the lock makes a sequence of changes. Your edit was waiting to acquire the lock, but it had to wait too long, so it gave up. This means something is holding the lock too long, or there are too many processes trying to get the lock at the same time. The system admins with access to the database can use the db administration features to diagnose what's going on with the locks, and then try to remedy it. There is nothing that can be done on our end except to let them know that the error is happening. — Carl (CBM · talk) 19:26, 17 May 2009 (UTC)
I'd like to search all non-File: pages for the text "[[Wikipedia:Non-free content|Non-free TV-related media rationale]]". Is there a way to do this? --Carnildo (talk) 23:01, 15 May 2009 (UTC)
Not an easy one. You could use a database dump; the next all pages/current version dump for enwiki has an ETA of about 15 hours from now. Or you can use the API with generator=backlinks, looking for links to Wikipedia:Non-free content in every namespace except for 6. According to a db query, there's 178491 links to it outside of ns:6, so that would be a minimum of 357 queries with a sysop or bot account. The vast majority of the links are in ns:3 and ns:1. Mr.Z-man01:37, 16 May 2009 (UTC)
Thanks. I was hoping to find the template that produces that text, but it looks like it's either generated by a meta-template, or it's off-wiki. --Carnildo (talk) 23:28, 18 May 2009 (UTC)
I believe it is generated by some edited version of Template:Filmr, possibly copied and pasted. I'll look to see if I can find anything more, but it seems unlikely. The best bet would be to as User:Dr. Blofeld if he remembers where it was. — Carl (CBM · talk) 23:33, 18 May 2009 (UTC)
Table of contents: because I do not use the TOC contents within list body as section names, they do not link. What I would like, is to have the "48" in the TOC to link the "48" part of the table (but not to the section "First generation (1948–1956)"). Is there some way of linking this way, and if so, could someone please do one for me so that I can see how it is done?
Width: of the table found under the section "Tenth generation (1997–2007)" is rendered much wider that the other tables when viewed in Safari and Opera and IE 6, but is fine under Firefox 3. From what I can see, the "width" parameters are identical in each table, and I can not work out the problem. Basically, I need to have all the tables and all the cells of each separate table to line up exactly. Thanks in advance. OSX (talk • contributions) 08:05, 17 May 2009 (UTC)
I have one more request on top of these two: it it possible to have a link at the bottom of each section titled "Table of contents" and a link back to the TOC? OSX (talk • contributions) 08:13, 17 May 2009 (UTC)
Cheers. That works well. I moved the "anchor" template to the table heading so that these are included. Any way, many thanks. OSX (talk • contributions) 09:11, 18 May 2009 (UTC)
Usability Study
Is there some place that we are discussing this? I think it's a big deal, but since it's over at wikimedia there doesn't seem to be a place to discuss it. The only thing I've found so far is this. Thanks. - Peregrine Fisher (talk) (contribs) 00:46, 18 May 2009 (UTC)
The displayed plot spans the entire download, 5.9 seconds. Notice that the load of the page itself takes 0.99 seconds. Halfway through its download (t ~ 0.5 sec) there begins a spurt of parallel dowloads, apparently CSS files, which end at t ~ 1.5 sec. Then there is a long chain of serial downloads that takes off where MediaWiki:Common.css (?) ends. The chain begins with wikibits.js and ends with MediaWiki:Wikiminiatlas.js, at t ~ 4.0 sec. Finally all the images are dowloaded, almost in parallel; the last one to finish is the wikipedia logo.
Thus the chain that begins with wikibits.js is responsible for about 3.5 seconds of the total 5.9 sec download time.
The useful contents (main text and images) could have been downloaded in about 2.5 sec, or perhaps as little as 2.0 sec if images could be downloaded in parallel with the text.
It looks like one could cut the user-perceived download time by 50% or more just by repackaging the js/css files in that hain info a single file. Is that feasible?
I am using Firefox 3.0.8 under Linux Madriva 1.9.0.8-0.1mdv2009.0. The timings were obtained with the Firebug add-on for Firefox. The page was fetched with shift-reload which (AFAIK) ignores the browser's cache and fetches every component from the net.
By the way, for the last week or so, WP downloads often hang halfway and stay there forever. I still cannot figure out the culprit, but "waiting for upload.wikimedia.ord" often appears in the status bar at those times. Has this been discussed anuwhere?
I can confirm the behaviour as observed by Jorge; it doesn't happen consistently, though. Often it is so bad that it makes editing impossible, and then pages load with normal speed again. -- Michael Bednarek (talk) 06:36, 18 May 2009 (UTC)
Yesterday, it was slow on saving long pages too. I supposed that it might had something to do with the filters. -- User:Docu 08:24, 18 May 2009 (UTC)
This has much to do with the way how Firefox processes Javascript btw. On Safari, everything runs much more in parallel. However change is comming. Actually for a large part this is already done. It's just that the website has not been synced with the sourcecode for over 2 months now (Busy developers). —TheDJ (talk • contribs) 09:46, 18 May 2009 (UTC)
On another project, I did just that with some JavaScript, for sysops only. It is actually quite convenient, so you might try suggesting it at MediaWiki_talk:Sysop.js. Meanwhile, it's been implemented in MediaWiki (mediazilla:18423), but we haven't seen MediaWiki updates live for quite a while. — AlexSm19:08, 18 May 2009 (UTC)
Hiding Template
I have a template. {{Cotton processing flowchart}}. I want to display it on some pages- in full (such Cotton mill and on other pages collapsed (such as Lancashire Loom), using a simple syntax {{t1|Cotton processing |Show}} {{Cotton processing |Hide}}. What is the syntax?
