Wikipedia talk:Content review/workshop/Archive 2

From Wikipedia, the free encyclopedia

Archive This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Contents

New straw poll for current topic

Here is the list of topics from the old straw poll. You may want to visit the old discussion to look at some of the arguments there; there were some excellent points made there. I've added one additional poll option: to delay selecting another topic until after the current PR automation effort has progressed. That option is just to allow people to indicate a concern that we will outpace the energy level of our participants.

I also propose to renotify many of our previous participants, in particular those who have not posted in the last week or two. I am particularly thinking of people we asked to join us who had a specific focus area: Walkerma, Awadewit, LaraLove and a couple of others. Please notify anyone else you think might be interested.

Here's the list, with my supports in place. Mike Christie (talk) 01:09, 4 December 2007 (UTC)

I've left messages for User:SandyGeorgia, User:Jayron32, User:Ling.Nut and User:Derek.cashman. The thought occurs that it may be worth inviting one or two from the LoCE as well (maybe User:Galena11, User:Happy-melon and/or User:Unimaginative Username?); these editors are currently driving a reorganisation there and their input as copyeditors might be valuable here too. EyeSereneTALK 10:25, 4 December 2007 (UTC)
I think it would be a good idea to notify anyone from LoCE that you think might be interested. Mike Christie (talk) 11:57, 4 December 2007 (UTC)
OK, done. EyeSereneTALK 17:47, 4 December 2007 (UTC)

One or two of these questions seem a bit dated to the origins of this workshop. Although this is cause for a wry smile on a couple of occasions, I wonder if we could rephrase them? Suggestions below. Geometry guy 20:34, 4 December 2007 (UTC)

