Jump to content

Wikipedia talk:Wikidata/2017 State of affairs/Archive 3

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1Archive 2Archive 3Archive 4Archive 5→Archive 10

A poor discussion

This page is a terrible contribution to the discussion of the underlying issues.

I feel I could contribute both on "data integrity on Wikidata" and "reliability of English Wikipedia content", and the relationship between those two issues. But I have no intention of doing so here. I would actually be glad to have Wikipedia:Wikidata/2017 State of affairs deleted. The personal attacks on this page could be deleted under talk page guidelines.

The scope of the subpage title has shown itself, pretty much, to be too broad. We can surely do better than this. Charles Matthews (talk) 12:19, 19 January 2017 (UTC)

Feel free to start any page about this you like. You are also free to nominate this page for MfD, but I don't see a good policy reason for it, nor a good non-policy reason either. The page, poor as it is, will be used in an upcoming RfC to give at least some background to what happens already with Wikidata on enwiki, and on how people perceive this. The value of this talk page is not really impressive though, but is indicative of the fact that this issue divides (parts of the) editing community and needs more facts and less ill-informed contributions all around. Personal attacks can also be removed of course, just make sure that they are real personal attacks and not general criticism of edits or something similar. Fram (talk) 13:13, 19 January 2017 (UTC)

Well, I suggest a "refactoring" of the page would be beneficial. You do realise that some of the discussion above is "pot-kettle"?

I have been working on Wikidata since 2014, on projects which have nothing to do with infoboxes. I know certain benefits have arisen, for the English Wikipedia (and not only that one). I would say your comment above about limiting Wikidata to the indexing function (interwiki) is not really accurate. I think the verifiability issue on Wikidata is important, but not as simple as is sometimes implied.

I'm concerned that you say the page "will be used". As you say, in fact, it reflects mainly that there are divisive arguments used here, about Wikidata. It is currently fashionable to force the issue in divisive areas of politics: and we don't see the argument from factual evidence getting much credence there.

WP:VOTE, full title Wikipedia:Polling is not a substitute for discussion, makes the point in its nutshell: "Some decisions on Wikipedia are not made by popular vote, but rather through discussions to achieve consensus. Polling is only meant to facilitate discussion, and should be used with care." I am arguing for care, given that this Talk page shows not the slightest sign of emerging consensus. Charles Matthews (talk) 13:40, 19 January 2017 (UTC)

"I would say your comment above about limiting Wikidata to the indexing function (interwiki) is not really accurate."? What comment do you mean?
As for "this page will be used": I plan an RfC (probably first a stage to decide, if possible, on a number of questions for the RfC, and then the actual RfC) to get some consensus about the use of Wikidata on enwiki (not a simple yes or no, but more of a "this, this and this is allowed, but that is not allowed). At the moment, all I note are discussions about this without any resolution, with one project embracing it and another rejecting it, with people implementing e.g. Wikidata lists on their own, with others making phabricator tickets to add wikidata links to enwiki, and so on. It would be great if you could start a page that would create a policy or guideline for this which has broad consensus and which is achieved by simple, congenial discussion. I doubt this can be achieved though, and think that an RfC with a number of neutral questions can see if there is a consensus or not. Such consensus in any case is not decided by simple votecount, but by the closing admin(s), so the end result may still be a "no consensus". But then at least we have tried.
So, feel free to start another page and gather all factual evidence that you can. I have little faith that it will fare better than this one, but I would gladly be proven wrong. Fram (talk) 14:03, 19 January 2017 (UTC)
Fram, one function related to Wikidata that I have seen used to great effect is called "mix 'n match". It addresses some of the objections you raised above about getting links correct. See meta:Mix'n'match and the actual interface here. I hope someone else can explain better than me the benefits of that. Carcharoth (talk) 14:14, 19 January 2017 (UTC)

Let's take the usability discussion, for example. How could that be made "factual"? It is more suitable for a separate page, I would say.

I was actually thinking yesterday about the "Richard Burton's wives" issue you raised, because I was adding education histories to Wikidata items about people, and it is annoying if the order doesn't follow chronology. For Burton, you are actually "supposed" to add a "start time" qualifier in the "spouse" statements; and I would say that you should reference it. That is, the marriages should come at least with dates, if we are going to attach significance to any ordering at all. It is certainly is troublesome. if the marriages are out of order, for the human reader. You can only fix it by deleting and retyping. But actually, with the dates, the system could fix that. This doesn't yet happen in the maintenance of Wikidata, but clearly it could.

This sort of thing strikes a newcomer as a usability issue. Wikidata statements are added to the bottom, sort of - there is a division into substantive statements, and identifiers; and there is a setting in Preferences to display the substantive statements in a standard order.