You just need something like class="collapsible {{#ifeq:{{{1|}}}|Hide|collapsed}}" on the table. HOWEVER. this also requires the table uses a header. Atm, your infobox is using a caption (! vs. |+) Also, Infoboxes tables and collapsible tables don't really like eachother. You might consider using a collapsible table in one cell of your infobox table. However, in general, it is a bad idea to hide CONTENT (opposed to hiding meta-/navigationdata which we commonly do). —TheDJ (talk • contribs) 09:10, 18 May 2009 (UTC)
Thanks. It is not actually ever used as an infobox- it just happened to work- so that can be hoiked out. It has served its purpose- proving the concept that this may be useful on multiple pages and giving a home for the .svg graphics in a table. The opposition to hiding content needs thinking about- it could provide an alternative solution. --ClemRutter (talk) 19:50, 18 May 2009 (UTC)
WikiBiff
Ladies and gentlemen,
I submit, for your approval, a new application for all you WikiAddicts out there -- WikiBiff.
You know it's happened to you -- someone left a message for you on your user talk page. Normally, the next browse to another page on Wikipedia would have alerted you to the new message, but you weren't on Wikipedia at the time -- you were off typing up some paper in Word or playing some flash game on the numerous websites that have flash games on them. You finally saw the message, five hours later, and you realized that the person who left you the message went to bed 15 minutes after they left you the message, sorely disappointed that you never answered them.
Now those days are over, because now WikiBiff is here! WikiBiff will sit silently in your system tray, and alert you with a friendly balloon whenever you have new messages waiting for you! Multiple accounts on the same Wiki site? NO PROBLEM! Multiple accounts across multiple Wiki sites? NO PROBLEM! Just plug your account info into WikiBiff and relax!
And if you tell me someone else already did this, after I searched all over for it, and spent all this time writing it, I'm going to be really mad. :-P
That's pretty neat! A couple of things, from a glance at the code:
For SUL accounts, having to enter the same account & password for a dozen accounts is a bit stupid, in particular if one changes ones password regularly. The accounts screen should let me tick a whole number of projects where I want to check for messages. Personally though, I'd want to check for messages per default only once per day, on a handful of wikis once per hour, and only on en-wiki every couple minutes, so that would still need to be configurable somehow. :) Even better, I'm sure there is some toolserver tool somewhere that lets me query on which projects my SUL account is registered, so that I don't have to check them myself at all. Amalthea15:51, 18 May 2009 (UTC)
I think something about that (having to enter information for each account) had passed through my mind. I'll take a look at how I can fix that. Matt (talk) 18:32, 18 May 2009 (UTC)
There is an option in the main window to have it check for messages every X seconds, but that's a program-wide setting, not a per-project setting. Are you saying you'd rather have a per-project setting? Matt (talk) 18:32, 18 May 2009 (UTC)
Yes. Thinking about it a bit more, I probably don't ever want to check for messages at some projects where I have an account at all, e.g. fa: or no:. And checking on mw: or meta: more than once per day, or commons: more than once per hour, would be a waste for me. I'd only want to check en: every ten minutes or so. Amalthea19:01, 18 May 2009 (UTC)
Passwords are stored in the registry, I gather? You might want to offer an option to not store it anywhere, but to ask it on startup. Again, for my SUL account, I would much rather input it once on startup. Amalthea15:51, 18 May 2009 (UTC)
You know that there is System.Net.HttpWebRequest, right? You don't have to write the HTTP header yourself, and don't have to parse and decode the response yourself. Unless you really really want to of course. ;) Amalthea15:51, 18 May 2009 (UTC)
I know. :-) I tried HttpWebRequest with my bot and was disappointed with its behavior -- basically, it was throwing exceptions for things it shouldn't have. A portion of my code was copied from that bot, so the old saying "if it ain't broke, don't fix it" kicked in. Matt (talk) 18:32, 18 May 2009 (UTC)
When I just tried it and left myself at User talk:Amalthea2 a message, the tray notification claimed that both my de.wiki account and my en.wiki account had messages, and a click opened both pages. I've entered account credentials for both projects. Was that wrong? Also, how do I get rid of the notification once I read it? It shouldn't still pop up once I clicked it, right? Amalthea15:51, 18 May 2009 (UTC)
*smack*. Whoops. That would probably be a problem with the HTTP request that I stupidly left setting at "en.wikipedia.org" in a couple of places. I'll post the updates in a few minutes. Matt (talk) 18:32, 18 May 2009 (UTC)
*double smack*. Just discovered another bug that would explain why the balloon keeps popping up. I posted a new binary at SourceForge. Matt (talk) 22:36, 18 May 2009 (UTC)
"User talk" is the canonical namespace name, and works on all projects, by redirecting to the localized namespace name (e.g. de:User talk:Amalthea2). So you wouldn't have needed to figure out the localized namespace. Amalthea15:51, 18 May 2009 (UTC)
Good to know! I'm not terribly familiar with all the international wikis. Didn't know they did that. Matt (talk) 18:32, 18 May 2009 (UTC)
I have recently "upgraded" my Windows Vista PC to Internet Explorer 8. My first problem was that popups stopped working, but that was quickly fixed due to some great work by User:TheDJ, and as a result I have been able to turn off "Compatibility View" mode. Another three problems with the edit textbox have now become annoying:
I have difficulty placing the caret at the end of the line. I used to be able to click the mouse in the white space to the right of the line and the caret would be placed after the last character on that line. Now, the caret is often placed at the start of the next line, though this behavoir is not consistant. Yes, I can click within the text of the line I want to edit and press the "End" key, but why should I have to type this extra key when I didn't before the upgrade?