  • How can reviewers be encouraged at peer review? This is the previous topic; it is reproduced here for completeness and in case anyone feels the previous discussion was unsatisfactory and we should revisit this.
  • How can reviewers be encouraged at the other content review processes?
  • Is there a way to reduce redundancy between the FA and GAcontent review processes?
  • How do we recognize the value and quality of good, short articles that are unlikely to grow?
    • Support - This should be an easy one to conclude. Lara❤Love 05:54, 4 December 2007 (UTC)
       :-) Geometry guy 20:34, 4 December 2007 (UTC)
  • How can we encourage the connection of subject-matter experts with articles that need their expertise?
    • Comment I'd support this if I thought there was any real chance of addressing it. Mike Christie (talk) 01:09, 4 December 2007 (UTC)
    • Support - I think this is doable, at least with some of the technical projects. Perhaps others will be willing to help once some of the other issues have been addressed. Lara❤Love 05:54, 4 December 2007 (UTC)
  • How can all the processes scale? Each has its own barriers to getting ten times more productive.
    • Support. Scaling any process would be a huge win. Raul just nominated a delegate at FAC, which is an excellent sign as it indicates that FAC submissions are on the increase. I would be interested in anything related to scaling content review. Mike Christie (talk) 01:09, 4 December 2007 (UTC)
    • Support. Although this ties in with several other issues, such as instruction creep, single reviewers vs multiple reviewers, automation and so on. I think we should bear it in mind in every other discussion, and perhaps come back to it as a specific topic later. Geometry guy 20:34, 4 December 2007 (UTC)
  • Acrimonious arguments at individual processes and in discussions between regulars at the different talk processes drain time and energy; is there a way to eliminate or reduce these source of friction? How can we encourage reviewers and editors to work together to improve content, and also reduce friction between different approaches to content review?
  • Can we avoid or mitigate the weaknesses of single-reviewer systems in which one user may mistakenly pass an article that does not meet the relevant standards?
    • Support - In addition to sweeps, there are other proposals that have been, well, proposed for this. Discussion and brainstorming would be good here. Lara❤Love 05:54, 4 December 2007 (UTC)
    • Comment. I think this is a slightly loaded question, and could be rephrased, and possibly united with the next question. Geometry guy 20:34, 4 December 2007 (UTC)
  • Can we avoid or mitigate the weaknesses of multiple-reviewer systems in which some undocumented and unrecorded form of consensus has to be decided on by a single individual?
    • Comment. This reads as an amusing call and response of loaded questions! The real issue, in my view, is how to minimize the disadvantages of any system while maximizing the benefits. For multiple-reviewer processes, I see scaling as a more significant issue than how consensus is determined. Geometry guy 20:34, 4 December 2007 (UTC)
    • Comment Is that a complicated way of saying, "Can we improve on the current system of having an FA director who makes unexplained decisions based on an undisclosed method of determining consensus?" Ling.Nut (talk) 13:01, 5 December 2007 (UTC)
  • What can be done in all processes to reduce bureaucracy and instruction creep and make things easy for the content-writers wishing to use the processes?
    • Support, for either FA or GA, whichever one we wanted to focus on. I suspect this would turn into another automation exercise, which may not be the right thing to work on next given we don't have a bot written for PR yet, but it would certainly be a good thing if we could get it done. Mike Christie (talk) 01:09, 4 December 2007 (UTC)
    • Support. This is an important question, and I think GA is the most interesting case here, because it is supposed to be lightweight, but isn't. This may involve some automation, but it is likely to be of the kind that can be done by VeblenBot (category listing), as GA does not involve any archiving. However, it might be better to tackle another topic first. Geometry guy 20:34, 4 December 2007 (UTC)
  • Can statistics be improved for any of the processes, in order to help measure their success? No process keeps complete statistics on both articles reviewed and reviewer participation, but there is great variation among existing processes.
  • Are there opportunities for collaboration between the various review processes?
  • Should there be a unified review hierarchy?
  • What can be done to make it easier for editors to make articles comply with the somewhat technical manual of style?
    • Support I like this idea, because MOS is controversial. I think if we want to tangle with controversy in a way that might be helpful, this is a decent starting point. Mike Christie (talk) 01:10, 4 December 2007 (UTC)
    • Comment. In what way is MOS controversial? I'm not aware that there's any serious argument against wikipedia having a style guide, as other respectable publications do. But there may be an issue here about reviewing only for MOS issues? --Malleus Fatuorum 02:02, 4 December 2007 (UTC)
      Yes, I didn't phrase that well -- MOS is not controversial in itself, as far as I know, but I've seen a few complaints about content reviews that focus on hyphens and dashes rather than on content. That's what I meant by controversy. If there were a MOS-fixes task force which only dealt with MOS-compliance, and which had a quick response time, then that might be one way to assist editors in coming up to speed on MOS issues. Mike Christie (talk) 03:12, 4 December 2007 (UTC)
      My initial reaction was yes, that sounds reasonable. But then I wondered whether articles not yet ready be nominated for GA/FA would benefit from that kind of review. I can't remember where I saw the suggestion, but I do think that MOS compliance is an important part of the GA criteria, even though it's only partial at present; how can you be MOS-compliant until you know what it is that you want to say? --Malleus Fatuorum 03:50, 4 December 2007 (UTC)
      Hyphens and dashes, for the most part are only an issue for FA, and should remain that way. In GA, it's usually mentioned, in my experience, as something that would improve the article, but not something required for GA. Lara❤Love 05:54, 4 December 2007 (UTC)
    • Support: I think the discussion Malleus is referring to is this one (where SandyGeorgia offered to put together a list of the most common FAC-MoS issues that could perhaps be screened for by GA). Anyway, no objection to taking this next. EyeSereneTALK 09:41, 4 December 2007 (UTC)
    • Support: I think the MOS should remain static for particular lengths of time (6 months? 1 year?) and then be re-released with a list of revisions that editors can see. Right now, it is hard to keep track of all of the changes. A MOS isn't a problem - an ever-changing MOS is. Awadewit | talk 15:05, 4 December 2007 (UTC)
    • I'll actually oppose this. Tony1 and others have already recently over-hauled the MoS. What would we be approving here? Getting into the specifics of grammar as described by the MoS? How many non-technical ways are there to discuss, say, quotation marks? The MoS is big and technical because there's no way manuals of style cannot be. We can discuss, but I don't see it as primary now. Marskell 18:16, 4 December 2007 (UTC)
    • As an addendum: what might happen with the MoS—describing the most common mistakes, for instance—can work through SandyGeorgia's and Tony1's talk pages. I don't think the workshop needs to take it on. Marskell 18:19, 4 December 2007 (UTC)
      Comment the proposal is 'how can we help editors to comply with the MoS'. This certainly involves the review processes, even if only to the extent that we should perhaps be checking MoS compliance more thoroughly earlier on in the review hierarchy, and also doing more to educate article writers. EyeSereneTALK 20:17, 4 December 2007 (UTC)
    • Comment I agree with Marskell, that we should not attempt to discuss the content of MoS. However, there is some scope for us to discuss how to help editors to comply with it, and broader issues over how to stabilize changes in MoS, and publicize changes where they are necessary. The main controversy, as I understand it, is the question of just how prescriptive the MoS should be. I'm not sure if we have something useful to contribute here, but who knows? Geometry guy 20:34, 4 December 2007 (UTC)
  • Should labels themselves be examined—stub, start, B, A, Good, Featured?
    • Support per above thread. Marskell 09:21, 4 December 2007 (UTC)
    • Support as above. EyeSereneTALK 09:51, 4 December 2007 (UTC)
    • Support I'm not sure that ratings without thorough reviews are meaningful and I'm not sure that distinctions between stub and GA are meaningful. Awadewit | talk 15:05, 4 December 2007 (UTC)
    • Comment This topic may be worth an airing, as there appear to be some misunderstandings about what the WikiProject assessments are for. Geometry guy 20:34, 4 December 2007 (UTC)
      • Comment - I agree. These different classes may not appear to be worth much, but for the assessment purposes of a Wikiproject, there is a difference between stubs, starts and Bs. If you want a task force to start expanding stubs, it's helpful to have that category there. Another for those working to expand starts or build on Bs to go for GA. Lara❤Love 04:55, 5 December 2007 (UTC)
      • Comment Am I the only one who actually likes something? I like stub, start, B, GA, A and FA. I like the system. I like the ratings. I like the idea (my idea, anyhow) that GA is a surface check; A is a content check; FA is both (and more intensively examined). I like the idea that Wikiprojects should be able (but not required) to claim the exclusive right to eval. an article within its scope to an A. Ling.Nut (talk) 13:01, 5 December 2007 (UTC)
        You are a true and faithful hand sock, and I agree with everything you say! I like it too. Geometry guy 19:56, 5 December 2007 (UTC)
      • Comment. If the present categorisation actually worked in the way that Ling.Nut described, then there may indeed be something to be said its favour. But as GA is more than a surface check, A is often skipped, and FA is not infrequently less than an "intensive" check it clearly doesn't. Stub, start, GA and FA, along with the parallel peer review process, seem like quite enough categories. As things stand, I'd have to admit to being rather unclear even about what the intended purpose of the A and B classes is, particularly if the A class is considered to be some kind of optional peer-review process. Perhaps the bottom line is to consider that every classification exists to prepare an article for its FA nomination, even where there is no possibility or intention to ever submit the article to FA? Perhaps FA ought not to be the tough review at the end, but almost a rubber-stamping of what has been done before? Of course I realise that once again raises the question of the purpose of GA; is it a separate goal to be aimed for, for those articles that have no chance of achieving FA for whatever reason, or is it a step on the way to FA? The answer to that question ought to inform the creep that seems to have taken place in the GA criteria. This comment has already grown longer than I intended, so I'll stop now. :) --Malleus Fatuorum (talk) 20:29, 5 December 2007 (UTC)
        These are interesting comments, and maybe there is some mileage in this topic. I'd like to discuss them further, but meanwhile, for clarification: I agree GA is more than a surface check; also for many articles, WikiProject involvement is unnecessary, and so stub-start-B-A is a luxury or an irrelevance. Within many WikiProjects, stub-start-B-A work very well as Ling described; it is only when these ratings are interpreted outside of the WikiProject context that confusion arises. This is something that needs to be clarified. Geometry guy 21:34, 5 December 2007 (UTC)
  • Can we improve the mechanism for nominating articles for review, either for each process individually or across the project as a whole?
    • Strong support - What of the possibility of a screening panel? All articles listed at the same place by article editors. A panel of users familiar with GA and FA determine, individually, where the article should be listed based on its current quality. The panel then lists as PR, GAN or FAC. Lara❤Love 05:54, 4 December 2007 (UTC)
    • Support per Lara - this might need to be addressed before taking other discussions further, as it might impact all review areas. It may also be fairly brief ;) EyeSereneTALK 09:51, 4 December 2007 (UTC)
      • Comment: In my opinion, the reviewers won't be strict enough. I've had so many people tell me that articles were ready for FAC that I knew were not. Awadewit | talk 15:05, 4 December 2007 (UTC)
    • Comment. This sounds a bit creepy to me... Geometry guy 19:56, 5 December 2007 (UTC)
    • Comment: sounds creepy to me too, very much against that idea. --Malleus Fatuorum (talk) 22:39, 5 December 2007 (UTC)
  • Rather than selecting another topic now, we should wait until after the current PR automation effort has progressed.
    • Support. If there's enough interest we can go ahead and start another discussion, but I'm sensitive to the fact that we lost a couple of participants last time, I think partly because of the high level of effort needed to participate. (I.e. there were a lot of posts to read and keep up with.) I think it's more important that this workshop have stable long-term participants than that it is continuously active. So I've supported other options above, but this would be acceptable too. If we select this option I'd suggest revisiting in early January. Mike Christie (talk) 01:09, 4 December 2007 (UTC)
    • Comment. This is usually a time of year when contributions drop, so it may be that in-depth discussion of the next topic will have to wait until January. On the other hand the enthusiastic response to this straw poll raises a question mark over that. :-) Geometry guy 20:34, 4 December 2007 (UTC)
  • Comment, I don't know where to put my vote, so I'll put a comment and let you all interpret :-) The biggest problem I see throughout all processes is that, with the exception of the automated peer review bot, *no one* routinely reviews articles for WP:V and WP:RS, and *no process* reviews articles systematically for WP:MOS compliance. I've said before I think that GA could be useful in this process; if it's not possible for GA to take on some of that role, we need to strengthen peer review, as it does work well when it works. I sense some "between-the-lines" criticism of MOS reviews above; well, someone has to do it, and if someone doesn't review sources for reliability somewhere along the way, we can end up with articles with important deficiencies. My experience with A-class reviews is that some of them at least tend to make sure sources are reliable and the highest quality sources are used, so to that end, I find some (for example medical or MilHist) A-class reviews useful. On the other hand, when an article has passed GA, I can't assume that means anything about MOS compliance or reliability of sources, since those aren't necessarily covered. SandyGeorgia (Talk) 18:07, 4 December 2007 (UTC)
  • Support I was just asked to give "leeway" to sources being used in a GA article, something I find highly disturbing. Awadewit | talk 18:57, 4 December 2007 (UTC)
  • Scary thought, huh? We don't "give leeway" to sources used in any article, even stubs :-) And some people look at all those little footnotes and think an article is cited, and never even click on the links. SandyGeorgia (Talk) 19:02, 4 December 2007 (UTC)
Hm, but why does any of that dictate a vote to discontinue until after PR bot coding is done? Marskell 20:35, 4 December 2007 (UTC)
I don't think Sandy intended it to -- I think she simply placed another bullet at the very end of the list to hold a comment she couldn't find another sensible location for. Mike Christie (talk) 20:38, 4 December 2007 (UTC)
Yes. It was Awadewit's support that threw me. Marskell 20:46, 4 December 2007 (UTC)
Correct on my count; I just stuck my comment at the end, and will go along with whatever you all choose, but I don't think you need to discontinue and wait. SandyGeorgia (Talk) 21:26, 4 December 2007 (UTC)
  • What can be done to ensure that content review processes evaluate articles for verifiability and to check the sources used are reliable? (This topic is extracted from Sandy's comments above; Sandy, please edit if this doesn't reflect your intentions. I would count this topic as having Sandy and Awadewit's support, per the comments above.)
    • Strong support Wikipedia's bad rap among the people of the non-wiki world is "unreliable." Ling.Nut (talk) 13:04, 5 December 2007 (UTC)
    • Strongest possible support. While "merely" copy-editing, have occasionally checked a source to verify the intent of a phrase and possibly re-word it, only to find that what is in the source doesn't match the article. Example and example explanation on Talk. Not singling this particular article or editors out; there are many such examples. I had assumed that the GA process would have involved some fact-checking. Ideally, it should, or at the very least, FA should. This "shouldn't" be difficult: shouldn't articles be sourced such that any interested reader can consult the source and easily verify the statement? Of course, this isn't as easy when the source is (e. g.,) a copyrighted book whose text is not available online. Nevertheless, copy-editing shouldn't include fact-checking, but surely GA and FA should. As Ling.Nut points out, credibility is the crucial issue. Certifying articles as good, or as our best, then having discrepancies found between source and article, loses credibility and supports the critics. Unimaginative Username (talk) 04:21, 6 December 2007 (UTC)
      • Hezbollah is a heavily-edited article; editors routinely add text before a citation, separating it and making it appear to cite something different. Articles like that have maintenance problems unless multiple editors check daily for such changes. Gimmetrow 04:25, 6 December 2007 (UTC)
        • I can't emphasize strongly enough that I wasn't picking on the Hezbollah editors. The example was not a complete error, it was a slight mis-interpretation, probably made by a non-native speaker of English. Such writers should be commended for their fortitude in writing in a second language. All of us make misteaks -- all the more reason for some kind of fact-checking, perhaps random sampling, in FA and GA process. Sorry if it sounded like criticism of those editors. It was just the first example that came to mind. Unimaginative Username (talk) 06:12, 6 December 2007 (UTC)
          • No problem. Gimmetrow 17:37, 6 December 2007 (UTC)
  • p. s. Wouldn't it be better if each of the topics in this category had its own subsection, and hence its own "edit" button, for easier conversation and easier editing? Unimaginative Username (talk) 04:21, 6 December 2007 (UTC)

