Wikipedia talk:Good article criteria

From Wikipedia, the free encyclopedia
Jump to: navigation, search
ArchiveThis page, a part of the Good article talk page collection, is archived by MiszaBot II. If your discussion was mistakenly archived feel free to go retrieve it.Symbol support vote.svg
Current Archive location: Wikipedia talk:What is a good article?/Archive 3

Stand-alone lists being nominated as Good Articles[edit]

--Redrose64 (talk) 22:58, 1 December 2016 (UTC)

Note that one of the three proposals in this RfC involves creating a separate "Good Lists" rating and process independent of the GA process. BlueMoonset (talk) 18:17, 5 December 2016 (UTC)

Readable prose threshold[edit]

Hi, looking over the editorial assessment criteria, I feel like good articles should have some guideline that dictates a general threshold of kilobytes of readable prose. For example, one of the criteria for C class is that the article be substantial in length, thus it follows that good articles should also be substantial in length. Would having a guideline for readable prose character count be considered acceptable? For the sake of argument, say 25 kB? Note that the example GA in the editorial assessment criteria is 36 kB readable prose. (Disclaimer: I've never nominated or reviewed a GA. I want to get into the process soon; in the meantime, this was just a thought that crossed my mind). Looking forward to knowing what everyone thinks. Icebob99 (talk) 02:00, 8 December 2016 (UTC)

I would like to add that in the case of articles like Morocco at the 2014 Winter Olympics, which gives a complete coverage but is only 5.5kB readable prose, the threshold could be waived. Icebob99 (talk) 02:03, 8 December 2016 (UTC)
  • Comment I think #3 (Broad in coverage) covers this well enough. Different articles will require different lengths to cover major aspects and arbitrary length requirements can create worse articles. I have always felt is better to keep these criteria open as much as possible and trust the judgements of the reviewers. AIRcorn (talk) 02:30, 8 December 2016 (UTC)
  • According to PetScan, there are over 9,000 good articles with less than 20 kb total article size (much less prose only), so prose minimum of 25 kb would be out of line with actual practice and could require re-evaluation of thousands of articles. --RL0919 (talk) 16:51, 8 December 2016 (UTC)
    RL0919 didn't know that stat. Presumably, some articles will have a lot less material than Roman Empire, just by nature of importance. I thought my example mentioned above (Morocco at the 2014 Winter Olympics) would only cover the exceptions, but obviously these 20kB or less articles have a pretty substantial majority. Icebob99 (talk) 17:38, 8 December 2016 (UTC)
  • Comment - Coming up with a count of "kilobytes of readable prose" might be difficult without a tool to generate that. Perhaps Icebob99 meant "characters" instead of "kilobytes". One could do an article of two sentence of prose, with the rest images, templates and an infobox, and still generate a sizeable article if all that counts is kilobytes. — Maile (talk) 17:22, 8 December 2016 (UTC)
There are a couple of tools available for calculating prose size, for example User:Dr pda/prosesize. --RL0919 (talk) 17:27, 8 December 2016 (UTC)
Interesting tool. — Maile (talk) 17:29, 8 December 2016 (UTC)
(post EC) Maile66 Sorry, that was an unclear wording, I meant characters of readable prose. I got the idea from the Africa destubathon criterion of 1.5kB readable prose. Readable prose according to the Dr pda tool doesn't include iboxes, templates, lists, or images. Icebob99 (talk) 17:38, 8 December 2016 (UTC)
  • Comment I'm going to withdraw my idea/proposal based on what RL0919 brought up. I think we can all agree that we want good articles to be nice and long so that they cover everything broadly, but given the difference between the topic depths of many articles, I think a threshold would not be very useful unless it were to contain ridiculously complicated exceptions, which defeats the purpose of simplicity that a threshold would have. Icebob99 (talk) 17:38, 8 December 2016 (UTC)

Proposal for 2e[edit]

Our criteria require that reviewers check all the listed sources to verify that they support the assertions. However, I'm concerned that many reviewers are skipping this altogether. Number 2 in the criteria has a link to WP:V which says: "The cited source must clearly support the material as presented in the article". It's not enough for a reviewer to see that there is a citation, but that the cited source actually has to be examined. Our criteria about verifiability currently has a note which says (in part): "At a bare minimum, check that the sources used are reliable (for example, blogs are not usually reliable sources) and that those you can access support the content of the article". My proposal is to promote this note into criteria 2e (or maybe make it 2a and move the others down the list). I recommend the following:

All of the sources checked substantiate the claims of the content they support. The reviewer is expected to check a majority if not all of the sources cited, with special attention to every quote and every statement likely to be challenged.