When I place the caret in the middle of the text and I start typing, the text box often scrolls so that the caret is on the last line in the window. For example, suppose the textbox is displaying the last 30 lines of the hundreds of lines in the article and I want to edit the 5th from last line. I click on that line and start to type. The textbox will scroll so that the line I am editing is the last line displayed in the textbox. That is really annoying when I still need to read the following few lines.
I keep getting undesired scrolling in the edit textbox if I move the mouse quickly (eg. to move it out of the way, to click preview, or to move the caret). Thankfully, the editing caret stays in the same place, but the textbox scrolls by a few lines so that I can no longer see the caret.
Obviously, this behavoir could be unrelated to the IE8 upgrade, but it only started happening after the upgrade. It would be good to know that I'm not the only one experiencing these kinds of problems. Any help would be appreciated, but please no "use Firefox/Opera", "switch to Mac/Linux" or "reinstall Windows" suggestions, thank you. Astronaut (talk) 18:19, 17 May 2009 (UTC)
What do you want us to say other than "get a proper browser"?? These sound just like the sort of quirks and 'features' of Internet Explorer that piss so many people off with it. There hasn't been a change to the MediaWiki software running WMF wikis for weeks now; the only thing that's changed is your browser. If you'd raised these issues with Firefox, I'd say to go over to Mozilla's bugzilla, and in all likelihood they'd be fixed in good time. But oh wait, Microsoft doesn't have anywhere for users to file complaints in an organised way.... :DHappy‑melon18:28, 17 May 2009 (UTC)
Well, I sort of expected a "We know IE is crap, but have you tried setting the xyz option in Advanced Options" or maybe a "Since IE8 is now (supposedly) standards compliant, we are working on removing some of the IE-specific kludges we had to put in the Media Wiki software". Trouble is, the standard response from too many people seems to be "get Firefox". And yes, I would like to take it to Microsoft's equivalent of Bugzilla but as you correctly pointed out they don't have one. I'm not looking for you to fix the problems with my computer, however, I am trying to let someone know of a potential problem that could affect high percentage of Wikipedia editors. Astronaut (talk) 19:20, 17 May 2009 (UTC)
You have got a point with the IEfixes; maybe try editing without javascript and see if any of the problems go away? Other than that, and given that IE8 is pretty new, there aren't any fixes specifically designed for it. I hope it fixes some of the more egregious IE7 bugs to compensate for this whole new raft of errors... :DHappy‑melon19:32, 17 May 2009 (UTC)
I am curious about these supposed bugs in IE7. Apart from not being able to use some of the flashier Wiki gadgets due to deficiencies in IE7, I haven't experienced any unexpected behavoir while editing or simply surfing the internet in over a year of using it. Astronaut (talk) 20:02, 17 May 2009 (UTC)
I can confirm the behaviour of IE8 as observed by Astronaut. After the admirable fix by The DJ, which allowed WP:POPUPS to work in IE8 in strict mode, IE8's scrolling behaviour in Wikipedia edit boxes is highly erratic. Aside: I find the attitude towards IE users shown above inconsiderate and counterproductive. -- Michael Bednarek (talk) 06:35, 18 May 2009 (UTC)
It wasn't my intention to contemn IE users, although my disdain for the browser itself knows no bounds. As RTD notes, web developers have for many years (and it appears, will for many years still) going far out of their way to hide as much of IE's faults from its innocent users. IE's reputation for brokenness is well-deserved; it seems that IE8 - heralded by Microsoft as finally being a standards-compliant browser - is no different. Sooner or later, we'll identify exactly what problems can be fixed by web developers, with JavaScript, wierd conditional hacks, whatever, and fix them; and even further down the line Microsoft will bring out an IE9 which might fix some of these issues in core (although in all likelihood introducing a host of new ones). My comment is essentially that: we have no IE8fixes as yet, so there is nothing I could recommend to fix these particular issues. But there's a world of difference between bitching about IE being broken, and in abandoning it for that reason. We may do the former, but we'll certainly not do the latter. Not while it remains the world's most popular web browser, anyway. Happy‑melon07:57, 18 May 2009 (UTC)
I'm almost certain that this is caused by fixIEScroll in MediaWiki:Common.js. Please test disabling Javascript when you visit wikipedia. If that fixes it, then I'm relatively sure that it is the cause. Then I only need to find a way to limit this Javascript to Pre 8.... —TheDJ (talk • contribs) 09:36, 18 May 2009 (UTC)
@Happy-melon: Thank you for distinguishing between IE users and the product. I have tested FF (3.0.10 & 3.5b4) and it doesn't suit my habits as a quotidian browser (unintuitive tab switching, non-sizeable dialogues, plugin-hell, no support for Table Data Control (I know that no support for ActiveX is considered a plus, but in the Real World™ … )).
