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

Copyvio/close paraphrasing check[edit]

During the discussion at Wikipedia:Did you know/Good Article RfC, I asked what exactly would be involved in the Copyvio/close paraphrasing check proposed there. The RFC has now been closed, with a direction that there should be such a check, but that further discussion is needed to sort out the details. What should be expected of a reviewer conducting the check? Monty845 21:47, 6 September 2013 (UTC)

  • Obviously, if during the course of a GA review, the reviewer detects copyvio, plagiarism, or close paraphrasing, they shouldn't pass the article till the issue is resolved. But that is already covered by the long standing requirement to respect copyright laws. What additional steps are expected now? Having a bot do an automated check? Randomly googling phrases in the hopes of finding something? Something else? Monty845 21:52, 6 September 2013 (UTC)
  • I pressed Gatoclass for a more specific version of the proposal during the discussion. We never quite got specific language out of him, but so far as I understand it, he wants plagiarism to be independently listed as criterion 1c instead as a part of 1a, which makes sense to me. It's more of a format change than a change in the actual criteria, or as Gatoclass put it, a change in emphasis. -- Khazar2 (talk) 22:11, 6 September 2013 (UTC)
  • I'm not sure precisely what is being asked here. Good article criteria are not decided by DYK any more than we tell DYK how to do their reviewing. Are DYK trying to tell us how to review all our nominations, just the nominations that end up at DYK, or is this checking to be done by DYK on GAs that end up at DYK? DYK is very much about box ticking. To give just two extreme examples, North Carolina Highway 2A which is probably larger than many DYKs could be checked fully for COPYVIOs, Ethanol fuel in Brazil which took a lot of effort to review can't realistically be checked fully for copyright violations. It would be ludicrous to write a tickbox procedure that could handle these two extremes. Pyrotec (talk) 22:20, 6 September 2013 (UTC)
  • To return to this Copyvio/close paraphrasing/plagiarism comment above, if I copy some thing out of a document that has copyright protection, without acknowledging that I've done so, then that is a COPYVIO. In contrast, Wikipedia, for instance, does not claim copyright so if I copy something from one wikipedia article to another without referencing it, that is plagiarism. So what? Is some boxticker at DYK now demanding that every wikipedia article that has material copied from another wikipedia article needs a formal citation; and also a formal plagiarism check at GA to verify that such an event has not happened. It makes no sense to me. It's not a FA requirement, neither does it seem to be a DYK requirement. I've reviewed GANs that were DYKs and none of them had formal citations for material copied or duplicated in other wikipedia articles. So why should DYK impose it on us? It really means that the DYK boxtickers regard (quite wrongly) that their checking is superior to theirs and that our standards have to be raised to theirs; or, that they don't know what they are talking about. In actual fact the DYK standards are far far lower than ours. Pyrotec (talk) 22:42, 6 September 2013 (UTC)
  • Smile when you call people that. Pyrotech, unattributed copying of Wikipedia content is a copyright violation. Notice how Wikipedia text is not PD, but CC-BY-SA (emphasis mine): You have to attribute, otherwise you are violating one or more other people's copyright. — Crisco 1492 (talk) 00:24, 7 September 2013 (UTC)
And there is a reason FAs have spotchecks for new FA writers / nominators. — Crisco 1492 (talk) 00:25, 7 September 2013 (UTC)
Thanks for that. In that case, there are articles classified as FAs that have copyvios since they were created by "copying and pasting" from other wikipedia articles. That is how some prolific editors produce articles and get them through GA and FA. They are well referenced and well sourced, since they were copied from other articles of at least GA-level. Pyrotec (talk) 15:57, 7 September 2013 (UTC)

(undent)I'm unclear as to what exactly is under discussion here. A mandate against copyvio is already part of the good article criteria, and reviewers should be performing at least a brief spot check for every article they review. As has already been pointed out, copying without attribution within WP is also a form of copyvio, and so is covered by the good article criteria. Copying from a PD source without attribution would fall afoul of the good article criteria sourcing requirements, so again, that is already taken care of. Copying within WP or from a PD source with attribution is neither plagiarism nor copyvio, although the text of many PD sources is inappropriate style-wise for inclusion on WP, and could go against the the prose requirements for GA. So, TL;DR version: I don't see what DYK wants that is not already in the criteria. If reviewers are not performing checks required by the criteria, then that is a problem, but it is a problem with the reviewing process, not the criteria. Dana boomer (talk) 00:45, 7 September 2013 (UTC)