The attached note would then be moved to this new line and would then read:

Ideally, a reviewer will have access to all of the source material, and sufficient expertise to verify that the article reflects the content of the sources; this ideal is not often attained. At a bare minimum, check that the sources you can access support the content of the article (for example, inline citations lead to sources which agree with what the article says) and are not plagiarized (for example, close paraphrasing of source material should only be used where appropriate, with in text attribution if necessary).

I'd like to hear the consensus view on this. GA is an important step on the way to A-class and FA and I think this new wording would reinforce our existing standards. Chris Troutman (talk) 14:07, 17 January 2017 (UTC)

Maybe also mention that reviewers should check if the sources actually exist.--3family6 (Talk to me | See what I have done) 04:27, 18 January 2017 (UTC)
  • Question—I'm not seeing how what is being proposed above is different than the requirement already in 2b:

    all in-line citations are from reliable sources, including those for direct quotations, statistics, published opinion, counter-intuitive or controversial statements that are challenged or likely to be challenged, and contentious material relating to living persons—science-based articles should follow the scientific citation guidelines

  • @Chris troutman: Could you clarify? Grondemar 04:44, 18 January 2017 (UTC)
    @Grondemar: I want to specify what WP:V already says: that the citations have to actually be checked by the reviewer to ensure the citations support the content. This is already required as criterion 2 is verifiability. I think the wording I suggest makes this requirement clearer. Chris Troutman (talk) 13:37, 18 January 2017 (UTC)

"In addition to meeting the policies regarding content for all Wikipedia articles"[edit]

Per discussion many years ago on the featured article criteria (wish I could pinpoint the actual convo in the short time I have right now), not all content policies are outlined in the criteria. It's mostly for obvious reasons (e.g. the criteria would become bloated). A GA should adhere to all content policies, even those not included in the criteria (WP:BLP, WP:NOT, WP:TITLE, etc.). —Deckiller (t-c-l) 06:04, 23 February 2017 (UTC)