@The DJ: After disabling Javascript (or Active Scripting as IE calls it) the above described symptoms still exist. The most common symptom is that starting to type somewhere in the editing window will scroll the text in one big jump so that the cursor is now in the bottom row of the editing window. At other times the text scrolls one line further down with each character typed, until it stops scrolling when it has reached the bottom line, or sometimes it stops at the fourth-last line. I'm not reporting this with an expectation of it getting fixed, only as tidbit for the terminally curious. -- Michael Bednarek (talk) 13:15, 18 May 2009 (UTC)
After setting the columns to "100" in my preferences, I could not detect any marked improvement. After playing around with various settings, I arrived now at a preferences setting of "132" :-) —anything >100<200 would do— and a width of my editing window of 900px, which accommodates 111 characters; this gives a relatively stable scrolling behaviour. By relatively I mean that erratic scrolling occurs only about 50% of the time. The discussions you pointed to above seem to me to indicate that this is a fundamental problem of IE8 and I suspect there's not much that can be done about it, but I may of course underestimate your ingenuity. -- Michael Bednarek (talk) 10:12, 19 May 2009 (UTC)
I have raised the issue with Microsoft's technical support. Maybe they will get back to me with something or send out another update/patch. I'm still waiting for a reply - hopefully something along the lines of "Congratulations, you have found a genuine bug. Here's your bug ref. number" - but it could take them months, so I'm not holding my breath. Astronaut (talk) 13:30, 19 May 2009 (UTC)
Would you mind leaving a note on my talk page if/when you hear something? I would like to take this page (VPT) off my watch list. Thanks. -- Michael Bednarek (talk) 13:41, 19 May 2009 (UTC)
No. If the words are separated by slashes (i.e., if they are subpages) then you can split them with the #titleparts parser function, and you can theoretically try to use the string manipulation templates to split it at the spaces, and if you have a limited list of words you expect then you do some tricks as well. What is it you are trying to do? Amalthea11:20, 19 May 2009 (UTC)
I would like to insert a #switch in a table, based on a disambiguator. Say I have articles about athletes, then the heading of a table's column should change depending the occupation given in the article's name. Originally I had tried to use {{#switch:{{PAGENAME}} | (footballer) = football | (basketballer) = basketball | #default = No athlete }} for this, however, as PAGENAME obviously requires the full article name, it didn't work.
Which template exactly? If it's an infobox template, they usually get the name without the qualifier already, so that you can compare {{#ifeq:{{{name}}} (qualifier)|{{PAGENAME}}|does end with (qualifier)|does not}}. Two templates that do this and similar things, with somewhat different methods, are {{Taxobox name}} and {{italictitle}}. There also is {{str endswith}} for generic input, but for your case it could be written a fair bit more efficient. However: For athletes, the page title does not usually have any qualifier, right? So an automated function like that would only match on a minority of all athlete bios. If the template you have in mind is {{Infobox Athlete}}, then it already has a |sports= parameter that could be used to switch, I assume? Amalthea12:55, 19 May 2009 (UTC)
Sorry, I forgot to mention that I have this rather not intended for Wikipedia, but mainly for my own PC, and I should have also accentuated that athletes was merely an example as I want use such switches preferably in various topics. However I will take a look at your suggestions and will check if I can adapt those to my purposes. In any case thanks for your efforts.
Hello, I am trying to get Miszabot to archive my talk page automatically. I believe I have inserted the code correctly there, and have waited the 24 hrs to make sure the bot has 'run' a cycle. Did I do something wrong? (Is this the wrong place to ask?) Thanks for your help Dougofborg(talk)20:19, 15 May 2009 (UTC)
I seem to recall there's also a minimum number of threads parameter, which defaults to 5. Since you have 5 threads at present, the bot might leave your talk unchanged. I'll add a thread as a way of testing. --AndrewHowse (talk) 15:26, 16 May 2009 (UTC)
Especially since both pages existed as redirects until a few days ago. I suspect that Google does not index redirect pages, thus it may take a bit longer for these to index. ---— Gadget850 (Ed)talk15:01, 16 May 2009 (UTC)
I'm just going to go with "I don't know" here. I dealing with a user page that was created yesterday and is now indexed, but the two articles in question here still don't show in Google. ---— Gadget850 (Ed)talk15:30, 20 May 2009 (UTC)
Done, with |link= in the image tag rather than imagemap (which is for complicated link structures). In future, you can use the {{editprotected}} template to draw administrators' attention to such edits. Happy‑melon19:29, 17 May 2009 (UTC)
Search Engine Problem
Using the wikipedia search engine for 'centrifugal force' sometimes leads straight to the 'centrifugal force' article. But mostly it will lead to a branch article entitled centrifugal force (rotating frames of reference). Is there anybody who knows how to fix this problem? David Tombe (talk) 07:19, 19 May 2009 (UTC)
Recently, I noticed that some of the formulas changed size for me. In particular simple formulas with only a subscript or only a superscript got smaller. I heard that equations are displayed differently depending on settings. Is there a way that I can fix the equations so that they display the correct size. Barring that what settings do I need to change. TStein (talk) 18:19, 19 May 2009 (UTC)
Failed to parse (<math_output_error>): \{g|p_0\rangle|g\in G\}.