Timing

I propose to let this conversation run through to tomorrow and then summarize and suggest next steps. Mike Christie (talk) 12:51, 5 December 2007 (UTC)

Next topic redux

The topics which received the most support for current topic are the following three:

  • What can be done to make it easier for editors to make articles comply with the somewhat technical manual of style?
Three supports, and an oppose from Marskell, though I'm not sure that what Marskell was opposing was the intended meaning of this.
  • Should labels themselves be examined—stub, start, B, A, Good, Featured?
Three supports.
  • What can be done to ensure that content review processes evaluate articles for verifiability and to check the sources used are reliable?
Four supports, if I am correct in counting Awadewit and Sandy as supports.

There isn't a really clear preponderance of votes for any one choice, but I think the third topic, verifiability and reliability, is in the lead. I propose to make this the next current topic unless I hear any objections over the next 24 hours (i.e. post in this section if you feel this isn't sensible).

Once that's settled, I propose to start a brainstorming thread: see this thread from the PR discussion for comparison. I'll do that sometime on Friday (US Central time); however, I will be a bit short of editing time tomorrow so if it doesn't happen by mid-morning Friday, will someone else please jump in and start it? Thanks. Mike Christie (talk) 13:30, 6 December 2007 (UTC)

My comment on topic #2 was an implicit Oppose (viz., the topic is unnecessary). Sorry I didn't make it explicit. :-) Evaluating the implementation of the ratings might be OK; but I see no meaningful reason to examine the labels themselves, or alter them in any manner. Ling.Nut (talk) 14:02, 6 December 2007 (UTC)

Current Topic set to "What can be done to ensure that content review processes evaluate articles for verifiability and to check the sources used are reliable?"

I've updated the "current topic is" sentence on the project front page lead; and I've also updated the "current topic" section there.

The next step is brainstorming ways to address this. I may refactor as we go along, in order to group ideas together. As before I suggest the ideas should not be signed; this is brainstorming, not detailed critiquing. The goal of this step is to generate a bunch of ideas for discussion. Mike Christie (talk) 23:31, 7 December 2007 (UTC)

Brainstorming list

Please add ideas below. In order to get us started I have created a couple of ideas below, though I don't think they're particularly good ideas. Please think of better ones!