  • My reading is the same as Khazar's reading: give more emphasis to the role of copyvio/plagiarism checks, make it even clearer, maybe even as "criteria 1c" — Crisco 1492 (talk) 00:49, 7 September 2013 (UTC)
But again, "More emphasis on copyvio/plagairism checks", what exactly are we talking about when we say a check? What would a GA reviewer who is faithfully complying with a criteria that requires such a check do to satisfy the requirement? My view on the old language, is that we should keep our eyes peeled for violations, which we might detect as a large one diff expansion of the text, become suspicious based on an inconsistent tone while reading through the text for the other criteria, notice when we are checking the references to make sure they are proper, and that there is no OR, and then check the images for proper licensing tags. But other then the image licensing check, those are all passive detections, noticing something out of place and investigating. What additional steps should be taken? Monty845 01:37, 7 September 2013 (UTC)
  • I should think requiring an active review (i.e. actually opening the sources and checking; FA does it [on all first time nominations, at least], and DYK reviewers should do it, and some GA reviewers do it [I do], but having it codified would be better). Smart copyright violators know how to change it enough to get past a cursory inspection. — Crisco 1492 (talk) 01:45, 7 September 2013 (UTC)
  • So, Crisco 1492 seems to be saying that DYK are seeking to impose a higher standard on GAN than FA have in place (I've only had experience twice at FA as a co-nominator). The statement was that "FA does it [on all first time nominations, at least]". As a reviewer I do that checking in part on electronic sources that are available (I can't do it on those I can't access) and on books and journals (that I have to hand) as part of the verifiable sources process checks. I don't do random checking with google or similar. Is Crisco 1492 making claims here that for every nomination he (and DYK and FA) reviews, the reviewer has access to every book, journal, newspaper, web site, subscription-only site (including foreign language editions where appropriate) referenced in the article and does a full copyvio, close copying check and plagiarism checks? In fact that is not sufficient, since the copyvio, etc, might have come from an undisclosed source. In point of fact, my two MSc thesis (I have two MSc each with a requirement for a 15,000 word thesis) had to be submitted in electronic form (as well as three bound copies) as both universities used "checkit" (I'm not sure about the spelling) to check for copyvios and plagiarism. I saw parts of the printout for my second one, my tutor admitted there were massive quantities of "false positives"; but it could only check what was publicly available in electronic format. I think I will go further, unless Crisco 1492 (and any other reviewer) has access to all the books, journals, newspapers, web sites, subscription-only sites (including foreign language editions where appropriate) cited in the nomination, full checking of citations is not being done, then neither am I (nor is DYK and FA). I think there is a great degree of uninstantiated claims going on here. DYK are setting themselves as examplars of best practice and seeking to impose these on GAN, when it is clear from my example above, full checks for this and for verifiable claims can't be done without all the sources cited (as well as those not cited) and are certainly not being done by those that claim to do so. To clarify "best efforts" need to be made in carrying out such checks, but anyone claiming to do these checks in full probably either has a "trivial length" nomination or is making claims that aren't justified. Just in case anyone disagrees, do a full copyvio, close copying, plagiarism and verifiable sources check on Ethanol fuel in Brazil and state the time that it took to do it. It would be even better if more than one editor edit those checks and a league table of times published. Pyrotec (talk) 10:42, 7 September 2013 (UTC)
  • I am not saying anything of the sort, and suggest that you try reading my reply again. I'm saying that Gato's proposal was more to make sure spotchecks (nowhere did I say full spotchecks) were done, something which is required at both DYK and FA. — Crisco 1492 (talk) 13:41, 7 September 2013 (UTC)
  • At 00:25, 7 September 2013 you stated "... And there is a reason FAs have spotchecks for new FA writers / nominators." and 01:45, 7 September 2013 you stated "FA does it [on all first time nominations, at least], and DYK reviewers should do it". So, FA do spot checks some of the time and DYK should do it. This is your first mention of Gato's proposal, but in fact Gato (as far as I can see) has made no proposals at GAN. What he might have done was to make proposals at DYK in respect of what he/they intended to impose in GAN. So on the basis of what you have said here there appears to be some doubt as to whether FA and/or DYK do spot checks all the time. So give or take some uncertainty, the proposals being imposed on GAN are spot checks. Pyrotec (talk) 16:15, 7 September 2013 (UTC)