in the math formula on the page, which I copy here (don't know if it will work):
I don't know what's wrong with the formula, but fine. Here's the weird part. When I view the exact same version of the page through the permanent link to that version, here, or the difference between that version and the previous version here, it parses correctly! It also parses correctly when I view the exact same page in "Edit --> Preview" mode. Do other people reproduce this? Any ideas why this would happen or how to fix it? Thanks!! :-) --Steve (talk) 16:49, 20 May 2009 (UTC)
I'd guess there is a problem with the math parsing on one of the servers; which caused it to parse incorrectly. Previewing or purging would likely hit a different server, that doesn't have the problem. -Steve Sanbeg (talk) 18:23, 20 May 2009 (UTC)
When trying to edit the page Template:@@ I get the following error:
A database query syntax error has occurred. This may indicate a bug in the software. The last attempted database query was:
(SQL query hidden)
from within function "RecentChange::save". MySQL returned error "1205: Lock wait timeout exceeded; Try restarting transaction (10.0.6.22)".
Probably it is my fault (a rather weird wikilink) but I am nowunable to fix it. Could someone please help ? Thanks... --Jorge Stolfi (talk) 04:47, 21 May 2009 (UTC)
I don't know whether this is a bug with me or a bug with wikipedia (and if it's with wikipedia then whether it's with <ref> or with {{subst) but there's definitely a bug in here somewhere.
Here's the test case code which I entered, escaped so you can see it:
Today is {{subst:today}}
<ref group="refexample">Today is {{subst:today}}</ref>
Which turns out to be bugzilla:2700. With a title like "Pre-save transform skips extensions using wikitext (gallery, references, footnotes, Cite, pipe trick, subst, signatures)", no wonder I couldn't find it. — Carl (CBM · talk) 23:59, 17 May 2009 (UTC)
Indeed. Makes me feel better about not finding it too.
Given the nature and age of that bug, is it worth adding a Known bugs section to the documentation of the affected areas, such as WP:Subst, to explain the restriction, or at least to link to the bug report?
Extending the above query: How can I get a template to expand the contents and the name of <ref name="{{{2}}}">{{{1}}}</ref>? --Jorge Stolfi (talk) 04:42, 21 May 2009 (UTC)
Found the answer (I Think): {{#tag:ref|{{{1}}}|name={{{2|default}}}}}
Is it possible to suppress the link generated by <ref>STUFF</ref> (that is, the [12] marker) without supressing the STUFF entry in the {{Reflist}}? Or to replace the number within the brackets by something else? --Jorge Stolfi (talk) 05:45, 21 May 2009 (UTC)
No. You could hide the whole reference with inline CSS, but then you'll end up with a backlink from the References section that goes nowhere. Which, BTW, is why I think {{source list}} is a bad idea. Anomie⚔10:54, 21 May 2009 (UTC)
Well, resolved it by commenting out the transclusion of {{Special:Newpages/4}}. Still don't know what was going on though.--Kotniski (talk) 11:07, 21 May 2009 (UTC)
hello,
some months ago, a Wikimedia announcement talked about planned implementation of a feature where display of video start points/images/other media could be triggered by text links, quite similar to those green links in http://proteopedia.org .
As I can't find the proposal anymore, can you please point me to it or new information about it? --Ayacop (talk) 17:43, 21 May 2009 (UTC)
Constant server hangups
About half of my operations on Wikipedia begin normally but hang indefinitely at some random moment. There are no error messages; the cursor just keeps spinning forver, as if the server had just "forgot" about a request that it just accepted. I then have to abort the operation and retry. The retry almost always work fine, but sometimes it hangs again, at a different point along the transaction. I cannot see any definite pattern, but it seems . It is not general slowness: the transactions that do not hang usually complete at the normal speed. These hang-ups happens while reading pages, editing and saving, history and watchlist lookup. It started one or two weeks ago and happens at at all hours of the day. It happens with Firefox and Epiphany. I cannot exclude a local or network problem, but this behavior does not seem to occur when I access any other site. At first I assumed that it was due to the WP dump in progress, but the dump is over (so it has been said) and the hang-ups are still there. Is this happening to anyone else? Where --Jorge Stolfi (talk) 02:45, 22 May 2009 (UTC)
Archive policy
Can I put a notice that the archive policy applied at Wikipedia:Village_pump_(miscellaneous)#Miscellaneous_archive also applies here? Currently, there does not seem to be a clear policy. If I don't get any objections within a week, I'll do it. JesseW 04:55, 8 Nov 2004 (UTC)
What's the best way to change/edit the name of an article I created? I thought I would solve the problem of a non-uniquely named poet by titling the article using the middle name. This creates more confusion, so I'll learn how to create a disambiguation page. Meanwhile, taking baby steps, I need to rename the article from Tom Poeller Mandel to, simply, Tom Mandel, the non-unique name. Many thanks, Gaelyn
Click on the "move" tab at the top of the article page. This moves the article contents and history to a new name and leaves a redirect in place of the old title. EncMstr22:36, 20 June 2006 (UTC)