  • Track the verification of sources independently of the GA/FA work; have a separate flag to indicate this has been done.
    • Doubt seriously that a means of doing this could be found.
  • Expand the role of WikiProjects to require that a project performs the relevant verification before some status (e.g. FA) can be granted.
    • This relies on having a strong WikiProject backing an article, which is not always the case.
    • Very few WikiProjects would be up to this task.
  • Write a primer, similar to Tony's redundancy reducing exercises, explaining to reviewers how to spot and quickly check that reliable sources are used, that sources verify info, and that sources are correctly and completely formatted. It's not rocket science, but hardly anyone reviews sources and it's critical; perhaps a tutorial will help editors understand how easy it is.
    • Look for an existing primer (available in many universities)
      • There is something like that at Wikipedia:Reliable source examples, but I feel it could be improved. Awadewit | talk 13:39, 8 December 2007 (UTC)
        • That's more of a definition of reliable sources; I mean, explain to reviewers how to run through the citation list to easily spot problems and do a random check. The same thing that should be done to easily and quickly spot MOS problems; no one is doing either, and it's not that hard. I just don't understand what the issue is, so maybe a quick and dirty list will help, but when I did the quick and dirty on MOS issues, no one even paid attention, so <shrug> ... deficient FACs and GANs will go through. Lead a horse to water and all that. SandyGeorgia (Talk) 16:06, 8 December 2007 (UTC)
          • Oh, yes, I realize that. That is why I said "something like". However, I do feel that definitions are key to understanding whether the sources are reliable. You can't run the spot-check unless you understand what constitutes a reliable source, right? :) By the way, I didn't even know you did the MOS list - where is it? That should be more prominently advertised at the Community Portal, FAC, and GAC, in my opinion - like in the list of things to do before submitting. Awadewit | talk 16:11, 8 December 2007 (UTC)
            • Since no one paid attention to it or even commented on it, I db-authored it. It's long gone, and I don't even remember where it was. I figured either I was the only one who cared, or it was so bad no one wanted to say anything. SandyGeorgia (Talk) 16:54, 8 December 2007 (UTC)
  • Perhaps not everyone is doing either, but that's not quite the same as no one. But like Awadewit, I don't recall seeing any quick and dirty MOS list anyway. --Malleus Fatuorum (talk) 16:31, 8 December 2007 (UTC)
  • Correct :-) My frustration speaking. SandyGeorgia (Talk) 16:39, 8 December 2007 (UTC)
  • SG, are you saying that at one point, you had a condensed, quick-check version of MOS? I haven't been coopy-editing here that long, but have longed for such a thing (if it's what I think it is). I didn't dare suggest it here, as it seemed like a lot of work to ask someone to do. But it's a lot of work learning, and consulting, the entire MoS. Not to mention that for numbers, there are two completely different sources, MoS (numbers subsection), and MoS:Dates and Numbers. Please please bring back your quick guide. No one else may care, but you will have this editor's eternal gratitude. My frustration in c/e articles to the MoS joins with yours. (Be sure to link it prominently everywhere, especially at the top of the full MoS). Unimaginative Username (talk) 20:54, 8 December 2007 (UTC)
  • ah, ha ! Tito to the rescue. No, I don't mind, but I lost my edit conflicted response saying that I'm embarrassed about the state it's in because my prose stinks and it really really needs work, and when no one commented or offered to help, I assumed it was really bad and needed to be deleted. Most editors either generally comply with MOS, or have no idea about MOS, so it's not that hard to spotcheck an article to see if an in-depth look is warranted; the common errors stand out. That page wasn't *at all* ready for primetime; I threw it together very quickly, knowing that it had problems, just to see if there was interest. When there was no interest, I assumed it was so bad that people didn't want to hurt my feelings by saying so :-) I could easily write up some tips for spot checking sources, as I'm so frustrated at the number of FACs sitting there unreviewed for reliability or MOS. SandyGeorgia (Talk) 21:13, 8 December 2007 (UTC)
  • PS, I really wouldn't want it prominently mentioned anywhere until it can be polished and tweaked. SandyGeorgia (Talk) 21:17, 8 December 2007 (UTC)
  • I would be willing to help out with such a project, but, um, I can't see the page since I am a lowly content editor and not an administrator. :) Awadewit | talk 21:19, 8 December 2007 (UTC)
  • Same here; let's see how bad it is once Tito resurrects it :-) SandyGeorgia (Talk) 21:21, 8 December 2007 (UTC)
    Done: although the first time I tried it got deleted again as a zealous admin spotted the db-author within seconds! Geometry guy 21:36, 8 December 2007 (UTC)
  • I'd like to express my strong support for this too, and I'd also be willing to chip in. Is the resurrected page going to stay at User:SandyGeorgia/Article review? There's some excellent stuff there (nice work Sandy!); perhaps we should continue development in Sandy's userspace for now? EyeSereneTALK 14:49, 10 December 2007 (UTC)
  • I don't know where to put it, but it sure shouldn't become anything "quasi-official"; different reviewers review different ways, and FAC works best collaboratively—one editor checking sources and MOS, someone else is copyediting and looking at prose, someone knowledgeable on Images, and so on. I wouldn't want that page to go anywhere official, lest it sound/become prescriptive and other methods of reviewing go by the wayside. It's just one way of reviewing for some aspects of WP:WIAFA. For now, I suppose we can use the talk page there for tweaking it? SandyGeorgia (Talk) 18:12, 10 December 2007 (UTC)
The idea of a primer makes me think that there are two issues we have to deal with here: A) how to do the task, and B) who will do it. If the answer to A is "it has to happen at FAC", then perhaps it would be up to Raul and Sandy to check someone has done it. If the answer to B is "a certified WikiProject has to tag it as done", then they may develop different guidelines in different projects, since validating an article on Quantum chromodynamics would be quite different to validating one on slash fiction. Mike Christie (talk) 21:58, 8 December 2007 (UTC)
I think option (A) may be the most workable one. Looking at the other processes... expecting GA reviewers to verify references may be asking a little too much - not necessarily for reasons of workload, but it would seem to me to be stretching the credibility of a one-reviewer come-one-come-all system too far. The WikiProjects solution would be ideal, except that not all articles come under projects that are active, credible and have available members to do this kind of thing. Actually, I'd assumed this sort of check was performed as a matter of course at FA. EyeSereneTALK 14:59, 10 December 2007 (UTC)
  • There are some basic checks on references that even single GA reviewers probably ought to be doing routinely. Like checking for dead links for instance. --Malleus Fatuorum (talk) 16:44, 10 December 2007 (UTC)
  • For anything to be performed as a matter of course at FAC, someone has to be doing it as a matter of course at FAC. My review method was to spot check on reliability of sources and spot check that sources verified text before I spent time on article content. I saw one article while I was reviewing that got 25 Supports right out of the gate with unreliable sources. SandyGeorgia (Talk) 18:12, 10 December 2007 (UTC)
  • I absolutely agree - it relates directly to the factual accuracy/verifiability criteria, and it's no good asking editors for cites if a reviewer doesn't bother checking, if possible, where they go to. Having said that, I know I'm often guilty of randomly sampling links rather than checking them all. However, re-reading my above post, I didn't mean to imply that no reference-checking should take place at GA; many articles never get taken further, and it'll be no good for WP's credibility if the only articles we can say are fact-checked are featured ones. EyeSereneTALK 17:48, 10 December 2007 (UTC)
  • Malleus, on checking for dead links, I haven't fully figured out how to make this work yet, but there's a new tool installed at FAC and FAR that looks *very* promising. See this. SandyGeorgia (Talk) 18:12, 10 December 2007 (UTC)
  • Gimmetrow deciphered it for me: you check an article by changing the page parameter. Examples: [1] or [2] SandyGeorgia (Talk) 18:24, 10 December 2007 (UTC)
  • That's a really useful tool I think. I've just been running it raw" so to speak, by typing in the url and then the page name. --Malleus Fatuorum (talk) 18:33, 10 December 2007 (UTC)

[← & edit conflict]I am a little concerned with this, but only based on the sorts of articles I write and sources I use. Writing about events in the Peninsular War, at a level likely to achieve GA/A/FA ratings, pretty much requires me to use sources that aren't exactly common. Using sources like Sir Charles Oman's 7 volume history of the Peninsular War means that not many people can actually check that source to verify I've cited & interpreted it correctly. I've received comments on my talk page from respected and knowledgeable Peninsular War article editors saying they've never read Oman, for example, and having visited my local library today, my own personal library is possibly better than provincial libraries on this subject. Add to that the fact that it can cost well over £100 (and even as much as £250) in the UK to get hold of even a second hand version of Jac Weller's book on the subject (which I also use), it just gets worse. I fully support the idea behind the proposal, but requiring reviewers anywhere, whether GA/project A-class/FA, to verify the sources I use is a recipe for stalled nominations.