  • First of all, nobody is "seeking to impose a higher standard on GAN than FA have in place": you have a fundamental misunderstanding of what is being suggested below, and seem to think "close paraphrasing" or "copyvio" checks require all sources to be scanned (as opposed to random spotchecks used everywhere else). Another of your quotes: "DYK are setting themselves as examplars of best practice and seeking to impose these on GAN, when it is clear from my example above, full checks for this and for verifiable claims can't be done without all the sources cited (as well as those not cited) and are certainly not being done by those that claim to do so." - Nobody has said anything of the sort except for yourself. And if you didn't know Gato made the initial proposal, then you haven't read the other discussion yet. — Crisco 1492 (talk) 16:31, 7 September 2013 (UTC)
  • Let us consider this in small blocks: there is a voluminous discussion at Wikipedia:Did you know/Good Article RfC, which I've made minor contributions to be, but I've not really followed. Secondly, these reassurances are rather "wooley": you state that "nobody is "seeking to impose a higher standard ...", but that is all. The whole of the discussions and decision making about what standards are required at GAN to check for COPYVIO/plagiarism took place at DYK, not at GAN (and that seemed to be intentional). I'm also well aware of your comment there I disagree with the suggestion that passing a GA review automatically means the DYK review will be easier; I've seen some very shoddy GA reviews which, if done at DYK, would have led this talk page to catch fire. — Crisco 1492 (talk) 23:16, 28 July 2013 (UTC) Yes there are crap reviews at GAN, we can delist them with another review, we sometimes ban editors. DYK have a system of captive labour, whereas GAN relies on volunteers, so perhaps we don't need to light fires. I also assume (without any knowledge) that having copyvio material on Main Page is rather embarrassing. It appears to be that much of the proposals are coming from people with a similar mindset and they seem to be coming from DYK. I don't really know Gato and I don't know Monty. They both seem to have done five GAN reviews each. Monty has already asked his question at Wikipedia:Did you know/Good Article RfC as long ago as 7 August 2013 (UTC) with deathly silence; and he has asked it again here with this rfc. There is also deadly silence on the purpose of these "suggestions", I'm tending towards the "conspiracy theory" approach, and that they are going back to DYK for approval, it's already on-file at DYK that there is to be another discussion in the near future to review the progress and the operation of the Main Page system. Pyrotec (talk) 18:22, 7 September 2013 (UTC)
  • Secondly, why am I the only one making these statements? I have to assume that either DYK don't know about them, or that the type of nomination that you get at DYK means that they are not needed. If that is the case why are DYK trying to run or manage our quality systems for copyvio checking, etc? You gave me a short lecture on copy violations, but then I've had the advantage of two days of university lectures on copyright violations and plagiarisms, once at Salford and once at Southampton, so I'm returning the favour. Perhaps my grammar was poorly presented, so I'll rephrase it. If you have a typical DKY or a small GAN with six perhaps, ten or even 20-ish citations, running a check on google and amazon might work and produce some "hits". I wrongly called it "tickit", what I should have written was turnitin. That is what some UK universities use to check student dissertations. The use of a professional fully automatic checker such as turnitin produces a massive quantity of "false positive hits", and its also a two-day course on how to use it (so I'm told). It's effectively not all that useful for us, but then its not available to use. My point was that where articles rely on the use of books (only some of which have fragments on-line, and some of these are clearly copyvios), newspapers, journals, etc, spot checking can't work if only google and amazon are used. I gave Ethanol fuel in Brazil as an example of the type of nomination that appears to GAN, but is unlikely to appear at DYK. Since the use of Google Books, Amazon is propose below, why not have some fun and play with it, do some checks, see how long it takes, etc. Since, I assume you do this all the time at DYK, try a bigger article. Pyrotec (talk) 18:22, 7 September 2013 (UTC)
  • I'd also think some spotchecks would be sufficient (by which I mean simply looking up a few of the citations through links, Google Books, Amazon "search inside this book", etc.); spotchecks of the sources should be happening anyway to verify accuracy. Like you, Monty, I also try to doublecheck anything that specifically seems tonally odd or problematic. I'm resistant to writing a precise "how-to" here, however, because articles differ in their numbers, types, and uses of sources. I think it's better to leave it up to the reviewer what method (spotchecks, automated tools, etc.) would be best.
To put this RfC in perspective, Gatoclass (the proposer) stated several times that his main concern was that a copyvio check was omitted from the templates summarizing the criteria for reviewers (that part's already been addressed). He also proposed making copyvio a separate criterion so it wouldn't be overlooked by hasty reviewers. As most of the supports emphasized in their comments, this is a clarification rather than a change; I think it would be overthinking it to put a major new procedure in place. -- Khazar2 (talk) 02:07, 7 September 2013 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── It's interesting to look at what FA requires, as FA is a higher standard than GA, unlike DYK. The requirements are at Wikipedia:Featured article criteria, but I can't see them. Since today's featured article already appears on wikipedia's Main Page, we go for the higher FA standard. Pyrotec (talk) 11:55, 7 September 2013 (UTC)