So those are some facts, and explanations. Do you agree that such matters deserve a page of their own? Charles Matthews (talk) 14:26, 19 January 2017 (UTC)

A page of their own, or a section on this page (not the talk page, the actual page). This page is intended as a central repository, if some aspects need more detailed explanations then subpages may be the way to go. But while we need such pages or sections as background, they will never decide whether we (the broader community) trust Wikidata enough to let them insert X into Wikipedia (where some X will be widely accepted, and some others will be generally refused, to take an extreme example, I don't think many people will let Reasonator generate articles, or will let a bot delete articles here when for some reason the Wikidata entry gets deleted). Fram (talk) 15:07, 19 January 2017 (UTC)
I don't think anybody calls for Reasonator creating articles on Wikipedia or call for bots deleting Wikipedia articles when Wikidata entries get deleted. That's a strawman.
We seem to disagree about the usage of Listeria, data import in templates, links from Wikipedia to Wikidata and the article placeholder tool.
There also an open discussion about what tools should be build to improve the integration. That goes for direct editing of Wikidata via Wikipedia, Watchlist integration, new versions of listeria and also further projects like Librarybase. ChristianKl (talk) 09:09, 23 January 2017 (UTC)
[1] This Article Placeholder discussion certainly saw some prominent Wikidataians calling for Reasonator-created pages on Wikipedia. "I agree with GerardM. We should be getting rid of substubs and replacing them with reasonator pages not creating more substubs." Fram (talk) 09:57, 23 January 2017 (UTC)

Sorry – "they will never decide whether we (the broader community) trust Wikidata enough ..." – did you get elected to office here? I feel this kind of rhetoric is misplaced. Maybe half a year of Brexit does that to people.

I would say, with the best will in the world, that what is needed is a "Wikidata FAQ for Wikipedians". And I think the effort to get one by choosing this particular route in project space has probably failed. Which is why the page should be deleted, or edited very seriously.

But anyway, isn't a FAQ what we are talking about? Charles Matthews (talk) 15:22, 19 January 2017 (UTC)

Rhetoric? A FAQ, a page about Wikidata usability, this page, ... are (at best) interesting and helpful, but "they" (these pages and sections, not the people editing them) can not decide what the policies about Wikidata will be. That will happen through discussion, probably an RfC (as the issue seems to me to be too acrimonious to be decided by simple discussion), where the broader community (you, me, ... i.e. "we") can have their say. In my view, the ultimate point for what will be accepted and what won't be accepted is whether we (again, you, me, all others participating) trust Wikidata or its implementations enough. Do we trust Wikidata to provide lay-out, or do we want to keep that local? Do we trust Wikidata for sourcing? Do we trust Wikidata for BLP-sensitive things? You may describe the decision-making process here, and what will be the driving force, in your own words, and you don't have to agree that trust will be a deciding factor, but there is no reason to air your disagreement with "did you get elected to office" and references to Brexit. My "rhetoric" was directly related to the issue at hand, your rhetoric though was rather out of the blue and indeed misplaced. A bit sad to see you coming here to complain about a poor discussion which should be deleted because of the personal attacks, and then reply in such a manner to what is a rather normal and relevant statement, even if you don't agree with it. Fram (talk) 15:34, 19 January 2017 (UTC)

How about the ultimate point being the Wikipedia mission statement? Say as on q:Jimmy Wales. Or the WMF statement of purpose: "The mission of the Wikimedia Foundation is to empower and engage people around the world to collect and develop educational content under a free license or in the public domain, and to disseminate it effectively and globally." As far as I'm concerned, the latter has quite a lot to do with Wikidata.

How about addressing my point on the need to build consensus, per the important WP:VOTE essay, rather than laying down sweeping preconditions?

I apologise, as a UK citizen, for everything about Brexit. I did think you might include under "general criticism of edits or something similar" a criticism of tone.

I do think, if you will try not to take this as a personal attack, that the non-specific use of "we", as if representing a constituency, was not helpful. It anyway gives me a clue as to why this debate is not going well. Normally editors here speak for themselves. Charles Matthews (talk) 16:24, 19 January 2017 (UTC)

Quoting Jimbo Wales is unlikely to convince me ;-) First, I think I have caused the friction or misunderstanding here (at least in part) by using "they" instead of "these" in the "they vs. we" statement above. I meant that information pages don't decide consensus, but the community does. "We" was not meant as as a group separate from some undefined they, but as "all of us". That's why, what you call a non-specific use of "we", had the parenthesis "(the broader community)", as an attempt to explain what I meant. I clearly failed there... Anyway, there are different ways to achieve the statement of purpose, and I am not convinced that the use of Wikidata is in most cases the best way to achieve this. Ccriticism of Wikidata and not wanting to use Wikidata for some things doesn't mean that one doesn't support the WMF statement of purpose. And an RfC is a medium to gauge consensus, even though that essay may not believe it to be the best or even a good one. Like I said, you are free to try a different approach to get consensus about Wikidata use. Fram (talk) 17:43, 19 January 2017 (UTC)

OK then, it is better if the opposite of "us" is not "not one of us". Let's try for some consensus starting from GIGO. That's an important principle. So is the principle that one has to live in the real world. I find it interesting to draw a parallel with Commons. The metadata on Commons is, from some points of view, in a bad way: it can be 10 years old. It is essentially never footnoted. There are mistakes made here with images misidentified, so being placed incorrectly in articles. More worrying, if quite rare, is that apparently good metadata from a GLAM is wrong: I know this happens. It is unlikely that in all cases independent, third-party referencing could be provided: we tend to accept the GLAM metadata as authoritative.