Of course, I realise that this is purely my personal experience, based on the stuff I write about, but I'm sure other article topics suffer similar problems: in order to get the level of detail required for the higher classifications requires specialist sources that simply aren't available to the average reviewer. The basic stuff that Malleus mentions wouldn't be a problem, of course, but any proposal that results in an FAC hanging around for months with no comments, just because no-one can verify the sources, is not a good thing. Just my tuppence worth, anyway. Carre (talk) 17:51, 10 December 2007 (UTC)

It's Joe'sPersonalWebsite.com and MarySmith.blogspot.com that I worry about. SandyGeorgia (Talk) 18:12, 10 December 2007 (UTC)
Ahh, well, if that's all this is about, then I withdraw my concern. I was thinking more along the lines of what happened with the Jacques Le Gris FAC (which I admit to supporting), where more dedicated reviewers than myself actually checked the reliability of the (published) source, and raised very legitimate concerns. At first glance, the source wouldn't have been questioned - it needed someone to actually go and look for reviews of the source/get the source to verify it. Other cases have also occurred recently where the citations were to reliable sources, but the cite didn't actually support the cited statement. This is, for me, the main problem with the whole citation system; WP:RS already disallows personal websites and blogs (although yes, many reviewers don't actually even look at the references), but what should be done about checking definite reliable sources to verify the use of the citation? Carre (talk) 18:28, 10 December 2007 (UTC)
You raise legitimate concerns, there is a clear limit to how much checking of sources can practically be done at GA/A/FA. But the saving grace is that GA/A/FA isn't a forever grading. Both GA and FA have their review processes, in which any editor can raise their concerns about sourcing and have them looked at in more detail. And articles are often demoted from both rankings as a result. --Malleus Fatuorum (talk) 18:43, 10 December 2007 (UTC)
(EC) Fact-checking is a problem where the refs are printed sources, no doubt about it. Having subject experts review will help, but as you say this will also backlog the backlogs, and we can't guarantee being able to match articles to suitable reviewers. On the other hand, do we then deny status to articles simply because we can't be expected to put in the same amount of work in reviewing as the article editors in writing? The best we can realistically achieve, I think, is to check what can be checked, and if that's OK then take the rest on trust... until (as Malleus says) someone points out otherwise. EyeSereneTALK 18:50, 10 December 2007 (UTC)
Fairy Snuff, AGF it is. Dedicated copy-editors like you, Eye, serve to keep me honest at least ;) Carre (talk) 19:07, 10 December 2007 (UTC)
Ha, compliments! I'd love to be able to take credit for your articles, Carre... but 'tis my lasting regret that I'm honest too :D EyeSereneTALK 20:19, 10 December 2007 (UTC)

FACT

The section break above was arbitratily imposed. Happy‑melon 13:02, 12 December 2007 (UTC)

We do all realize that Wikipedia will have to bite the bullet and establish a fact-checking project someday, right? People will actually have to check the sources for all these articles, particularly the FAs, I think. Everyone makes mistakes and citations get muddled and lost over time in the editing process. Awadewit | talk 23:26, 10 December 2007 (UTC)

WP:FACT. Gimmetrow 23:27, 10 December 2007 (UTC)
I'll rephrase - a visible fact-checking project. :) You would think I would have heard about this project after a year...Awadewit | talk 23:29, 10 December 2007 (UTC)
Maybe it would be a worthwhile goal to make this project more visible or active? Gimmetrow 23:33, 10 December 2007 (UTC)
I suggested a relaunch as "Fact and citation testing". Geometry guy 23:39, 10 December 2007 (UTC)

← Very good :) I agree with the above - this project must be one of WP's best-kept secrets - providing the project participants are willing. EyeSereneTALK 17:39, 11 December 2007 (UTC)

I'm very strongly in favour of relaunching WP:FACT. I think with the proper infrastructure and mission-statement it can easily form a parallel to the LoCE, but for verifiability. Imagine: at FAC, reviewers ask for a LOCE-copyedit for brilliant prose (1a), and a FACT-check to ensure perfect referencing. Duties of a FACT-check could include:
  • Checking all sources for reliability
  • Checking all sources for verifiability (ie that they actually support the statement)
  • Ensuring that the references are formatted properly and that there are no errors (this would include use of cite templates, checking ISBNs, etc)
Once this check has been completed, the talk page could get a nice little shiny messagebox like {{LOCEcomplete}}, which would indicate to reviewers that the verifiability of that article (or at least the reviewed version) is as high as possible for an amateur review. Sound good?
As an aside, who would support adding LOCE reviews to {{articlehistory}}?? Happy‑melon 12:16, 12 December 2007 (UTC)
I'd broadly support the above, with the reservation that requiring a copyedit and 'FACT' check at FA will inevitably introduce further bottlenecks into the process. However, it's my personal view that we need to seriously consider doing this anyway, and that FA is probably the place for it.
I also think it's a good idea to include LoCE copyedits in the AH (as further points in its development where a stable, well-written article existed), although if the above suggestion is adopted there would be no need at FA as this could be taken as a given. EyeSereneTALK 09:31, 17 December 2007 (UTC)

Section break

How much is it reasonable to expect of a reviewer? I wonder how many even do basic things like check that ISBN/ISSNs match the claimed source, never mind have access to the sources to check what they say? Not such a problem though when there's a strong WikiProject backing an article.

I'm inclined to believe until told otherwise that there are three sorts of poor referencing:
  1. Deliberate deception - an editor deliberately supports a statement with a source which either does not exist or does not actually support the statement.
  2. Innocent deception - an editor accidentally uses a nonexistent source, or makes errors in referencing (like using an incorrect ISBN).
  3. Use of questionable sources - an editor supports a statement with a source which does not actually support the statement, and/or is not a reliable source.
Of these, number 1 is both the most damaging and (fortunately) the most rare. Deliberate scams can be hard to spot - look how long we put up with Upper Peninsula War because the sources looked credible. Number 2 is more common but is fortunately fairly benign - I'd consider reference formatting to be a similar task to MOS-compliance or copyediting. Number 3 is (IMO) what we should really be looking to spot at the earliest possible stage, and certainly by FAC. I'll sleep on possible means of doing that. Happy‑melon 22:27, 11 December 2007 (UTC)
Another problem is when online references for an article change or are no longer available. Larrys Creek became a featured article in August 2006, and in August 2007 I went through and checked all the refs again. There were no dead links when the article passed FAC, but just one year later six of the 42 refs cited no longer worked (one in seven or about 14%). Three had a changed URL (so I just had to update to the new URL) and three were no longer on the web. I found one of these at the Internet Archive, but two newspaper article URLs were just gone (one for good, one at least for free access). Here is the diff [3], be aware I also added a new ref at the same time.
Any sort of FACT check tag on the talk page would be great, but things change enough online that this would have to be something that is checked again at periodic intervals. I am also someone who uses a fair number of realtively obscure print sources that would be difficult for most users to check, but that is a whole 'nother issue. Ruhrfisch ><>°° 19:15, 19 December 2007 (UTC)

Wikipedia:Reliable sources

I thought I'd start a new section on this as its central to our new topic. One of the great secrets of the P&Gs is that RS sucks. I argued at length here that it should be put out of its misery. Our principal wording on reliable sources is not and has never been Wikipedia:Reliable sources but Wikipedia:Verifiability. We were inches from redirecting RS to V, with about a dozen supports, until a couple of editors strongly objected and, because of exhaustion, nothing happened. The page is currently locked.