Tentative summary/arbitrary break[edit]

To address the concerns of the RfC, I propose that criterion 1a. be broken into two criteria and expanded as follows:

1a. the prose is clear and concise, and the spelling and grammar are correct;
1c. it complies with Wikipedia guidelines on copyright and close paraphrasing.

Does this seem like a reasonable summary? (Re-reading the RfC, by the way, the formal proposal doesn't mention plagiarism--Gato only discusses it in his later comments--so I've omitted it here, too.) -- Khazar2 (talk) 02:38, 7 September 2013 (UTC)

Oh, heck, I just saw that this was closed as being about "plagiarism" rather than "copyvio". Let me try to clear this up with the closer. -- Khazar2 (talk) 02:40, 7 September 2013 (UTC)
Yes, indeed the DYK ditact would need:

1c. it complies with Wikipedia guidelines on copyright, close paraphrasing and Wikipedia:Plagiarism.

However, I believe that they want more that just a specification, they seem to what a full method statement and that is what I want to get onto.
DKY claim to do this fully on all their articles, well perhaps they do, but let us look at today's DYK's: Kumarakottam Temple , Sybil Campbell, Darling Downs funnel-web spider, Bigger Hair, Prisoners of Hope, and North Radworthy. They are mostly "trivial in length" articles (sorry to express it that way, since individual editors have worked hard to improve them up to DYK level) but most of them either have three or six citations, one has ten and one has 21 citations. It's probably not going to take much effort to do full copyvios, close paraphrasing and plagiarism checks on any or all of them. I've issued a formal challenge to DYK for them to do full copyvios, close paraphrasing and plagiarism checks on Ethanol fuel in Brazil since it is one I reviewed, just those checks nothing else, and to state how long it takes them it do it. I choose that one since, it is representative of articles that appear at GAN and in its current form its unlikely to appear at DYK since I very much doubt that anyone could double it in five days (the DYK requirement is little more than that, however I believe that a piece of software is used to count the words, so they don't have to do that themselves). GAN certainly does get nominations of a comparable size to today's DYKs and in backlog reviews, where points are issued for each review, they are popular to review since the effort is fairly trivial in time. However, they are not representative of the whole of the GAN spectrum, and just because DYK deals with "trivial in length" articles and nothing else they appear to have a "blind eye to the telescope" approach to such checking. I suspect that some of them are aware of the effort that is required, since they are very very unwilling to carry out their own copyright, close paraphrasing and Wikipedia:Plagiarism checks on GAs being passed to them under their impose system. That "hidden agenda" can be found on the DYK talkpages. So why should we carry out these checks for them? Pyrotec (talk) 08:42, 7 September 2013 (UTC)
Clarifying with Newyorkbrad (the closer), he states that he didn't specifically mean we needed to add checks for plagiarism to the GA criteria; he only used the term it because he sees it as interchangeable with COPYVIO. We're free to work out the form of that on this page. -- Khazar2 (talk) 14:07, 7 September 2013 (UTC)
Well, this is what wikipedia states in Plagiarism: Note that plagiarism and copyright infringement are not the same thing.[6] Copyright infringement occurs when content is used in a way that violates a copyright holder's exclusive right. Giving credit does not mean the infringement has not occurred, so be careful not to quote so much of a non-free source that you violate the non-free content guideline.[7] Similarly, public-domain content is plagiarized if used without acknowledging the source, even though there is no copyright issue. For advice on how to avoid violating copyright on Wikipedia, see Copyright violation. Pyrotec (talk) 15:45, 7 September 2013 (UTC)
Following on from the Ethanol fuel in Brazil example, if our concern is over whether the checks will actually be done, then perhaps we should split 1a into new criteria 2d, instead of 1c. Admittedly I haven't reviewed in a long time, but in practise doing a copyvio check goes hand-in-hand with checking that the sources actually back up the article content. —WFCFL wishlist 16:18, 7 September 2013 (UTC)
It depends on what is meant by "our concern" - its more a case of "their concern". It appears that there is a wish for new GAs to be listed on Wikipedia's Main Page in the same way that there is a Featured Article of the day and there are a blocks of "Did You Know" hooks. This seems to be under the control of DYK, who regard their checking procedures to be superior to GANs. DYK wish to impose on GAN, or perhaps force GAN to submit proposals to DYK for them to approve. The problem appears to be that DYK have concerns over whether GAN do these checks. There are voluminous discussions at Wikipedia:Did you know/Good Article RfC. Pyrotec (talk) 16:33, 7 September 2013 (UTC)
  • Pyro, rather than DYK bashing could you focus on the discussion here? We know how you feel. Remember, sometimes even a group you disagree with can have good ideas. — Crisco 1492 (talk) 16:36, 7 September 2013 (UTC)
  • In that case, perhaps you fully understand why I have those feelings. Its the "DYK missionaries" and the "we know better attitude" that I can't stand. DYK as a process (and an award) is good. I've reviewed at lot of GANs and some of the better GAN nominations I've reviewed had previously gained DKY. Quite a few of these become GAs and some of them later became FAs. Pyrotec (talk) 17:25, 7 September 2013 (UTC)
  • By "our concern", my point is that I care less about whether the DYK folks are happy (good articles on the main page will happen whether "they" are happy or not), and more about whether GAN is reasonably good at checking for copyvio. A few bad apples will slip through any net, FAC included. For me emphasis is merely on making sure that reviewers know that copyvios are something to specifically look for. This isn't about changing standards, it's about ensuring that the standards we have are applied as consistently as possible. —WFCFL wishlist 16:58, 7 September 2013 (UTC)
  • Overall I agree with Pyro's proposed wording. @Waiting: I usually do my spotchecks towards the end of a review (see Talk:Oliver Bosbyshell/GA1 for one example). By that point the article is roughly stable and so there will not be any major changes between the time I start my random spotcheck and the time it finishes. — Crisco 1492 (talk) 16:36, 7 September 2013 (UTC)
  • Whatever we put in it should be concise and relatively flexible. One of the strengths of GAs is that it allows for many different reviewing styles as long as the minimum requirements are met. Saying that I don't really see too much of an issue with this proposed wording at 1c, although I don't see an issue with keeping the status quo either (unless there are some examples of bad copyright violations that have got through). Any advice on how-to check for this criteria should not be on this page and instead presented at Wikipedia:Reviewing good articles (which already contains quite a bit on this subject). AIRcorn (talk) 22:19, 8 September 2013 (UTC)
  • I put the relevant part of the guidance on this page, some ten hours before you wrote your comment (see below): firstly, because DYK appear to be under the misapprehension that we don't have a written policy (we do it goes back two years to 4th August 2011); secondly, it would be nice to see what their written procedure is and how the two compare. Pyrotec (talk) 18:25, 10 September 2013 (UTC)

Frequent GAN reviewer here. First off, I always check for close plagiarism when I review GANs, and I have failed a few for this reason. I spotcheck a few sources to see if they actually support the statements in the article, and CP (close paraphrasing) concerns are usually immediately apparent when I do that. (I also always check FACs as well, and I frequently find I'm the only FAC reviewer who seems to care about such issues.) Having said that, however, I don't think we should put up requirements that are so strict that they will discourage people from reviewing GANs. The backlog is already too long, by any standard. I think it is the duty of the reviewer to investigate whenever they have reason to believe that a copyvio or CP issue is present, and it is a very good idea to check even when you don't have a reason to be suspicious, but I wouldn't make anything more than that mandatory. I think we should particularly avoid any suggestion that the reviewer is responsible for copyright violations that he/she did not detect; that would be poisonous to the GAN process. – Quadell (talk) 18:43, 6 October 2013 (UTC)

Draft proposals pending[edit]

As the original proposer of this change, I have some specific ideas about what changes need to be made but I haven't time to draft them yet. I probably should have drafted them as the RFC was coming to a close but I was intending to open a discussion here at that time and wasn't expecting someone else to do it right away. I will probably propose some changes to the wording of the GA review tables in the next day or two. In the meantime, we may need to have a discussion about whether only a spot check should be required or a thorough check - if only a spot check, we may need to define what that means exactly. Gatoclass (talk) 13:16, 7 September 2013 (UTC)