So, why do we accept the use of Commons images here routinely, instead of insisting that everything be uploaded here, and the metadata scrutinised? I would suggest this is not a matter of "trusting" Commons: because the mass uploading probably makes that (a) meaningless and (b) implausible. There could be plenty of photoshopped nonsense: copyvio is more of a concern. There are two things acting in the other direction: Commons is very helpful in developing articles; and images, we feel, tend to provide internal evidence that we use in place of good metadata. "In the real world", don't we mainly just accept Commons on that basis?

To sum up: some of the arguments deployed in the case of "trusting" Wikidata, to adopt your term though I don't like it, can be seen in action here. One kind of argument would say "Wikidata is not very useful in developing articles or lists here". I don't agree, but we can have a rational discussion about that—and the WMF point I was introducing is that enWP and deWP are the hardest Wikipedias on which to make this argument, because they already have so many non-stub articles, the case being clearer on smaller wikis. The other part is clearly more interesting. It is rational to say that adding an image of a dog here and giving the caption "cat" is not really troubling: it is child-like vandalism and a small child could spot it. I won't try to complete the discussion here: I did say in my original posting I wasn't going to do that. But I think GIGO can be used in a nitpicky way, as well as sensibly. Charles Matthews (talk) 18:48, 19 January 2017 (UTC)

The other problem with insisting that everything be uploaded here, and the metadata scrutinised? is that it won't happen; hosting files here does not increase the scrutiny or quality in any way. The issue with Wikidata is that there is a widespread feeling that in information hosted on Wikidata there is less scrutiny and quality than information hosted here. Jo-Jo Eumerus (talk, contributions) 19:30, 19 January 2017 (UTC)

Yes, that was the "widespread impression" I said I wasn't going to address here.

Why I said that there was a pot-kettle argument in progress is that I don't think Wikipedians have the right approach, if they think denigration of Wikidata helps anything at all. For reasons User:Carcharoth alluded to above, amongst other things, Wikidata can be helpful with Wikipedia's quality of information, and has been criticised for its importation of data from Wikipedia. The part that might be taken seriously of all that is any issue of circular referencing.

Putting Wikipedia "factoids" into Wikidata where they are scrutinised ... well, honestly, if people can't see some good might come it, I don't want to spend much time on further exposition. Suffice it to say that machine-readable format could help with detecting anomalies. Charles Matthews (talk) 19:46, 19 January 2017 (UTC)

Blocklist

The page currently says "We can block links from being added to enwiki (through the blacklist); what if these links are inserted from Wikidata (in infoboxes or so)?" This is incorrect. We have a global blocklist that's also blocking edits on Wikidata. ChristianKl (talk) 09:42, 23 January 2017 (UTC)

And on every Wikimedia project. Meaning that we need a damn good reason to have it added to the global list. Jo-Jo Eumerus (talk, contributions) 09:53, 23 January 2017 (UTC)
Yeah, it is not easy (or quick) to get something on the global list (which is why individual wiki's have their own). Only in death does duty end (talk) 10:36, 23 January 2017 (UTC)
Indeed. We now have a discussion on whether the Daily Mail should be added to the enwiki blacklist; to get this added to the global blacklist would be almost impossible, I think. The local one is [2], the global one is [3]. Fram (talk) 10:57, 23 January 2017 (UTC)

Is the page communal or a collection of opinions?

I was going to edit the page, but then I noticed that is full of signatures. If the page is communal, shouldn't signatures be removed from it?--Micru (talk) 08:06, 19 January 2017 (UTC)