One obvious thing the great minds working here could do is figure out what's best for the guideline. Caveat: I am against guideline creep and we should not accidentally set it in motion. Marskell (talk) 12:30, 12 December 2007 (UTC)

It .. doesn't really add any new or distinct content. It should have been redirected, with any lingering bits and pieces of useful explication scattered to other pages. So.. that's what should be done... plus add a word or two to WP:V to make it abundantly clear that the "challenged or likely to be challenged" bar is a minimum threshold, not a gold standard. Then pass out the beer and cashews. Celebrate a day's work well done. Ling.Nut (talk) 12:56, 12 December 2007 (UTC)
My personal preference is for WP:ATT, but that's a different can of worms ;) I agree with Marskell and Ling that there's little on WP:RS that isn't covered elsewhere, and would certainly support any further moves to trim it out. EyeSereneTALK 09:44, 17 December 2007 (UTC)

Supposedly we didn't need the WP:RS page, it was useless and redundant and needed to be merged to WP:V. Yet now WP:EVALUATE is being created, with a strange emphasis on defining primary and secondary sources (see commentary on my talk page), yet covering essentially the same territory. I can't determine that this new page is anything other than a change of address for WP:RS, since RS is locked in disagreement, except that it seems to be attempting to elevate news media sources to the same level as the highest quality, peer-reviewed journal sources for scientific topics. If we didn't need WP:RS, why do we need WP:EVALUATE, which is re-creating the same territory? I've asked that on the talk page and gotten no answer. SandyGeorgia (Talk) 15:16, 20 December 2007 (UTC)

{{ArticleHistory}}

Other content templates

It strikes me that {{ArticleHistory}} completely deprecates a lot of templates:

And probably a fair few others. I'm not at all inclined to miss them - they're cumbersome, take up a huge quantity of space at the top of talk pages (the existance of {{skiptotoc}} indicates we put too much on the top of talk pages) and aren't nearly as easily machine-readable as {{articlehistory}}. However, I'd want to get a lot of support before trying to deprecate these entirely, as they're used on so many pages (1585 for {{GA}} alone). Does anyone else think this is a good idea? Happy‑melon 16:26, 18 December 2007 (UTC)

I'd support, but would comment that you really should ping Gimmetrow about it, and possibly SandyGeorgia too, for Gimmebot and FAC archive/promotion factors. Carre (talk) 11:05, 19 December 2007 (UTC)
I can see your point, but don't think it would be a good idea to delete these. ArticleHistory is a complicated template: it is hard to use, time consuming to update, and easy to make mistakes with (which generates work for the maintainers of Category:ArticleHistory error). The other templates are easier to use. Any editor, at any time, is most welcome to replace any use of these templates on an article talk page by an article history action, and some replacement is done automatically by Gimmebot. It would be good to expand the role of Gimmebot to autogenerate more article history actions. Ideally, these templates would simply be signals to the bot to generate an article history action, but that may be hard to implement in some cases. Geometry guy 18:25, 19 December 2007 (UTC)

Additions

On a similar vein, who would like to see what added to {{ArticleHistory}}? I'd very much like to see support for LOCE copyedits, and FACT checks once that project gets up off the ground again. Any other suggestions? Happy‑melon 16:26, 18 December 2007 (UTC)

I'd certainly support LOCE. Carre (talk) 11:05, 19 December 2007 (UTC)
That sounds like a good idea to me. I think I know the template code well enough by now to add this feature (if Gimmetrow is too busy) as long as you let me know what the possible action(s) and action-results are, and whether there is a useful action-link. Geometry guy 18:29, 19 December 2007 (UTC)
The template supports wikiproject approved versions (see Talk:The KLF and Talk:W. S. Gilbert for examples). I think it was envisioned (oh so many months ago) that LoCE and ACID and FACT work would be recorded as wikiproject reviews. (Related discussions here and here.) Gimmetrow 19:08, 19 December 2007 (UTC)

Quiet, isn't it? DrKiernan (talk) 12:40, 4 January 2008 (UTC)

Yes it is very quiet here, but this is a quiet time. I imagine that over the weekend and next week, activity will increase. Perhaps Mike will start a discussion about the next topic in a few days time. Geometry guy 00:09, 5 January 2008 (UTC)
I've been thinking about this; I'm busy at the moment with a couple of article-space (and non-wiki) things, but hope to get back to this some time over the weekend. One related question: are the peer review changes now thoroughly implemented and tested, and have they led to any insights about likely next steps that would be beneficial? For example, is automated archiving now higher priority? Mike Christie (talk) 00:32, 5 January 2008 (UTC)
Much has been learnt from the automation of the peer review page, and there may still be further insights. The main ones so far are positive: simplifying process reduces errors, and hence reduces time correcting errors; also editors seem to respond well to a change in process if it is clearly explained. Automation of peer review has raised issues about WikiProject peer reviews, but these don't apply to other content review processes. It has also provided the opportunity to make the automation more robust. However, I would not say it is thoroughly tested, as it has only been running for a couple of weeks. Nevertheless, I am confident enough in the technology to suggest it might be used at good article nominations. As for automated archiving, this is definitely a high priority now, and I will be pestering Gimmetrow soon about it ;-) However, extending archiving to a process like GA (which has no tradition of archiving GA reviews in a permanent place) is another story. Here the priority is probably automating ArticleHistory actions, and Happy-Melon has been working on this. Geometry guy 01:32, 5 January 2008 (UTC)

Facilitation approach?

I am not sure if my idea of how to facilitate this workshop is the best thing for it. Before we start another topic, here are a couple of comments.

When we have active contributors who have the energy to discuss something, and some of whom will volunteer to implement the resulting idea, then I think I can help by guiding the conversation through the steps we've agreed on. This gives everyone a sense of where we are in the process, and makes things more purposive. I believe (and hope) that this is a useful contribution to the workshop.

I believe that the key to a successful outcome of a conversation in this workshop is the participation of editors with the energy to implement ideas. Hence I'm not very comfortable with starting conversations via nudges; I am not sure it's ultimately productive for the project.

On the other hand, our process does assert that we will pick a topic, without waiting to see how many people are interested. We also have to remember that there's a "dead time" after an energetic discussion; we can't be productive on content review discussions all the time -- we have content to write too. So there are likely to be lulls.

Given all the above, my inclination is to wait, after a topic is discussed and reaches some end point, for either (a) a new discussion to start here, not necessarily involving me; (b) an editor to ask me to start another round of topic discussions (and perhaps Geometry Guy's and DrKiernan's notes above count as that); or (c) wait two months (or perhaps longer; experience will tell). If any of those happen I'll get involved and try to help facilitate. I don't want to facilitate a discussion that won't lead to anything, though.