This is only an issue if neither the question nor the response have timestamps (right?). I imagine it's pretty common for questions not to have one, but responses should always be signed and timestamped so I'd think this would very rarely come up. -- Rick Block (talk) 16:37, 5 September 2006 (UTC)
Always tell what browser you use when your question is about what you see on your screen. This information shall improve help. People here are used to some details or bugs, or are able to reproduce your case in some of the most popular browsers. --82.227.17.3019:23, 23 October 2006 (UTC)
It is possible that the User:Curtisaallen account was registered some time ago but never utilized till today. When such accounts are layed away for nefarious purposes they are usually known as sleeper account on Wikipedia. (→Netscott)16:01, 22 November 2006 (UTC)
What would it require for the program to be changed so that users with no edits so far would be treated as if they were brand-new? Wahkeenah16:07, 22 November 2006 (UTC)
Done. Make sure you type #REDIRECT [[Meridian Charter Township, Michigan]] in exactly one line. Although you probably wanted to put this on WP:VPT, not its talk page... Titoxd(?!?)03:30, 1 December 2006 (UTC)
Click Meridian Township (you'll be redirected to Meridian Charter Township, Michigan). At the very top, under the title, the words "Redirected from ..." appear, with a link that will actually take you to Meridian Township (without being redirected). Click this link and then "edit this page". This page's contents are a single line starting with #REDIRECT, as indicated above. To make another one of these, edit the article name you want to be a redirect and add just one line, like the one in Meridian Township. -- Rick Block (talk) 02:20, 8 December 2006 (UTC)
When I posted that, the user was posting every few minutes. As soon as they found their text had been reverted, it was back. It took longer when they found they had been blocked, but only a couple of hours. So it would have been more than a "occasional revert". Will (Talk - contribs) 00:28, 6 January 2007 (UTC)
Extremely slow tools.wikimedia.de
I find that any link to tools.wikimedia.de has an extremely slow response time, I think minutes to tens of minutes. (I haven't measured it lately; it takes too long to bother waiting.) This didn't matter so much when it was just Kate's tool, but now the geographic coordinate links have moved to t.w.d as well.
It seems to me that others must be getting reasonable response times, or no one would have accepted the migration. Can anyone explain why t.w.d is responding so slowly for me? I'm using Mozilla on Debian stable. I recall once long ago getting a reasonable response time from Kate's tool on t.w.d, when accessing it from Windows. --Trovatore20:05, 27 January 2007 (UTC)
I can't find it in about:config (Mozilla 1.7.8). Thanks for the suggestion, though. Could you give me an idea why you think it might help? --Trovatore06:00, 30 January 2007 (UTC)
I have found network.dns.disableIPv6 in 1.7.13. But I don't think that's the problem, I have it set to false and I don't have problems with the toolserver. Tizio15:30, 28 April 2007 (UTC)
Picture Placement
Hi, how is it I can choose where on the page I'm editing where the picture I want to add is placed? Whenever I try to place one, it always appears at the top of the text, not in the desired location. Cheeseah (talk • contribs) 21:08, 19 March 2007 (UTC).
What is the difference between this section of the Village pump and the MediaWiki space? I would like to have a discussion on managing conversation sprawl, but I would like to make sure that I am using the right channels for such a discussion. --Aarktica13:04, 26 April 2007 (UTC)
The MediaWiki talk namespace is for discussing a particular interface message (for instance, if you want to change the wording of the warning below the edit box); WP:VPT is for general technical comments, questions, and discussions (such as asking questions about how to achieve something complex in wikimarkup, asking about a particular sort of login problem, or wondering whether a particular proposal is technically possible). --ais523 15:42, 27 April 2007 (UTC)
Thanks for the feedback. Just so that I understand, the proper course of action is WP:VPT -> WP:VPR -> MediaWiki ? --Aarktica15:17, 28 April 2007 (UTC)
No. Pick the most applicable page, and expect to get an answer. If you don't, or if a knowledgeable editor suggests you post elsewhere, do so. The only reason to plan to go to multiple places is if you're checking about the technical aspects of a proposal before proposing it elsewhere. -- John Broughton(♫♫)15:37, 23 January 2008 (UTC)
Finding out if articles are watched
Adding this here as well since feedback at Help:Watching pages is slow:
From a maintenance/anti-vandalism point of view, it would be helpful to know how many people, if any, are watching a given article. Is there any way to ascertain this information? I can find nothing on Help:Watching pages suggesting it is. If there isn't, I think it should quickly be added as a feature so that editors can locate unwatched articles to help maintain the quality of the encyclopedia, as unwatched articles have little defense against drive-by vandalism.
Unregistered users should not be able to see this information as it would help vandals locate unwatched pages and make a mess of them. Only established (semi-protection criteria) users should be able to view this information.
Thanks for your quick reply. Could you specify to what extent it exists for administrators? Can they tell who watches the page as well as how many?
