User talk:Christopher Parham/FA-GA

Merge GA with FA? A-Class?

See revised proposal section below

I just wanted to moot some discussion here before I did anything rash and MfD'd WP:GA, but it seems to me from reading the GA criteria against the FA criteria, there is a rapid trend in the continually evolving GA project policy towards convergence with the FA process. The key catalyst that caused me to notice this was the relatively recent strict rules adopted by GA requiring adequate citation for all GAs. It appears to me that the only major substantive difference between the two mechanisms is the approval process; for all other intents and purposes the content requirements are nearly identical.

If this is so, why not consolidate these two units together, and gradually review all GAs, a la Wikipedia:Featured articles with citation problems, for promotion to FA status. Because otherwise all I can differentiate between GA and FA is that one is better for instant gratification.

(edit): I'd also suggest that GA's which fail FA criteria in a merge be re-classified as A-class articles. This means that all articles can be individually assessed at any class level, with only one (final) candidacy step in the process, for FAC.

Thank you for your time, Girolamo Savonarola 22:48, 17 February 2007 (UTC)

Seems that a link to Wikipedia:Compare Criteria Good v. Featured is appropriate here.

Generally, I agree with this proposal to merge the two, but I would keep separate criteria for each on the same "guideline". (Why are none of the Featured and Good article pages tagged as guidelines?) —Doug Bell ^talk 23:47, 17 February 2007 (UTC)

I like the idea. HighInBC ^{(Need help? Ask me)} 23:49, 17 February 2007 (UTC)

The difference in approval process is a major distinction and quite likely a useful one. The GA process is far more scalable than the FA system, which is never going to assess more than a tiny fraction of the articles on Wikipedia. We should see much quicker results from GA. We haven't thus far, however, for which one can offer two reasons:

1. The GA process is bloated. It is not clear why a central candidate page exists, when all the work takes place on the talk page, but having this page certainly substantially increases the difficulty in nominating articles.
2. Further, for many articles its purpose has been obviated by effective WikiProject assessment systems.

At the moment the GA process is probably too similar to the FA process to be adding much benefit to Wikipedia. As far as a "merger" goes, it's not clear what that would entail, but assuming that we don't want to change the FA process, it must basically mean scrapping GA. I think it would be worth at least attempting to go back to the original, simpler GA system, thus creating a clearer distinction between the processes. Christopher Parham (talk) 01:13, 18 February 2007 (UTC)

I oppose any merger; I think there's a definite difference in actual quality, even if it's not expressed clearly in the criteria. Have a look at GAs which have failed FACs to see some of the differences. Trebor 01:52, 18 February 2007 (UTC)

So you just think that nothing whatsoever should be done? Girolamo Savonarola 15:25, 18 February 2007 (UTC)

I don't know; I'm not really involved with the GA process (other than putting a couple of articles through). But many Good Articles come to FAC and fail, so there is a definite difference in standards (even if it isn't apparent from the criteria). FACs only work if there is an editor willing to work on the article, and if GAs haven't already been put through FAC then it implies there is not. Trebor 15:51, 18 February 2007 (UTC)

I don't doubt that there is some difference, but I am questioning if it is enough of a difference to warrant the entire GA apparatus. At the current time, I'd say not. I'd merge the two projects as per my original comment, and abolish the GA class. This would benefit the assessment structure as well, since the articles can still be assessed as Stub, Start, B, or A class by a single individual, based on criteria. So I'd presume that most of the GA's which fail FA in a merger would likely be recategorized as A-class. So structurally, the article is individually assessed from Stub thru to A class, and then if the article is deemed good enough, it only has to jump through one candidacy process - the FAC. Much more linear, simple, and less bureaucratic. Girolamo Savonarola 16:03, 18 February 2007 (UTC)

That sounds less like a merge and more like an abolition of the GA-class and framework, which would warrant an MfD (note I'm not saying I'm necessarily against that, just saying what it sounds like). There has been a slight lack of clarity in the assessment criteria with GA- and A-class being on different scales, and overlapping to a large degree, so it would be good to clear up where they fit in. But a blanket put-through of all GA articles to the (already fairly overloaded) FAC process, regardless of whether there are editors willing to improve them, would be a mistake in my eyes. Trebor 17:54, 18 February 2007 (UTC)

• fix, don't merge. The GA idea was to have a quick and unbureaucratic way of assessing articles that are clearly good, without meeting the formal FA requirements. Now, it appears a whole bureaucracy has accreted around GA assessment as well. The solution is to get rid of it, and turn "GA" back into what it was supposed to be. There are many, many articles on WP that are good without being likely to become FA anytime soon. "GA" to my mind is a tool to facilitate measuring of the distribution of quality on Wikipedia. I say, leave FAC as it is, an assessment on the very best on WP, and turn GA back into something simple and unbloated. dab (𒁳) 16:03, 18 February 2007 (UTC)

That simple process, in my mind, is more or less equivalent to giving an A-class assessment. See my above comment and my revised proposal. Girolamo Savonarola 16:04, 18 February 2007 (UTC)