Any comments on the above? No comments necessary, really, unless someone thinks I'm doing something wrong.

The current state of play here is that we were discussing verifiability, but had not conclusions yet that seemed to lead to a next step. I will see what I can do about restarting that conversation and the one about ArticleHistory; I'll probably post a note this weekend. Mike Christie (talk) 03:52, 5 January 2008 (UTC)

Peer review

OK, so I see the PR system is operational. Is CBM thinking of doing the archiving too?

Again, the issue is that there are almost exactly 3000 pages using {{oldpeerreview}}. The archiving approach currently in place would involve moving all those pages to /archiveN, creating 3000 more pages. Is this worth it? Or would it be better to work out a template-based solution that would handle archiving without involving a bot at all? Gimmetrow 08:22, 5 January 2008 (UTC)

Perhaps this is a server sacrifice we'll have to make to clean up the system. Don't forget, of course, that there are many more old peer reviews recorded in ArticleHistory, which also have pages that are incorrect. I think this is definitely a task for a bot, to work page-by-page through the PR archives, moving pages and then fixing the links as it goes. Happy‑melon 11:23, 5 January 2008 (UTC)
Archiving isn't on CBMs agenda as far as I know: VeblenBot doesn't specialize in that kind of activity (in fact, it doesn't edit outside of its own user space).
There are two issues here. The first is to clean up the current archives, and I agree with Happy-melon that this is worth doing. It probably wouldn't amount to 3000 new pages, as some have already been moved to /archiveN. (Also, I've added and documented an "archive=N" parameter to {{oldpeerreview}} to encourage proper archiving.) The second issue is to make archiving easier or even automatic, so that future clean-up is not necessary. The main source of difficulty is the page move from Peer review/ARTICLE to Peer review/ARTICLE/archiveN: I don't see how this can be done using an entirely template-based solution. Ideally, archiving should simply involve replacing {{peer review}} by another template on the talk page: a bot will then do the rest. Geometry guy 12:32, 5 January 2008 (UTC)

I had in mind something like this, where one template would be subst'ed in, and that would generate a template tranclusion with a parameter pointing to an appropriate empty page. Since the page would be fixed from the subst'ing, it wouldn't change. Automatic archiving from the get-go. Gimmetrow 01:42, 6 January 2008 (UTC)

Clever template. Yes, archiving from the get-go occurred to me too: it certainly has the advantage of no page moves. The problem with it is transcluding the archive to the PR page. There either needs to be a redirect to the correct archive, or the PR page needs to know which archive to transclude. Geometry guy 13:26, 6 January 2008 (UTC)
If you reorder the {{#ifexist: statements to run in descending order, and set up the subsitution right, when the template is substituted it will create a link to the highest-numbered existing page, which could be helpful. Happy‑melon 13:56, 6 January 2008 (UTC)
No need to reorder as far as I can tell. Anyway, this was just a concept. The idea would be to subst: a template called newpeerreview, and the subst code would generate code which looked like {{peer review|Somepagename/archive2}} (if /archive1 were taken). I'm not sure what the PR bot code uses to generate the page, but presumably it works from a category of PR requests. The bot could read all the talk pages and extract the parameter indicating the PR subpage, meaning another 100 or so page reads for each update. Another approach would be to have the PR category on the PR subpage. Not only would this avoid further page reads to update the main PR page, but removing the category could be used as a trigger to close the PR. Gimmetrow 15:56, 6 January 2008 (UTC)
This sounds promising. As you presume, VeblenBot uses the PR request category: in fact it doesn't read any pages, but simply queries the toolserver at tools.wikimedia.de. This allows for frequent updates (at the moment, one per hour). Removing the peerreview template results in the article being removed from the PR page, but the archiving has to be done by hand. However, if the category were on the PR subpage, then archiving could be done by changing the template to something like {{oldpeerreview|January 2008}}, and the archive page could also be automatically generated from a category. How does that sound? Geometry guy 16:22, 6 January 2008 (UTC)
I'm sure you mean changing the template to {{oldpeerreview|Somepagename/archive2|January 2008}}. But is that necessary? The bot can tell when a PR closes by its removal from the category and PR page. When that happens, just add whatever page disappeared to the archive. The bot could then change {{peer review|Somepagename/archive2}} to {{oldpeerreview|Somepagename/archive2}}, if that wasn't already done by hand. Gimmetrow 16:51, 6 January 2008 (UTC)
I had in mind that the template would be on [[Wikipedia:Peer review/Somepagename/archiveN]] (and would add that page to [[Category:January 2008 Peer reviews]]). Your idea is fine, but VeblenBot can only provide a list of the current contents of a category, so it would require a different bot to do that. However, something like GimmeBot is needed anyway if we want to update ArticleHistory when a peer review gets archived. Geometry guy 17:44, 6 January 2008 (UTC)
I can do a lot of different things with VeblenBot, and there are several other bot operators who would likely be willing to help. So please don't pick an inferior setup just to make it easier on the bot operators. On the other hand, if there is a good setup that requires minimal additional code, that would be great. Just keep me informed and I'll help with the bot code. — Carl (CBM · talk) 00:35, 12 January 2008 (UTC)

OK, so has anyone (GGuy?) had a chance to think through using the template to set up the subpages? At some point people wanted the peer review on the article page - this could be included in various ways, like a transclusion inside a hide/show box. A preload could be added. The main benefit is a bot wouldn't need to move pages around or update links in the log pages, much like AfD subpages. And all the old PRs and failed FACs wouldn't need to be moved around.

The disadvantage is that it would be a big departure from the way PR and FAC work. It would take some adjustment for people to "subst" the templates. The PR- and FAC-related templates would need some rewriting because most are assume the subpage will be at WP:PR/Articlename or WP:FAC/Articlename. Is the effort worth it? Gimmetrow 07:59, 12 January 2008 (UTC)

I am going to trial this approach at WP:GAR, which is in desperate need of any improvement to procedures. I've created a new version of {{GAR}} and four subtemplates to handle the listing of good article reassessments on a unique page. I will add archiving using pure VeblenBot technology. Lets see how this goes, and then ask whether this technology is the right way to archive peer reviews. Geometry guy 18:47, 12 January 2008 (UTC)

Automating GA

Just noticed the discussion at Wikipedia_talk:Good_article_nominations#Backlog and automation. There are ways to automate WP:GA without using a zillion categories. For instance, you could have people put a category in the {{GA}} template somehow (maybe using a parameter called something like "topic" ;) Whenever a page is found in CAT:GA which is not listed at WP:GA, the bot could grab the topic parameter from the article talk page and put the article in the right spot. Gimmetrow 06:54, 15 January 2008 (UTC)