I would like to suggest that if this is added, it be added either as a seventh criterion or as a line below the list of six, rather than an a priori (and highly duplicative) content policy list, since the bulk of the criteria do duplicate these policies. If on a line below, then "In addition, the article is expected to meet the policies regarding article content not already mentioned above, including on biographies of living persons, what Wikipedia is not, article titles, etc." If as a seventh criterion, then perhaps "7. Meets content policies: it meets all policies regarding article content not already mentioned in the other criteria, including on biographies of living persons, what Wikipedia is not, article titles, etc." BlueMoonset (talk) 18:44, 24 February 2017 (UTC)
[edit conflict] BLP and TITLE are unobjectionable. NOT is a big can of worms that I would prefer not to import wholesale here. And more generally, Wikipedia appears to have over 100 policies, judging by the size of Category:Wikipedia policies and its subcategories. Do we really want to demand that each GA reviewer be familiar with the nuances of and individually check each one of these? Or do you only mean the ones actually listed in Category:Wikipedia content policies? —David Eppstein (talk) 18:46, 24 February 2017 (UTC)
In 2006 or so, we added this line to the Featured Article criteria, because people were arguing that things like title policy and BLP did not apply to their FACs (or something along those lines):
"In addition to meeting the policies regarding content for all Wikipedia articles, it has the following attributes."
Basically, it was just a way to cover all bases in the criteria; it's honestly not that big of a deal. I think BlueMoonset makes a really good proposal. The criteria are transcluded everywhere, so it makes sense to incorporate it as such. —Deckiller (t-c-l) 00:42, 25 February 2017 (UTC)
  • Precisely what problem are we seeing that needs this solution? WP:CREEP seems to make this a poor idea. If there is a specific policy that is being missed, let us add it explicitly, rather than saying "and do everything else perfect, too". Jclemens (talk) 05:41, 25 February 2017 (UTC)
There are two things seriously wrong with this: (1) it duplicates essentially all of the existing criteria, as "CONTENT" isn't a criterion in itself, and worse (2) it is a bucket class (other people call this "etcetera", "miscellaneous", "dustbin") to which other criteria may be added at any time, introducing the risk of continuous creep (or confusion about it) into the GA criteria. Much better would be to add specific criteria, if indeed we need any more than we have already. Chiswick Chap (talk) 20:00, 26 February 2017 (UTC)
The primary reason I reverted Deckiller's original edit was related to Chiswick Chap's first point: it was not only a duplication, but in its addition to the preamble, repeated what we already had plus added a whole, indefinable extra group of stuff. Stated in many documents, including WIAGA itself, are GA's six criteria (a reason why adding a seventh was listed last in my original post here), but I wasn't endorsing this, just proposing a way to put any new criteria as an unnumbered addendum. It's a good point that this is WP:CREEP in two ways: one, by adding all the content policies, and two, that they can be expanded at any time.
It always seemed to me that the GA criteria had been deliberately constructed to take those things that were appropriate to a Good Article, and omit those things that weren't necessary. Hence essays like WP:GANOT, which explains what the criteria are not. Interestingly, that essay points out that WP:TITLE is beyond the scope of the criteria. This is not to say that an article's title won't come up during a review, since it sometimes does. An article can be moved there and then if nominator and reviewer agree, or there can be a disagreement over whether the move is appropriate. In the latter case, a move request is then made, separate from the GA review, which is ill-equipped to decide the issue—the ultimate name of the article is decided under those auspices. I can't count the number of times that a GA review page has been seemingly orphaned by an article move, and I've had to move the review page so it mirrors the new article name. So I'm not sure that this is a good criterion to add, though we might want to point out in the various GA process documents that the reviewer should feel free to discuss potential moves with the nominator and initiate a move request separate from the review.
WP:BLP doesn't seem to fit in anywhere among the criteria, except for the part of the verifiability criteria that touches on it, all in-line citations are from reliable sources, including those for ... contentious material relating to living persons. If we do wish to mandate this, perhaps it can be added as I suggested above. BlueMoonset (talk) 03:39, 28 February 2017 (UTC)

Fair enough. I'm fine with keeping the line out of the page if users find it unnecessary. GA process is different than FA; times have changed as well. I think the only point left to address is possibly incorporating WP:BLP into the criteria as a precautionary measure (per BlueMoonset above). —Deckiller (t-c-l) 04:14, 1 March 2017 (UTC)

Disallow deadlinks in GAs[edit]

As stated in the GAN talk page, it seems a bit silly to explicitly allow deadlinks for GA's. At a minimum, we should be following WP:DEADREF. While this may happen in practice, it is not explicitly stated in the criteria. As I read the criteria as written now, the URL can be dead as long as it is not a barelink. Thought? Pinging @Bungle and BlueMoonset: since they were in the discussion on the other page. Kees08 (talk) 06:00, 9 April 2017 (UTC)