By "thorough check" do you mean every source? If so, why on earth would we do that? FAC doesn't even require that... Spot checks (checking a few sources to make sure there's no obvious copyvio, with an especially close eye on any places where the prose/refs seem hinky) are more than adequate for both FAC and GAN. Obviously there is a scale of checks: new nominators and nominators who have had problems with copyvio before should have their work checked especially closely, but this is just common sense, not something that we need to codify and create instruction creep with. Dana boomer (talk) 18:08, 7 September 2013 (UTC)
Well said. A voice of reason at last. Pyrotec (talk) 18:23, 7 September 2013 (UTC)
I would also suggest that practicality is considered, something that is quite quick and easy to run on a typical 6 to 10 citations, and expectionally just over 20 citations DYK, based on today's "did you know" may well fall over on a large 200 to 250 citation GAN. Pyrotec (talk) 18:33, 7 September 2013 (UTC)
Okay, assuming there is consensus for a spot check only, we might want to add a few words to the guideline per Dana boomer's comment above outlining what a spot check consists of. Gatoclass (talk) 06:58, 8 September 2013 (UTC)
Also, what is meant tables, we don't have tables, do you mean the (optional) review templates at Wikipedia:Good article nominations/templates? Pyrotec (talk) 18:45, 10 September 2013 (UTC)
I think we adequately covered "spot checks" back in the August 2011 update to the quidelines see below. Pyrotec (talk) 18:45, 10 September 2013 (UTC)

August 2011 guidelines[edit]

OK. I'll go through this step by step. The guidance is given at Wikipedia:Reviewing good articles. In it there is a section Assessing the article and providing a review, and Step 1, item 7 states: "The article is free of obvious copyright violations. Reviewers can use several tools, as well as Google searches, to help establish whether material has been plagiarised or cut-and-paste from some of the electronic sources used; but this is not a trivial undertaking. 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.". That part has not changed very much in the last two years. The clause seems to have added by User:Chaosdruid during a series of edits on 4th August 1911. There was a very long discussion on copyvio checking this talk page in August 2011 and a extensive range of editors took part. (See here). Incidentally, 2011 was probably the last big war of words between DYK and GAN over copyvio checks, lack of them and which system was better. It would be nice it we could avoid a rerun / return match. Pyrotec (talk) 10:30, 8 September 2013 (UTC)
As a matter of interest, what is the written policy of DYK for doing such checks? Pyrotec (talk) 09:19, 8 September 2013 (UTC)
I'm make the point that I made above yet again. This is "spot testing" and it can only test against what the reviewer "asks" the search engine to look for. If a book (journal, newspaper etc) is only partially digitised on Amazon, Google books, etc, or not all, and someone has copied the material that is not digitised and/or they've copied something that the reviewer has not "asked" the search engine to look for then it's not going to be found. I'll make another point, let us say that there is a strong match between what is in the article and what is on another website (which claims copyright): which came first, it is not unknown for sites to copy from wikipedia and then claim ownership. Obviously, in such a case the article review gets put "on Hold" for the copyvio to be resolved, but it could result in the article being passed at a later date, on the basis that the material was on wikipedia prior to it appearing on the other site. Pyrotec (talk) 09:38, 8 September 2013 (UTC)

I don't think it is reasonable to expect that "a reviewer will have access to all of the source material". Ideally, sure... but no one should be discouraged from undertaking a GAN review simply because he/she is unable to check the sources directly. – Quadell (talk) 18:45, 6 October 2013 (UTC)


This keeps coming up, so I've expanded a footnote to directly tell reviewers to stop fussing about whether the citations are perfectly formatted. Compliance with CITE has never been required. It just seems in this instance, some reviewers assume that the omission of CITE means that it is one of the requirements. I kind of assume that this is because it's easier to figure out whether an article looks good rather than whether it is good, but at this point, it may well be one of those urban legends that new reviewers believe because some old reviewer told them that it was required.

If anyone objects, then of course we could have an RFC to see whether perfectly formatted citations should be added as a criteria. I'd personally oppose that, but perhaps others like the idea of failing articles because of inconsistent formatting or the failure to include a publisher's name in every citation. WhatamIdoing (talk) 20:31, 27 November 2013 (UTC)

Thanks; I ran into this in a review just a week back. It was definitely overdue for clarification. -- Khazar2 (talk) 21:55, 27 November 2013 (UTC)
I also think this change is a useful clarification. Quadell (talk) 22:36, 27 November 2013 (UTC)

Image and media criterion[edit]

I'll jump in to start the discussion, but I reverted this edit to the criteria earlier. The effect, as I read that edit, was to remove our preference for including media in articles. In short, they may not be required, but if we have the ability to include them, we should.

If an article I was reviewing lacked any imagery, I would ask the nominator if there was any that could be added. If given reasonable explanations why the article lacks imagery (or sounds, or video, etc), then I'd just move on without it.