I think this needs to be extended to all users that have held an account for a few days. It's harmless to let people know how many are watching an article without specifying who they are, and helps people identify unwatched articles so they can better focus their efforts. There is the risk of higher vandalism from those few cunning vandals though, so I don't think it should be applied to non-account holders. Is there any harm you can think of in extending it to account holders? Richard00101:04, 2 May 2007 (UTC)
As far as I can tell, it lets admins know what the first 1000 unwatched pages in alphabetical order are, and not how many people are watching watched pages. (As it only goes up to 1000, it's kind-of useless; I wonder if this is a bug?) --ais523 12:50, 2 May 2007 (UTC)
I wanted to wikilink Joseph M. Scriven (there is an existing article on Wikipedia under that name) in this article He Is Beautiful. I was not able to get Joseph M. Scriven to be recognised as a link (in blue) until I copied the heading Joseph M. Scriven from the original article's heading and pasted that into the article He Is Beautiful. This does not make sense to me except for the fact that Joseph M. Scriven appears as Joseph M. Scriv-en in the wiki tab and address. How could this happen and is it a common problem? Address will need fixing. --User:Brenont19:35, 30 September 2007 (UTC)
I have no idea what the problem was or why it occurred, but moving the page to its own apparent name appears to have solved the problem. Nihiltres(t.l)20:12, 30 September 2007 (UTC)
There is in fact a character in the middle of the old title; a soft hyphen. Software that makes use of them will not display them unless it needs to put a line break in the middle of the word, but you can see it if you copy and paste it into a more naive text editor like Notepad. It may be that the article was created not by typing the name in, but by pasting it from a word processor, or perhaps from an online article at a website that likes to right-justify its text and so uses hyphens to make them come out looking better. TCC(talk)(contribs)20:54, 30 September 2007 (UTC)
(conflicted) I was going to say the same thing. I suspect that it was some other article that was created in that way, and someone later clicked a redlink there to create the Joseph M. Scriven article. Anomie21:17, 30 September 2007 (UTC)
I can't use the buttons above
For some strange reason, I can't use the buttons that are above when editing on a page. I can't make the text bold, cursive, or sign my nickname, unless I do it by manually typing the script. This seems to be a problem on this computer, only. Does anyone know if I need to change something? When I move my cursor over the buttons, a textbox appears with the explanation for what it means, but I just can't click the damn thing! Ugh! -- User:Anittas —Preceding signed but undated comment was added at 17:31, 10 October 2007 (UTC)
What browser are you using? There's some code in place to fix problems with image transparency in Internet Explorer 5.5 and 6, and there were some reports that it interfered with the toolbar in those versions. I thought those problems were fixed, though; try bypassing your cache, and if the problem goes away, then that was the reason and the cache bypass will have applied the fix. --ais523 17:35, 10 October 2007 (UTC)
Yes, I thought we had that fixed for IE6...but in any case, IE7 will work a lot better for you with Wikipedia. It will actually load pages faster and in general work better than IE6. —Remember the dot(talk)21:05, 10 October 2007 (UTC)
For the record, it turns out that we did have them fixed, but the script was improperly modified (without being tested) and it broke the editing toolbar again. —Remember the dot(talk)00:36, 11 October 2007 (UTC)
Arabic text
Not for the first time I've run into problems trying to edit articles which contain Arabic text. On this occasion I've been trying to add the date of death (1992) into the brackets in the first sentence of Youra Eshaya. The preceding Arabic text interfers with the Roman text in the edit box - try it if you want to see what I mean. PS I'm currently editing with IE 7. Any ideas? Greenshed20:02, 20 October 2007 (UTC)
But if you were to attempt to change "(born 1933, died 1992)" to "(1933 – 1992)" (in line with the MOS?) then I think that it would all go wrong. Greenshed23:23, 20 October 2007 (UTC)
I use Windows 2000/IE6 and only had the "Japanese" font installed once. Since then, most languages requiring unicode display just fine. I had no trouble editing the first line. — Edokter • Talk • 00:09, 21 October 2007 (UTC)
Fun with bi-directional text. The Arabic name starts a right-to-left context, and when you remove the word "born" (which begins a left-to-right context), the remaining punctuation and numbers inherit the Arabic's right-to-leftness. One easy solution is to insert the character U+200E just after the name; this invisible character has the strong left-to-right property that is needed to prevent the punctuation from being interpreted as right-to-left. Anomie01:37, 21 October 2007 (UTC)
how can I make a link to a specific paragraph of some other article? For example a link to the paragraph "Tampa Jook Music" of the article "Music of Florida" from a different article.. Thanks--Lykantrop(Talk)14:35, 1 March 2008 (UTC)
I'm not sure if this is the place to report it, or if it has already been reported, but HotCat breaks on "view source" pages (example), causing all my other javascript to break also. The exact error is "document.editform has no properties", for obvious reasons. — Bob • (talk) • 22:32, 11 April 2008 (UTC)