Some people have signed, some didn't. Feel free to add your own opinions to it with or without signature: it's intended to collect opinions, it's not that important who actually added what. Fram (talk) 08:10, 19 January 2017 (UTC)
I find the mix confusing, as it discourages improvement of the page and the different statements. Specially "Uses of Wikidata on enwiki", should be factual and a collaboratively written, but as it is signed, it prevents discussion. Personally, I would prefer to remove all signatures, and allow edits provided that they keep the spirit of each corresponding section.--Micru (talk) 09:08, 19 January 2017 (UTC)
I would prefer the same as well, and intended it to be a signature-less page, but some editors have been edit-warring to retain the signature of someone else on some statements, to the point that the page got fully protected and the edit-warring editor only escaping a block because the 3RR report was stale whne someone finally got around to act on it. Fram (talk) 10:06, 19 January 2017 (UTC)
@Pigsonthewing and SlimVirgin: As participants in the previous conflict, are you ok in removing all signatures from the page and keeping the discussion in the talk page about contested points? I see the signatures as an obstacle to develop the page, but before proceeding I would like to know your opinion too.--Micru (talk) 10:43, 19 January 2017 (UTC)
Currently, they act as a barrier to collaborative editing. That said, I'm not sure that removing them will fix the underlying problems. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:32, 19 January 2017 (UTC)
As has just been amply demonstrated. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:45, 23 January 2017 (UTC)

Authority control

Template:Authority control is added to more than 500,000 pages so far. At first glance, this seems like a good thing. I do see some problems with it though:

  • The name: it suggests that either some authorities control the page it is listed on (false), or that the page has been controlled by checking it with these authorities (also false). What it does is list some authorities which could be used to perhaps control some aspects of the page, if you are so inclined. Perhaps it should be moved to the talk page of articles instead of being on all these articles?
  • The contents: many of the links in the template add no value at all to the article, and would not be accepted for that page in the "external links" or "further reading" sections. Then why do we add them automatically through a template anyway? Some examples:
    • Marc Sleen, which already has 14 references and an external link, has 9 AC links. WorldCat is interesting, VIAF doesn't add much that WorldCat doesn't, LOC is useless here, ISNI gives "Your data limit has been reached." (first time I visit this!), Deutsche Nationalbibliothek is useless for a Flemish author on enwiki, IDRef can be removed as well, BnF is somewhat interesting, and Libraries Australia is again pointless.
    • Cromwell Dixon: Worldcat again interesting, VIAF not, LCC not really interesting either
    • Jan Van Eyck gets loads of AC links, including things like the wiki Musicbrainz, which gives us... a copy of the Wikipedia article[4]. Useless stuff again includes Libraries of Australia[5], some Swedish site[6], some Polish site[7], a Japanese site[8], ... many of which point to Wikipedia if you want more information. Added value? Zero. Added clutter? Too much.

If Authority control is supposed to be useful, it should be curated and much more restrictive. The blind posting of every identifier someone adds to Wikidata just because it is about the same person, but not because it adds anything at all for readers, is something that would not be allowed in normal editing, but is somehow acceptable because it is Wikidata-driven. Worldcat seems to be the only one that is consistently useful, all others seem to be dependent on the subject (e.g. Deutsche Nationalbibliothek should only be added to people with some link to German or the German language) or perhaps even never useful at all. As it stands, the disadvantages seem to outweigh the advantages, which are rather small to begin with. Fram (talk) 08:59, 16 January 2017 (UTC)