If the desire is just to copy edit the note for clarity, that's fine, but the preference to include media should not be removed without a discussion. Imzadi 1979  20:17, 28 March 2014 (UTC)

I will point out that the original/revert version should not simply say of "appropriate copyright status" without clarifying free and non-free image and important issues with the later. I know the intent of this statement is considering free vs non-free and when non-free are appropriately used, but I think this needs to be more specific. --MASEM (t) 20:27, 28 March 2014 (UTC)
I disagree with this proposed change. Images should be included in a good article. What subject is an otherwise good article that cannot have any images? Chris Troutman (talk) 20:45, 28 March 2014 (UTC)
Images should only be included if they are actually helping to improve the reader's understanding of the article, even with free images. Some topics are very "meta" and thus cannot be demonstrated with images, so requiring images makes no sense. Given that the FA process does not require images, GA certainly cannot require them. --MASEM (t) 20:54, 28 March 2014 (UTC)
Many GA writers are not native speakers of English. I only understood the current footnote after reading it thrice. When I started editing, that footnote said "a lack of images does not disqualify an article from GA status". GA needs to remain less demanding than FA. A "preference for including media" will increase systemic bias against certain topics. Some GA writers may wish to avoid dealing with image policy and should be allowed to do so. --Hildanknight (talk) 08:18, 29 March 2014 (UTC)
I fail to see how a preference for including media will increase systemic bias. All this note is saying is that if there is a perfectly good picture of a jaguar on Commons, the jaguar article should not be completely unillustrated. If, however, there are no pictures of a particular breed of dog, then the article on that breed cannot be kept from GA for not having an image. All this note is doing is expanding on the criteria itself, which says "Illustrated, if possible, by images:" Dana boomer (talk) 12:30, 29 March 2014 (UTC)
@Dana boomer: A fairly new editor who is not a native speaker of English writes and nominates an article about a historic building in a Middle Eastern country. The article is well referenced, neutral and broad, with a few minor grammatical errors that are easy to correct. Commons has a decent free image of the building. However, the nominator did not include the image because he does not understand our image policies (which even native speakers may find confusing) and does not even know what Commons is. Perhaps he did not understand why a few images that he uploaded (in good faith) were deleted and is thus deterred from uploading further images. As the reviewer, what would you do?
Another editor writes an article about an African politician, which becomes a GA after she convinces the reviewer that there are no free images of the politician. Feeling motivated, she decides to write an article about another African politician. She can find only one free image of this politician, but deems the image unsuitable for the article. Perhaps the image does not show his face clearly (see Yip Pin Xiu for a real example), shows him playing football (misleading) or grossly violates an African cultural taboo (BLP issues). As the reviewer, what would you do?
--Hildanknight (talk) 13:56, 29 March 2014 (UTC)
For both of them, I would say "Hey, I found this image on Commons. What do you think of including it in the article?" The second one is less clear than the first, but if the editor had a good reason for not including it in the article, then there would be no problem. These would both be minor issues in a GA review, and would in neither case be grounds for failing the article's nomination. I have had image questions come up in a number of articles that I have nominated or reviewed for GA, and every time they have been easily solved through dialogue between the participants - sometimes ending up with including a certain image in an article, and sometimes not. I have written GAs that had images, and GAs that didn't, and GAs that had images later in the article but were determined to have no pictures illustrative enough to be associated with the lead. It would be ideal to have a high-quality, representative image in the lead of every GA, but it's not an ideal world, which is why the criteria is worded as it is. There is a major difference between "preference" and "required". Dana boomer (talk) 14:19, 29 March 2014 (UTC)
I'd echo Dana's comments on this one. Hchc2009 (talk) 14:34, 29 March 2014 (UTC)
@Dana boomer: For the first case, the reviewer should not fail the article simply due to lack of images. If the nominator is unfamiliar with image policy and the reviewer wants the article to have an image of the building, then the onus is on the reviewer to add the image. For the second case, free images are not "readily available" if the nominator has a convincing argument that the only free image is not suitable. Do we agree on these? If so, then how can the wording be made clearer to GA writers who are not native speakers of English? Reviewers may also misread the criteria and request images, even when they are not "readily available". --Hildanknight (talk) 15:27, 29 March 2014 (UTC)
  • @Hildanknight:: Reread Dana's reply. She never said she would fail the article. She said that she would mention that there is an extant free image of the building on Commons, and suggest using it, thus educating the GA nominator and offering him/her the chance to add the image to the article. Give someone a fish, they eat a meal. Teach someone to fish.... — Crisco 1492 (talk) 15:44, 29 March 2014 (UTC)
@Imzadi1979: @Chris troutman: @Masem: @Dana boomer: @Crisco 1492:
There used to be a footnote that compliance to all aspects of the Manual of Style is encouraged, but not required for GA. Is this "preference for including media" part of the criteria itself or simply a suggestion to use images that are available and suitable? In other words, if an article is well written, factually accurate, neutral, broad and stable, but does not contain images that are available and suitable, would its nomination be failed? A suggestion is perfectly fine with me, but the wording of the footnote is unclear.
Many GA writers are not native speakers of English (as GA criterion 1a is easier than its FA counterpart). I only understood the footnote after reading it thrice (my mother tongue is Mandarin Chinese). How the criteria could be interpreted (or misinterpreted) is as important as the exact meaning. If an editor mistakenly thinks that images are required, she may be deterred from writing GAs or worse, upload a copyrighted photo and claim the photo is free (due to her limited understanding of image policy). New reviewers may make the similar mistake of failing nominations due to lack of images, even when there are no available and suitable images.
The purpose of GA is to be a reasonably accessible standard for all articles and editors. Including articles about Chinese culture, African politics or Islamic law, written by new editors who are not native speakers of English. This poorly phrased preference for including images makes GA less accessible. In my opinion, editors who wish to avoid dealing with our confusing image policies should be allowed to write GAs without images, regardless of topic.
--Hildanknight (talk) 15:28, 2 April 2014 (UTC)
@Hildanknight: We meet again. I fondly recall our last discussion in February. The criteria is that images, where possible, be included. It's not a requirement to have images, only to include images if they're available. That's not so hard, in my opinion. Can you point to a GA nominee that didn't get promoted because it didn't include images? You claim that the "purpose of GA is to be a reasonably accessible standard". Please tell the GA reviewers because that's becoming less and less true everyday. English isn't the easiest of languages but I've not found the wording of the policy to be difficult to understand. If you can't speak English well then this project is probably not for you. I don't contribute to de.Wikipedia because my German-language skills barely qualify as basic. That's as it should be. Chris Troutman (talk) 17:38, 2 April 2014 (UTC)