Speak to TheDJ? I know how to fix it (if(document.editform) //blah else return;) but there's probably a better solution, one TheDJ would likely know. x42bn6TalkMess22:46, 11 April 2008 (UTC)
[ecx2]I'm assuming it's a permissions problem; as an administrator with the script installed, I don't have any problems with it. EVula// talk // ☯ //22:49, 11 April 2008 (UTC)
Problem should be fixed now, but you might have to trash your browsercache in order to get the latest version. Thx for the report. --TheDJ (talk • contribs) 23:02, 11 April 2008 (UTC)
Hello, I own the Username Ataraxie on the french wikipedia. It's taken here by someone who has not contributed for quite some time. Would it be possible to rename it in order for me to merge accounts ? Thank you.--Ataraxiel (talk) 13:54, 9 June 2008 (UTC)
Try asking at WP:CHU/SUL, that's where the English Wikipedia handles all SUL-related rename requests. --ais523 14:13, 9 June 2008 (UTC)
I would like to discuss with someone who is an EXPERT on PARSON FUNCTIONS how to get them "activated" on a Wiki site. - Ixthis888 (talk) 00:33, 17 July 2008 (UTC)
How about on top of the header for the page, have a list of ongoing technical issues (like the missing image thing) so we don't keep getting the same question? --Random832 (contribs) 21:05, 14 September 2008 (UTC)
My guess is that most new users wouldn't know to check it. It's just as easy to say "we know about that" as it is to say "look at the top of the page", but the first option is more friendly than the second. — Carl (CBM · talk) 21:11, 14 September 2008 (UTC)
In Template talk:Infobox VG you'll see that a particular topic has been archived separately in order to group like-topic threads and cut down on duplicate proposals. I think this would be the best approach, but also the hardest. (Probably near impossible for the Village Pump, unless someone were to step up and take a definite interest in the task.) SharkD (talk) 21:43, 30 December 2008 (UTC)
Talk archiving
If a talk page references itself, then the link(s) may go dead once a page has been archived. I was wondering if more sophisticated methods of archiving Talk pages have been pursued in order to prevent such errors? SharkD (talk) 21:38, 30 December 2008 (UTC)
We really should bring Wikipedia up-to-date. Semantic MediaWiki, a better design and increased HTML support would help - the internet's advanced, we haven't. Unfortunately cut-&-paste-ing the threads to a subpage is the only method we have. We should make Wikipedia more 'flowing', and all our problems will be solved! TLDR - No, there is no more sophisticated way of archiving talk pages. DendodgeTalkContribs21:56, 30 December 2008 (UTC)
For any Talk pages in any namespaces, I am seeing the following message when trying to view the page histories:
A database query syntax error has occurred. This may indicate a bug in the software. The last attempted database query was:
(SQL query hidden)
from within function "IndexPager::reallyDoQuery (PageHistoryPager)". MySQL returned error "1054: Unknown column 'ts_tags' in 'field list' (10.0.6.22)".
The current limit on number of category page items to display per screen (200) seems unnecessarily low to me. Would it be a big deal resource-wise if it were raised to, say, 500?--Kotniski (talk) 10:40, 18 May 2009 (UTC)
I imagined it could be - that's why I was asking, with a view to proposing that it be changed. It partly depends what people find most convenient when browsing categories (shorter or longer pages), but I wondered if there were any server-load concerns as well.--Kotniski (talk) 09:11, 19 May 2009 (UTC)
I don't see what's stopping us from allowing category urls to accept parameters like &limit=500 to override the config'd amount. However as far as aesthetics go it might be useful to allow the paging limit to be specified on a per-category basis, e.g. where the member pages have long titles, one may want fewer per page. — CharlotteWebb12:18, 23 May 2009 (UTC)
for starters, I'd have to go to each and every country data template {{country data England}}, {{country data Argentina}}, {{country data Australia}}, etc. - all of which are fully protected pages - and request the addition of a new parameter/data combo. well, I suppose I could get an admin to AWB it... It would make sense to do it that way, I'd just want to research it some more, first, so we don't screw up half a gazillion flag templates in the process.--Ludwigs2 21:18, 28 April 2009 (UTC)
You are a god among men. cheers, –xeno talk 14:37, 29 April 2009 (UTC)
Works for me , what's it doing for you? –xeno talk 20:26, 29 April 2009 (UTC)
--Ludwigs2 17:22, 24 April 2009 (UTC)
Resolved
Done issue fixed. —TheDJ (talk • contribs) 19:52, 2 May 2009 (UTC)
Fixed thanks to User talk:MSGJ. Regards, SunCreator (talk) 18:07, 7 May 2009 (UTC)
. Still I think the principle of writing good code for its own sake would apply regardless of the exact mechanism by which bad code is repaired. So when I say html-tidy I'm referring more broadly to any part of the software that makes broken code less broken. — CharlotteWebb 22:48, 9 May 2009 (UTC)
—Animum (talk) 01:44, 10 May 2009 (UTC)
/Archive_60#c-Ukexpat-2009-05-05T17:05:00.000Z-File:Daniel_Roth.jpg){kind=link}
/Archive_60#c-Stifle-2009-04-30T13:43:00.000Z-File:Daniel_Roth.jpg){kind=link}
![[11]](https://commons.wikimedia.org/wiki/File:Rugby_football_current_event.png){kind=link}
_dvd_box_set_cover.jpg&action=edit&redlink=1){kind=link}
_dvd_volume_2_cover.jpg&action=edit&redlink=1){kind=link}