I wonder, it is easy enough to make a selection of the external identifiers that are included. My question to you is, do you understand why it is a good thing to link to these external sources. Do you appreciate that what is enough for you may be too little for someone else? How do you find what it is that is useful. When the data is available it could be a personal option to increase or restrict what is shown. One caveat, the caching involved. Thanks, GerardM (talk) 12:07, 16 January 2017 (UTC)
As far as I can see, I can't make a per-page selection of which identifiers to include (e.g. leave DNB for German-related topics but remove from others). I think I have well explained why it is a good thing to link to some of these external sources, on a case-by-case basis, and why some others are not a good idea. Feel free to indicate which of the ones I didn't find useful in these examples would be useful for someone else on enwiki, and why. "How do you find what it is that is useful." Well, that's what we did for 15 years (and still do) in our external links and further reading sections. We decide this on a page-by-page basis, with talk page discusion if needed (though such a discussion is rarely needed); we don't dump the same indiscriminate list of reliable but in many specific cases useless links to all 500,000 pages with the same template. At least enwiki still gets to decide which entries get on the indiscriminate list and which don't, so we don't have to put up with really crap links liks Quora. But even so, getting an ID which says "yes, the subject exists, for more information check Wikipedia" will hardly be useful for anyone. Fram (talk) 13:04, 16 January 2017 (UTC)
"same indiscriminate list of... links to all 500,000 pages with the same template" This is not what the template does. Yet more bullshit. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:53, 16 January 2017 (UTC)
Then please give some examples of articles which don't do this. Don't just tell people "you're wrong", show it. As far as I know, the template has a list of what, 20+ possible links, and every one of them gets added to the article if a value is given in Wikidata, no matter how useful or useless that link may be for our readers (and no, this isn't just "personal preference", a link which gives nothing but a Wikipedia page is not a useful or even acceptable link). Fram (talk) 14:19, 16 January 2017 (UTC)
So, as you now explain, not the "same indiscriminate list of... links to all 500,000 pages with the same template". Your "nothing but a Wikipedia page" is yet again FUD. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:34, 17 January 2017 (UTC)
You keep bleating 'FUD' but dont actually provide any rebuttal of substance. If there is any FUD coming here its from your end. Only in death does duty end (talk) 10:42, 17 January 2017 (UTC)
Indeed. I gave the example of Jan van Eyck and his MusicBrainz link above. This is "nothing but a Wikipedia page". None of the tabs give any additional information at all. Oh wait, yes, it has a link to discogs.com[9] (in itself not a reliable site either): but the information on this page is not about Jan Van Eyck, the 15th century painter. The person who is meant here is Jacob Jan Van Eyck instead. So you get not just a Wikipedia page, but a Wikipedia page and, if you look carefully, a link to an incorrect Discogs page. Why should I welcome this link in any way? Fram (talk) 15:43, 17 January 2017 (UTC)
WP:EL. Only in death does duty end (talk) 13:23, 16 January 2017 (UTC)
":it suggests that either some authorities control the page" More FUD. Read the help page to which that phrase is linked in every instance of the template (and note that the phrase, our use of it, the template, that link, and the linked help page, all pre-date Wikdiata). "the blind posting of every identifier someone adds to Wikidata" this does not happen. More bullshit. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:51, 16 January 2017 (UTC)
At the template, we control (for now) which of the identifiers get used, and which don't. For all articles that have the template, indiscriminately. We don't only use Musicbrainz for music-related articles, DNB for Germany-related articles, and so on. If I add a musicbrainz identifier to Wikidata for Margaret of Valois, then that will automatically be shown here, without any change to the page history, and without any benefit for the readers. Ridiculous? Not really, such an ID exists[10]. Fram (talk) 14:16, 16 January 2017 (UTC)
"for now" More FUD. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:26, 16 January 2017 (UTC)
Thanks for your constructive criticism and helpful examples. Fram (talk) 19:08, 16 January 2017 (UTC)
See authority control. It is the standard term for efforts to assign unique reference identifiers or standardised indexing forms for things that may have many names. Jheald (talk) 13:01, 17 January 2017 (UTC)
Thanks. This explains why the name was chosen. I still think it is a bad name for the user-facing side of it (for readers, it is interesting that we have links to other, reliable insitututions with information about the subject: it is for most readers not interesting that abbrevation X uses ID Y for the same subject, and that this ID-giving is called "authority control"). Fram (talk) 15:43, 17 January 2017 (UTC)
  • So I spent 15 minutes trying to use the Authority Control template at Jan Van Eyck to exclude some links (which are inappropriate). If its possible I dont see how without removing the template entire, as there is no alternative (unlike infobox person). I can manually set them, but then it flags up big red errors. This is obviously a problem (not so much for a dead artist) on articles where the ACs may be incorrect or link to a different person, or just be completely useless links that shouldnt be there per WP:EL. Only in death does duty end (talk) 17:30, 23 January 2017 (UTC)

Fake news

I've just reverted an edit which made the bold - and, ironically, unsubstantiated - assertion that "Wikidata edits violate WP:V and WP:BLP.". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:28, 12 January 2017 (UTC)

And I've now reverted an attempt to re-word that as "The lack of reliable sourcing means that imported Wikidata text violates WP:V and WP:BLP.", which is still bunkum. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:34, 12 January 2017 (UTC)
Actually, it's one of the bigger problems, and I'm really sorry that you don't understand that. I note on reading this page, and many other Wikidata-related discussions, that individuals who are strong proponents of using Wikidata information in articles do not seem to give any weight whatsoever to the editorial policies of the recipient projects. Maybe the editorial community of a small wiki doesn't care and is more interested in getting *any* data regardless of whether or not it's even true (I've seen lots of Wikidata entries with bunkum in them). This is not one of those projects. We cannot afford to be one of those projects. We're the one that makes international headlines for incorrect information. Risker (talk) 15:32, 14 January 2017 (UTC)
"...you don't understand that" Bullshit. "individuals who are strong proponents of using Wikidata information in articles do not seem to give any weight whatsoever to the editorial policies of the recipient projects" Also bullshit. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:28, 14 January 2017 (UTC)
You have made it crystal clear that you do not in any way perceive this to be a problem, both here and elsewhere on this page, and I don't understand your rather extreme denial that there are BLP problems. I would identify you as one of the strong proponents of Wikidata on this project, and you seem to be denying there are any BLP and WP:V problems, while it is repeatedly demonstrated that there are. I think you would likely characterize yourself as being a strong proponent, and you're not giving it any weight. Risker (talk) 01:39, 15 January 2017 (UTC)
"your rather extreme denial that there are BLP problems... you seem to be denying there are any BLP and WP:V problems". Really, Risker, we expect this kind of rubbish from the usual trolls on Wikipedia, but you used to know better. Shame on you. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:30, 15 January 2017 (UTC)
Andy, having personally reverted Wikidata information that was indeed not only unverified but completely erroneous, at least on one occasion involving a living person for whom the error was particularly problematic, I will stand my ground on this one. I expect better from you too, Andy. These are core policies of English Wikipedia. If you're not on board with them, I don't know what to say. Risker (talk) 03:24, 16 January 2017 (UTC)
I too have personally reverted Wikidata information that was indeed not only unverified but completely erroneous. Such anecdotes are utterly irrelevant to the point at hand. "core policies of English Wikipedia. If you're not on board with them..." More bullshit. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:46, 16 January 2017 (UTC)