Absolutely, I am well aware of that, and it would be a much more flexible system. However, it suffers from one key disadvantage: it requires someone to have the time and energy to write and test the code, and then operate the bot.
In the absence of such a volunteer, I am trying to see what I can achieve with the much more limited technology currently available to me: a category listing bot. So far, I have automated listing at PR, and (following your clever ideas) listing and archiving at GAR. The next steps are: archiving at PR, listing at GAN, and listing at GA. Certainly I would use a parameter like the topic parameter in {{GA}}, etc., but without having access to talk pages, I would have to use categories.
Category listing has the small advantage that it is relatively simple and robust, with low server usage, but I readily admit it is not ideally suited to GAN and GA. So if anyone wants to develop a better system, please go ahead. I would be happy to comment and contribute ideas. If not I will continue to see how far I can get with category listing. Cheers, Geometry guy 09:16, 15 January 2008 (UTC)
Either a category or a topic would involve the reviewer adding a specific parameter. Same work. Gimmetrow 15:27, 15 January 2008 (UTC)
I agree entirely: no matter what technology is used to autogenerate GAN and GA, nominators or reviewers would use a talk page template with a syntax like {{GAN|subtopic=subtopic}} or {{GA|heading=heading}}. But you seem to be missing my point: I am talking about the work involved to set up and maintain the bot, which is completely different in the two approaches.
If we use a whole bunch of categories (generated by some simple template code such as {{GA/Subtopic}}), then we ask CBM to list them using existing VeblenBot code. If we parse the talk page, we need someone to write the code to do this and maintain it. I know this is similar to what GimmeBot already does, but GAN is a fast moving page, which probably needs updating every half an hour, so there is some development work to be done. If this is easier than I imagine, then please, just do it. But I know that the valuable editors who code and maintain bots have the same time limitations as every WP volunteer: indeed we've been discussing various forms of automation since mid-October and it is really time now to start doing some of it. I'm doing my best with category-listing and template technology, which is all I have available to me at the moment. If others can do better, or provide me with more flexible tools, I would be very very happy. Geometry guy 19:54, 15 January 2008 (UTC)
I could certainly work out a bot-template combined system if you really wanted to set me loose on it (I think it's fair to say that what I did at LOCE is a marmite thing), and I could write the bot for it in python, but I couldn't take on the responsibility of running the bot. Happy‑melon 21:47, 15 January 2008 (UTC)
I would be delighted if you would write and demonstrate such a bot, and explain the concerns you have about running it.
My own experience in this story suggests to me the value of modularity, i.e., considering that a bot is not a complete self-contained solution to a wikipedia problem, but a tool for constructing solutions. This has the advantage that if the original bot or application fail (e.g. an editor leaves WP), it is not so difficult to replace. And such a bot can be used for multiple purposes. We have category listing, and it has proven quite versatile: are there other general purpose bot activities that would make it easier to automate processes? The idea is a very wikipedian one: share the workload among multiple editors as far as possible, Geometry guy 00:31, 16 January 2008 (UTC)

FAC issue

I'd planned to return to the topic of verifiability, but a conversation beginning on the FAC talk page seems to be an opportunity to engage on something people have concerns about right now. Take a look at this section, in particular to Rjgibb's essay on his user page, and Outriggr's comments on Rjgibb's talk page. Maclean25's point is a good one: we keep identifying a problem but don't seem to have solutions. Or is he right? Is the problem really identified? Mike Christie (talk) 00:03, 9 January 2008 (UTC)

  • The problem I see repeated is that people expect all FAs to be thoroughly reviewed but there isn't the manpower to do so. It would require several people to become quick experts in the subject in order to judge accuracy and comprehensiveness and a few people skilled in professional writing and familiar with the house style, and most importantly people willing to apply those trades in active editing each article. Currently we have a hodge-podge of over-stretch experts creating unbalanced reviews, due to a variety of reasons. --maclean 04:16, 9 January 2008 (UTC)
  • As I see it the next step is determine what we want to be. Do we want efficient reviews (serving as many candidates as possible), or equitable reviews (a fair review of each criteria to each article), or comprehensive reviews (that create a few truly complete articles)? In this regard, Raul654 has kept the FA system remarkably flexible. It is the community who decides what type of review to give and what to enforce (people are allowed to support without revealing that they didn't fact check for accuracy). But the overlapping GA system (which uses the same elaborate criteria - increasingly bogging down single reviewers as they move away from efficiency towards equitableness - combined with a discouragement of editing articles they're reviewing and a lack of reviewer follow-thru to FAC) and expanding number of open FACs has led to "resource starvation" (I think he calls it that). --maclean 04:16, 9 January 2008 (UTC)
  • Well, the problem which these essays highlight is an unhealthy obsession with MoS and its associated pedantry. There is an obvious solution to this: remove the formal "guideline" status from MoS, and modify the FA criterion 2 to say that "It broadly follows the style guidelines..." Then those who care about things such as non-breaking spaces can get on with it without wasting everyone else's time. Geometry guy 10:50, 9 January 2008 (UTC)
  • I felt the "guideline" that FAs should conform to the MOS was reasonable until two things happened: I realised that the MOS wasn't even near stable, and the rules suddenly started being applied to the contents of citations, as in the pedantry of mandating non-breaking spaces in the page numbers given in citations. I'm not saying that ought not to be done, but is it a valid reason to oppose an FA until it's "fixed"? No way. --Malleus Fatuorum (talk) 20:42, 9 January 2008 (UTC)
  • In defence on the MOS reviewers, anything that improves an article, however minutely, should be done before it reaches FA. FAC reviewers are meant to be pedantic reviewers, as FAs are meant to be as good as possible. Should a spelling mistake in an FAC be ignored just because it's a minor issue? Fixing MOS breaches is no big deal; it's very quick to do and there's no shortage of reviewers willing to show nominators how to fix them. I think a problem is that FAC nominators aren't aware of the benefits of MOS, and feel they're being asked to make pointless arbitrary changes. Epbr123 (talk) 21:38, 9 January 2008 (UTC)
  • I partly agree. However, if the change is so minor as fixing a spelling, then it actually takes less time to fix it than to comment on it. This is why I think it would be a good idea to shift the emphasis from describing MoS failures to fixing them: Mike has made a nice suggestion about this at WT:FAC#Some thoughts on FAC Geometry guy 21:45, 9 January 2008 (UTC)
    • A counter-argument is that it's better to teach nominators how to fix MOS issues themselves, so that they'll then be able to fix them themselves in their future articles, saving others from having to do it. Unlike writing brilliant prose, fixing MOS breaches is quite easy to learn. Epbr123 (talk) 21:54, 9 January 2008 (UTC)
      • I'm sure that nominators must get very tired of the endless "On a quick look through I found a MOS breach/spelling mistake/grammar error/something I don't like, so I'm sure that there must be others. Please get this copyedited". It really might as well be automated for all the help it is to anyone. --Malleus Fatuorum (talk) 21:59, 9 January 2008 (UTC)
        • From my own FAC nominating experience, it was frustrating when reviewers just said "get a copyedit", but with MOS issues, reviewers do tend to explain exactly what needs fixing. No reviewers expect users to go read all the MOS. Epbr123 (talk) 22:10, 9 January 2008 (UTC)
  • For me this raises another major issue about FA, which is that it focuses too much on nominator Qudos, and not enough on article quality. I accept that the Qudos of FA is a great pull for getting editors to improve articles, but it really becomes something of a club or clique if one needs to be trained as to what are the accepted standards before one can be admitted. It would be much better for the goals of WP if editors concentrated on the article, not the nominator. And, as Malleus points out, learning how to fix MoS breaches is not that easy, because MoS is not stable. Geometry guy 22:08, 9 January 2008 (UTC)