Nomination and improvements from a topic banned editor[edit]

I reviewed Pharnavaz I of Iberia, and then found out that the editor who nominated the article, User:Jaqeli, was topic banned from articles that deal with Georgia and Armenia together. I requested an outside opinion from users involved with the topic ban of Jaqeli, and while I waited for that, the Jaqeli edited the article to satisfy the problems that I found during the review process. Because of this editing, Jaqeli has been blocked for two weeks. But Pharnavaz I of Iberia now satisfies the criteria for GA. How should I proceed? May I promote an article for GA even though the editor who nominated it and brought it up to snuff is topic banned?--¿3family6 contribs 13:57, 14 June 2014 (UTC)

Notability missing from GA criteria[edit]

Did you know that notability is not mentioned in the Good article criteria? A Good article is being considered for deletion (here) mainly because of lack of notability. Checking carefully, I agree with the lack of notability in this case, but I can also see how the article passed GA and I believe it (barely) met GA criteria. Why is notability missing from GA criteria? This seems to be a serious omission. Prhartcom (talk) 23:58, 15 November 2014 (UTC)

Indeed it is a serious omission. Back in June, there was a discussion at WT:GAN about notability and nominations. The discussion concluded that notability was an implicit requirement. However, it would probably benefit to explicitly include notes on notability. Snuggums (talk / edits) 01:13, 16 November 2014 (UTC)
Yes, otherwise an article that is not notable could become GA as this one did, while technically meeting all of the GA requirements. Have I placed this discussion in the wrong place? I have added a short note at WT:GAN. Prhartcom (talk) 16:40, 17 November 2014 (UTC)
How an earth can an article meet the GA criteria (particularly #2 - "verifiable with no original research" and 3 - "broad in coverage") and not be notable? A GA requires in-line citations to reliable sources, frequently tens if not hundreds of them, which are the basic definition of WP:GNG. This just sounds like silly wikilawyering. Looking at the article itself, I could easily see myself asking a question along the lines of "what has this got to do with the song in question, rather than the parent album" during a GA review. Ritchie333 (talk) (cont) 17:19, 17 November 2014 (UTC)
I agree, but the answer to your question is, "Because the criteria never mentions notability." That should be an easy fix. How do we modify the criteria to mention notability? Prhartcom (talk) 17:35, 17 November 2014 (UTC)
I think when this has been brought up before, it's always implicit that a GAN/GA has to meet all other policy and guidelines; the criteria of WP:WIAGA is to what degree specific policy and guidelines should be met, or above and beyond the bare minimum for the guideline. For example, WP:V doesn't absolutely require well-formatted reference citations, just that you have inlines and references, but GA can say "well, you should start approaching the MOS for citations, establishing a consistent style, etc.". Other policy/guidelines not explicitly mentioned should still apply - for example, if a GAN turns out to be a glaring BLP violation, it should be removed. Notability is the same way.
To Ritchie333's point: I have seen articles that fail notability but do meet WP:V and are "broad" (in as much as the subject is covered in sources), but these typically are ones using tons of primary sources and minimal or no secondary ones. I would suggest not trying to shoehorn in notability into this, but do explain that beyond the criteria GAs still must meet all applicable policy and guidelines that relate to content if not otherwise mentioned/covered in the criteria. --MASEM (t) 17:45, 17 November 2014 (UTC)
Okay, so can we explicitly add to the GA criteria "Must meet Wikipedia:Policies and guidelines? Prhartcom (talk) 17:56, 17 November 2014 (UTC)
Wikipedia uses community consensus to affirm or deny WP:N being met (at AfD and merger/redirect discussions), not the opinion of one reviewer. The GA reviewer should only be assessing the quality criteria of the article, not its right to exist. WP:GAN is not WP:AfD. maclean (talk) 20:22, 17 November 2014 (UTC)
However, it seems a waste of time for a reviewer to approve an article that will be deleted anyway.--¿3family6 contribs 20:48, 17 November 2014 (UTC)
This is simple: if you see an article in the queue at WP:GAN which you think should/will be deleted, then nominate it for deletion, and add a note to the GAN template that the article is currently at AfD. If it survives, then a reviewer can feel reasonably confident that their review will not be a waste of time. Adabow (talk) 20:57, 17 November 2014 (UTC)
(edit conflict) If the reviewer thinks the article shouldn't be here, he should nominate it for deletion instead of reviewing it. I don't think it would be appropriate to include notability explicitly in the requirements to be a Good Article, though, as that requirement applies to all articles, not just good articles and I really don't think that general article requirements should be explicitly listed here, except perhaps in a "please check that this article meets the general article requirements before beginning your review" sense. Doing otherwise could too easily lead to the infliction of an enhanced notability standard on articles by reviewers who don't understand the reason for the instruction. – Philosopher Let us reason together. 20:59, 17 November 2014 (UTC)
OK, agreed. Can we add this procedure to the GA instructions? The reviewer instructions currently do not mention "Ensure the article meets Wikipedia:Policies and guidelines." You know this will happen again unless we solidify this. Prhartcom (talk) 21:09, 17 November 2014 (UTC)
As a comment, it wouldn't just be limited to notability/AFD. Say the topic is better as a merge to a different article - that's a step to propose. If it's a major BLP issue, report it to the BLP/N and let them deal with it. Copyvio? Same thing. But one thing to consider: review the talk page of the GAN article and see if it has gone through anything like these steps, as you don't want to double-ding it if it's been cleared appropriately. For example, if you get a GAN that you think is non-notable, but find it has survived AFD in the last few months (and stable since), you should respect that decision and review the GAN without that concern. But do agree adding language that general policy/guidelines not spelled out in depth in the GA criteria should still be expected to be met. --MASEM (t) 00:11, 18 November 2014 (UTC)
I don't think I'd change the GA criteria, but the section above - "Immediate failures". Right after the numbered list, I'd add "Reviewers should be aware that just as some tags may be out of date, other tags may need to be added. If an article needs a cleanup banner (including, in extreme cases, an AfD or PROD tag), rather than reviewing the article, the reviewer should add the appropriate banner instead of reviewing the article." The phrasing isn't perfect, because an AfD tag obviously isn't a cleanup banner, but I'm sure someone will propose a refinement of it? – Philosopher Let us reason together. 04:03, 18 November 2014 (UTC)
I get the feeling we're wanting of concrete examples, so I'll provide one: I got Wisp (Sonic), a page about a race of aliens that are used as power-ups in a few Sonic the Hedgehog games, all the way up to FAC and had several supports for content until one editor came in and asserted that the subject was not notable. The community was at first rather hostile, but eventually came to form such a consensus, and instead of receiving a gold star, the article received a "merging" into the character list, much uglier than it once was. Because no one bothered to tell me beforehand that several full-length articles dedicated to the subject are necessary for notability (instead of just dozens of passing mentions and a couple articles that may have been more about the games than the species) until I'd spent that many hours working on the page, digging up the most obscure articles from the Internet to use, I basically threw up my hands and said, "F*ck it, enjoy your notability, everyone; I'm out." My interest in character articles has been completely soured now that I know they can be redirected so easily at the community's whim, as several of mine (including a few GAs) were right after this one, and at this point I just make it easier on myself by refusing to participate in notability disputes at all. Just in case anyone thought it was harmless not to be completely up-front about what's required for notability. Tezero (talk) 15:49, 20 November 2014 (UTC)