Agree with Kees08. For what it's worth, this issue was decided on 2013 RFC. It's referenced to WP:GANOT See "Mistakes to Avoid" under section "Factually accurate ... " I would suggest that if this is changed in Footnote 5 of the Criteria, then other GA instructions that refer to it should also be changed. — Maile (talk) 12:20, 9 April 2017 (UTC)
  • This will mean that many old GAs would now fail the criteria. Repinging BlueMoonset as fixed pings only work if you sign them. AIRcorn (talk) 17:25, 9 April 2017 (UTC)
    • Yes, and I think those articles should have the links repaired, or a new source. Kees08 (talk) 17:33, 9 April 2017 (UTC)
Actually, that might not be as many as one would think. I've seen a bot recently checking for dead links and repairing them if possible. Is it the GreenC bot ? Maybe we could ask User:Green Cardamom about this. — Maile (talk) 17:51, 9 April 2017 (UTC)
There are two bots currently operating: WaybackMedic by myself and IABot by User:Cyberpower678. IABot is the primary dead link tool for Enwiki and can be accessed from the History tab "Fix dead links" .. anyone can run it for a article on demand. WaybackMedic is a more specialized tool for difficult to solve dead links and not available on-demand. IABot will convert bare links to CS1|2 etc.. try it out on an article to see what happens. -- GreenC 18:50, 9 April 2017 (UTC)
To add on to that, you can also queue the bot on a list of pages with this tool.—CYBERPOWER (Chat) 18:54, 9 April 2017 (UTC)
I don't think any amendment as proposed should be retrospective for existing GA articles that would, under the proposed amendment, fail the core criteria. The task of doing that would simply be immense, though i'd like to think that many GA reviews over time would have picked up on dead links and either had them addressed or worked around anyway. I do support a change in the criteria that would explicitly disallow ANY dead link where there is no clear alternative given, or at the absolute very least, where an archive link isn't provided. I am not a huge fan of archive links anyway, but it's certainly better than not being able to verify anything associated with the respective reference(s). Bungle (talkcontribs) 18:41, 9 April 2017 (UTC)
Maile, so what you were originally saying is that per the 2013 RFC, dead links are considered verifiable so long as they aren't bare URLs? (That's what I got from the closure I read at the link there.) That would seem to agree with the current criteria, but disagree with Kees08, but you say you agree. Can you please clarify? Thanks. BlueMoonset (talk) 18:46, 9 April 2017 (UTC)
BlueMoonset I agree that it's policy as defined by the 2013 RFC that dead links are considered verifiable so long as they aren't bare URLs. I also agree with Kees08 that it should be changed. Any new policy about this should be applicable to new reviews being done, not to what is already out there. Just change the policy for new reviews. — Maile (talk) 19:18, 9 April 2017 (UTC)
Maile, how can you possibly have one policy for new reviews and another for existing GAs? A GA is a GA: either dead links are okay or they're not. If you could state for sure that a link was alive at the time the GA was listed and therefore it was presumed to have been checked (a dangerous presumption), then I might see it—if explicitly stated in the criteria—but I can guarantee you we'll see reassessments popping up if we simply disallow them. A reassessment has to be done according to the GA rules at the time of the reassessment: does this article meet the GA criteria as they are now? BlueMoonset (talk) 19:58, 9 April 2017 (UTC)
BlueMoonset Well, let's not overthink this. You could say the same about any GA criteria that exists. Are there many GA re-assessments requested now? I don't really want to get into this issue too much, because I realize that GA policy is not necessarily identical to other review standards. It's just different, that's all. But in an ideal practice, anyone who nominates an article for anything ought to make sure the citations are correct and any links are live, at the time of nomination. — Maile (talk) 20:16, 9 April 2017 (UTC)
Hey, guess what - that new Fix deadlinks tool on the article's history page is really easy. I just ran it at John J. Pershing to try it out. Neat. Simple. — Maile (talk) 20:21, 9 April 2017 (UTC)
  • Treatment of dead links needs more nuance and consideration of context than simply a blanket declaration of acceptable vs. unacceptable. If a dead link is used as the only source for something that requires sourcing per the GA criteria, and not enough information is provided to rescue the material, then we have a problem that could be reasonable grounds for objecting to GA status. But there are many dead link situations that don't rise to that level: if the material has multiple sources, if the link is a convenience link to a source also available offline, if the material doesn't require citation under the GA criteria, if an archived copy is available, etc. There are also cases where material has more pressing problems than a dead link, such as obviously unreliable sources, inclusion of irrelevant detail, POV material, etc. The current footnote attempts to call out the most egregious problem of dead links that lack key information, without trying to legislate every judgment call about more complex scenarios. --RL0919 (talk) 19:27, 9 April 2017 (UTC)
    • Right, and all those situations are covered at WP:DEADREF. Since it is already defined and clear, I do not know why we would not use that guideline as our requirement. Kees08 (talk) 03:41, 10 April 2017 (UTC)

@BlueMooset: How do I go about officially proposing this? I have never done an RfC before and do not know how, is that what I should be doing? Kees08(Talk) 05:56, 17 April 2017 (UTC)

"Quick fail" criteria[edit]

I have raised some concerns about the "quick fail" criteria here. Comments welcome there- it's best if this can be kept in one place! Josh Milburn (talk) 22:07, 3 July 2017 (UTC)