Request to fix my own post

I've tried three times to reduce my own post to bullet format without a signature, [11][12][13] but Pigsonthewing has reverted me. Could someone do that for me, please? It's my own post, so I should be allowed to write it as I want to; it's in the "disadvantages" section, where people were asked to list perceived disadvantages.

I would like it to say: "The lack of reliable sourcing means that imported Wikidata text violates WP:V and WP:BLP." SarahSV (talk) 23:45, 12 January 2017 (UTC)

As I thought I'd made very clear above, I would object most strongly to such a falsehood being included here. See also the recent comment by User:RexxS. As for "your own post", the page referred to is a communal page; see WP:OWN. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:58, 12 January 2017 (UTC)

Request to fix my own post (2)

I've tried three times to reduce my own post to bullet format without a signature, [14][15][16] but Pigsonthewing has reverted me. Could someone do that for me, please? It's my own post, so I should be allowed to write it as I want to; it's in the "disadvantages" section, where people were asked to list perceived disadvantages.

I would like it to say: "The lack of reliable sourcing means that imported Wikidata text violates WP:V and WP:BLP." SarahSV (talk) 23:45, 12 January 2017 (UTC)

As I thought I'd made very clear above, I would object most strongly to such a falsehood being included here. See also the recent comment by User:RexxS. As for "your own post", the page referred to is a communal page; see WP:OWN. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:58, 12 January 2017 (UTC)
I'm unclear as to why my comment was copied from the above section, into this one. Most irregular. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:06, 13 January 2017 (UTC)

And there goes the attempt to list the positions of both sides as they perceive them. Pigsonthewing, there are and will be "falsehoods" included in both the benefits and disadvantages section. If we only include things everyone can agree upon, the list will be very short and not useful at all. The (obviously badly failed) intention was to get to know the opinions, the perceived reality, which could then (here, or in a separate section) be discussed (politely, not by insulting everyone who dares to have a negative opinion of Wikidata). List what people see as the benefits and disadvantages of Wikipedia, not what the "real" benefits and disadvantages are, since no agreement on such a "real" list will ever be found. The "uses" and "discussions" sections should be factual (but there as well some Wikidata-promotors insist on adding "but it will improve" and "look what Wikidata can do" comments), but the disputed section not, as it can't be factual and useful at the same time. Fram (talk) 08:25, 13 January 2017 (UTC)

Yes, I think this dispute has occured because people don't know whether to treat the page as a project page (in which WP:CON applies) or a talk page (in which WP:TPG applies). I think it would be simpler if it were a talk page, but is it feasible to convert to such now? — Martin (MSGJ · talk) 08:51, 13 January 2017 (UTC)
Perhaps you could avoid straw men like "insulting everyone who dares to have a negative opinion of Wikidata"? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:09, 13 January 2017 (UTC)
There is a difference between slight hyperbole and straw men, you know? The number of insulting replies to anything somewhat negative about wikidata on this very page is staggering. Fram (talk) 09:30, 13 January 2017 (UTC)
Yes, I know the difference, which is why I chose my phrasing so carefully. Please give examples of these perceived insults. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:43, 13 January 2017 (UTC)
Then you chose carefully but wrong. As for insults, things like "It appears to have become merely a vehicle for some editors to regurgitate half-baked slogans and ridiculous inventions as if they were Gospel." perhaps? Fram (talk) 09:57, 13 January 2017 (UTC)
No. "Half baked editors" would be an insult; "half baked slogans" criticises - clearly - the slogans. Otherwise, your "cult-like behaviour" canard would be an insult, too. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:02, 13 January 2017 (UTC)
You don't seem to know the difference between "insults" and "personal attacks" (just like you don't know the difference between hyperbole and strawmen, and between reasoned arguments and I-can't-hear-you dismissals). The quoted sentence criticizes edits, but in an insulting way. "Criticism" and "insult", just like "source" and "wikimedia project", aren't mutually exclusive. Fram (talk) 10:19, 13 January 2017 (UTC)
It seems reasonable and logical to deduce from your argument that you intended your "cult-like behaviour" comment to be insulting. Should you wish to deny this, please explain the inconsistency of your argument. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:38, 13 January 2017 (UTC)
Obviously. It was reciprocal for "as if they were Gospel". An eye for an eye and all that nonsense. Fram (talk) 12:43, 13 January 2017 (UTC)