I thank everyone for their input. I believe we have gathered from this discussion that we should not add to the Good article criteria, but I believe we are seeing support for adding to the Good article instructions. Specifically, in Step 1 of both the Nominating and Reviewing instructions, simply include a new sentence stating, "Ensure the article meets Wikipedia policies and guidelines as expected of any article, including neutral point of view, verifiability, no original research, and notability." Being bold, I have just made this change to the instructions and I welcome and appreciate your comments below. I don't believe adding this sentence harms the existing instructions; I believe it will help prevent this kind of issue in the future. Cheers, all. Prhartcom (talk) 17:22, 20 November 2014 (UTC)

Notability is handled at AFD or in other venues. As has been mentioned above, FACs have even been deleted. --Rschen7754 07:26, 23 November 2014 (UTC)
We know AfD deletes articles. This discussion was to attempt to nip it in the bud, so to speak; to prevent editors from doing all that work of improving an article to an excellence recognition without first checking for notability. Hopefully this measure will help a tiny bit. Prhartcom (talk) 15:09, 23 November 2014 (UTC)
I think that rather than say "ensure that", it should be more along the lines of "be aware of" or "instead of reviewing you may wish to". The "ensure" phrasing seems to imply an extended scope of a GA review, while the purpose, as I see it, is merely to avoid reviewers wasting their time on articles with major non-GA problems, not to extend the scope of GA. – Philosopher Let us reason together. 19:15, 23 November 2014 (UTC)
Okay. Do you want to go ahead and make the change to the instructions? There are two places: Instructions for nominating and instructions for reviewing. Prhartcom (talk) 20:27, 23 November 2014 (UTC)
On further consideration, I left your change in place in the nominating instructions. "Ensure" does seem appropriate when nominating an article. On a related note, what were your thoughts on my suggestion for the "Immediate failures" section of the criteria page? Or do you think that the change to the nomination instructions is sufficient? I think it would be appropriate to put in some sort of warning for the reviewer as well as for the nominator. – Philosopher Let us reason together. 19:29, 25 November 2014 (UTC)
I do think what we did already is sufficient, and I'm glad that you helped propose it. We already added a new warning for the reviewer, in the Reviewing section of the Instructions. This idea is too wordy and the first item in the Immediate failures numbered list already addresses cleanup banners. However, after this first item's phrase "It has cleanup banners that are obviously still valid" you could consider adding "or needs new cleanup banners." I was wondering if we should add the entire "Ensure the article meets ..." sentence to a new number 4 in the Immediate failures list. So far we have managed to touch only the Instructions and have avoided touching the Criteria. Prhartcom (talk) 15:12, 26 November 2014 (UTC)
Heh - this fell right off of my watchlist. I've added the "or needs new cleanup banners" phrase - it's short and sweet and should address the issue. – Philosopher Let us reason together. 01:05, 13 December 2014 (UTC)