Can we please all agree to let people add their own idea of the benefits to the benefits sections, and of the disadvantages to the disadvantages section, without anyone else interfering with it? If you prefer, we can ask that everyone signs their entries to make it clear that it is the position of that person and not an "official" enwiki position. If you want to discuss these entries on the main page, start a new section at the bottom; alternatively, dissect any entries you want here to your (polite) heart's delight. Fram (talk) 08:25, 13 January 2017 (UTC)

While you're removing or altering posts with edit summaries like "Removed pro-wikidata POV" and "please keep your POV out of this"? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:12, 13 January 2017 (UTC)
Yes, these were in the "uses" section, not in the benefits or disadvantages section. Whether a particular use is "The ideal for Wikidata adoption."[17] is POV and doesn't belong there. Fram (talk) 09:30, 13 January 2017 (UTC)

Pigsonthewing, please don't add incorrect "unsigned" templates. The comment above was mine, not SlimVirgins. Fram (talk) 09:25, 13 January 2017 (UTC)

Redux

I see that this false claim has been reinserted. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:46, 18 January 2017 (UTC)

You mean the SimVirgin line? Let's see who last readded that line... [18] Hmm, a certain Pigsonthewing apparently. Ever heard of him? Fram (talk) 08:13, 19 January 2017 (UTC)
Apologies, I'd overlooked that it was the signed version. The whole page is a mess. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:11, 19 January 2017 (UTC)

We now have - despite my best efforts to make it actually say something factual and useful - an unsigned claim that " Wikidata edits/transclusions (where data is included from Wikidata without a reliable source being provided on ENWP) violate WP:V " As previously explained, this is still utter bollocks. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:47, 23 January 2017 (UTC)

WP:V requires that all information on a wikipedia article is verifiable by use of a reliable source. When data is transcluded from wikidata into an article, and there has been no reliable source provided, it is unverifiable. 'The data is sourced at wikidata' does not satisfy the burden required by WP:V. Especially since Wikidata a)has lax sourcing requirements itself, b)contains data from other wikis with equally different standards. While everything does not require an inline citation, it does require that a reliable source be used in the first place. Wikidata transclusions do not satisfy that except where the data is already sourced in the article or where someone manually adds the sourcing after. The first is sometimes true, the latter is rarely true. Only in death does duty end (talk) 17:55, 23 January 2017 (UTC)
Where did I ever claim "'The data is sourced at wikidata' satisfies the burden required by WP:V"? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:36, 23 January 2017 (UTC)
Strange, the policy which directly supports this has been quoted and explained to you. And of course, all claims on the page are "unsigned", not just this one. That's standard procedure. That you don't agree with a claim doesn't mean it should be removed, certainly not when it is a perceived claim in the first place, and one that isn't even subjective but actually based in policy secondly... Fram (talk) 17:57, 23 January 2017 (UTC)
It is not that I "don't agree" with the claim, is that it is over-reaching, and demonstrably false, and unsupported by both policy and agreed practice, as I have shown elsewhere on this page. That both you and Only in Death ignore that does not negate it. The "perceived" heading is a cop-out. We no more need to cater for bogus perception here than we do on pseudo-science articles. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:36, 23 January 2017 (UTC)
We don't ignore it, we don't agree with you. My reading of policy supports the disadvantage as written, your reading of policy somehow doesn't. That has little to do with pseudo-science. Do not use articles from Wikipedia (whether this English Wikipedia or Wikipedias in other languages) as sources. Also, do not use websites that mirror Wikipedia content or publications that rely on material from Wikipedia as sources. seems pretty clear to me. The only "cop-out" some people use is to claim that Wikidata is not a source somehow, and thus this policy rule is not violated. Which, to us your words, is bogus. Fram (talk) 07:48, 24 January 2017 (UTC)
Once again, you grossly misrepresent what I have said. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:03, 24 January 2017 (UTC)
Which part of my reply did this? "You are wrong" is hardly helpful without indicating what exactly is wrong. Fram (talk) 11:12, 24 January 2017 (UTC)
That would be quite a feat since despite asking multiple times you have yet to put forth a coherant argument other than 'its bullshit, fud' etc. From your rather uninformative opposition I have come to the conclusion you dont think content added on ENWP articles that has been transcluded from Wikidata counts as 'sourcing the data from wikidata'. You see wikidata as a middle-man where all the information is actually sourced from somewhere else. Where myself, Fram (and to a lesser extent Risker/SV above - correct me if I am wrong) see wikidata as the source from which the material originates - as that is where it comes from when it is used - regardless if it is ultimately referenced/originated on wikidata from elsewhere. This wouldnt be a problem if 100% of wikidata was reliably sourced according to our policies, but since barely 25% of Wikidata has a source that could be considered at all (let alone hit our reliability markers) and functionally even when there are sources available at wikidata, these are not used in ENWP at all because the material displayed/incorporated on ENWP is done absent references. Only in death does duty end (talk) 11:15, 24 January 2017 (UTC)
"you have yet to put forth a coherant argument other than 'its bullshit, fud' etc." Now that is simply a lie. Desisst. Having refuted the claim, it is not for me to do so again, every time you or Fram chose to ignore that. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:21, 24 January 2017 (UTC)
The closest you came to actually explaining anything is now in Archive 2 where you basically said that not *all* wikidata edits (interwiki links etc) violate the above policies. Since no one was actually arguing that in the first place it was rather a weak argument and still didnt address the issue of content in articles. I have now trawled through both archives (and on this page) trying to find an edit by you where you clearly state why it is wrong. Other than 'its false, lies, etc'. If you have done so, please correct me and provide a diff or I shall work on the assumption you have not. Only in death does duty end (talk) 12:49, 24 January 2017 (UTC)

Atomising of article writing

From my experience with editing Wikidata and en-Wikipedia, there is an element of feeling that rather than pulling together sources to build an article, information is being broken down into data elements for the purpose of putting it back together again later (i.e. machine writing).

Anyone who has written an article on Wikipedia (by which I mean a properly sourced and reasonably standard-length article) knows that this involves accessing different sources of information and distilling it into words and presenting this information. The information present can be broken down into discrete elements and carefully marked up and filed away as data (this is best seen in the construction of infoboxes and the metadata associated with references), but there is a balancing act between doing this and maintaining the flow of writing an article and understanding a topic.

What I am trying to say is that the process of processing and handling the underlying data can at some point go too far. There is, for want of a better word, an intuitive process that involves a human accessing a set of source(s) and transferring the information into an article (with references). Wikidata sometimes seems to try and make that intuitive process more robotic in nature, which doesn't really work. I am going to repeat what I said elsewhere:

"There needs to be a separation between database maintenance and article writing, and an interface between the two that is intuitive rather than opaque." (Wikipedia talk:Wikidata/Archive 4#Is this not a form of machine writing?)

Wikidata is great for organising data. When people then try and use this data for generating articles (rather than lists or discrete items of information), that starts to rub up against humans trying to put words together as language. Wikidata should be a resource, to be used by people writing articles, but it shouldn't be used to write articles in rote or automated fashion. Carcharoth (talk) 11:43, 24 January 2017 (UTC)

That atomisation of data is already being done when adding categories to the article, with the difference that the category selection cannot be sourced and the wikidata statement can be sourced. In fact many statements in Wikidata were added because the article was included in a given category.
I don't know why do you mention automatic writing of articles, afaik nobody suggested writing automatic articles, although some displaying of data from Wikidata is already being done (as an opt-in) by some small Wikipedias by using the Article placeholder extension (example: nn:Special:AboutTopic/Q845189 - Spotted turtle).--Micru (talk) 13:09, 24 January 2017 (UTC)
Article placeholders = automatic writing of articles. Listeriabot = automatic writing of articles. And I have seen (and linked here, I thnk) calls to use Reasonator to create articles as well. Fram (talk) 13:28, 24 January 2017 (UTC)
I don't understand how you can call displaying data "writing an article" when it is clear that it doesn't follow the same format as a written article. Reasonator does generate some short texts, but from that to saying that it "writes articles" is a stretch. Btw, one question, why is it requested that Wikidata generated lists should be individually sourced, when in general enWP lists are sourced globally?--Micru (talk) 13:50, 24 January 2017 (UTC)
You and I would perhaps know that the different format indicates that it is not an article. Our readers wouldn't. They go to a page in the mainpage and see information on it. For them, this is a Wikipedia article. As for Wikidata lists, how would you technically collect different items (different pages) based on a common property, but source them globally anyway? For me personally, one source to rule them all would be as good as one source per line, as long as we could be reasonably sure that every item came from that source. But just having e.g. a source "CIA World Factbook" at the item "capitals", and then assuming that all capitals are sourced from there on the individual pages, would be problematic (see e.g. the Sumatra example above). Fram (talk) 14:02, 24 January 2017 (UTC)
Regarding readers not knowing that a page is ArticlePlaceholder generated, there is T124191, which I hope it will solve it.
As for lists, in theory it should be possible to assign the same source to several statements at once. Perhaps what it is needed is a better interface for "mass-sourcing" items in a list.--Micru (talk) 14:28, 24 January 2017 (UTC)
Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia_talk:Wikidata/2017_State_of_affairs/Archive_3&oldid=763078047"