• No merge. The difference between GA and FA is the difference between one set of eyes and several set of eyes. Hence the FA process is more like a fine tooth comb that can get to the nitty gritty assessment pieces. The purpose and benefit of the GA process is that with a single reviewer you will obviously get assessment and feedback much quicker. Of course the quality of that review is dependent on the quality standards that the GA reviewer is upholding. The push towards stricter citation is a positive development because more articles that are actually good are being recognizes and more articles that are substandard are being improved and brought to compliance with simple policies like WP:V and WP:NOR. The GA assessment process is more approachable than the FA process and is a good way for new editors (or editors new to FA) to become familiar with a criteria similar to FA but only have to deal with the feedback of one reviewer. As the overall quality of GA reviews improve then you will see more GAs succeeding at FAC. Agne^Cheese/^Wine 16:26, 18 February 2007 (UTC)

So what would the argument be against simply folding GA into the A-class for assessment, then? Since they both require just one reviewer, and have similar criteria. Girolamo Savonarola 17:44, 18 February 2007 (UTC)

Revised proposal

Here is my revised proposal based on discussions above. There seems to be a view that the GA has become overly hampered down with bureaucracy, and to many extents apes the FA criteria and structure to a high degree, with the most notable difference being the single reviewer of GA versus the community of reviewers for FA. GA also has been a curious question in regards to its somewhat incongruous shoehorning into the assessment classes (it should be noted that the GA class was not originally proposed in assessment).

Given the more rigorous GA standards from the past, its single reviewer characteristic, and the unnecessary bureaucracy, what I propose now is a merge of GA into A-Class assessment. The standards for the two, content-wise, are nearly identical, and like GA, assessment only requires a single reviewer to judge the article against the class criteria. It makes article assessment classes more straightforward, with all classes up to A being solely based on assessment, with a final bureaucratic candidacy process only required for the top distinction, FA. Based on the current criteria, it is likely that most, if not all, of GA-class articles would qualify for A-class easily. It is also much easier to implement than kicking up the current GA's for (gradual) integration into FAC, which has been noted would be a problem without an active editor. Reassessing GA's into A-class would not face this problem.

I look forward to your comments! Thanks, Girolamo Savonarola 18:19, 18 February 2007 (UTC)

I would support that in principle, although I'm not really involved with the GA process (you may want to leave a message on GA talk). However, the assessment process is fairly haphazardly applied too, depending on the activeness of the Wikiproject (for instance the MilHist Proj seems to have a multiple-user assessment for A-class; others barely assess at all). Given my lack of familiarity with the issue though, I will see what others think (particularly if they can explain the need for, and differences between, GA and A class). Trebor 18:53, 18 February 2007 (UTC)

• We seriously do not need three different classifications for the quality of articles. >Radiant< 13:05, 19 February 2007 (UTC)

Since GA has already been shoehorned into the assessment system, perhaps the rating can be repurposed as effectively a 'B+' grade. There is currently too large of a chasm between A-class ratings (which are rare, and have an associated process in some wikiprojects) and B-class ratings (in practice used for a large range of article qualities that meet the general description of 'okay', which is not at all consistent with the rating's description). This classification could imply that the meat of the GA criteria are met (neutral, stable, referenced, reasonably complete) while allowing things like omission of minor content details, mixed referencing styles or minor formatting issues, or some prose problems, all of which (I think) would generally disqualify an article from an A rating.

I'd like to see most of the bureaucratic apparatus of GA scrapped, and what remains repurposed for its original intention: identifying excellent short articles. Current practice seems to be to call almost all short articles 'start' or 'B', on the assumption that they need expansion, but some topics just don't require more than a few paragraphs. Current practice also essentially blocks these articles from FA status, with the odd rare exception for a hurricane article. Whether this assessment class should also become part of the rating system is not obvious, as the existing ratings are not length-dependent and the 1.0 project might be too far along to permit adding or removing ratings at this stage. Opabinia regalis 17:46, 19 February 2007 (UTC)

You want another rating? Frankly, I think some simplification is necessary. The larger chasm (IMHO) seems to be between Start and B-class. I'm not so concerned with whether or not B and A are massively different, so long as the grading scheme offers some clear identifiers (which I believe it already does, for the most part). These aren't real grades, and no one is being judged, so the need to be so precise is actually adverse to the idea of a grading asssessment (especially given that the articles are constantly evolving). It's a very coarse-grained way of tracking the general progress of a group of articles even moreso than it is used to track any individual one. Any more attention paid to how to exactly quantify the ever-shifting state of an article seems like it would only divert energies more properly spent working on the articles themselves.

The idea is to have a coarse-grained and efficient system requiring a minimal amount of effort. Assessments shouldn't require long perusal of a given article - it should generally be obvious on a quick skim. Making more subtle gradations simply adds more time to the assessing editors' evaluations, and furthermore is likely to adversely affect the willingness of many of them to continue assessments, especially if more rules and grades are being added. The only level at which any prolonged effort should exist is the FAC. Thank you, Girolamo Savonarola 17:59, 19 February 2007 (UTC)

Well, yes. There are currently six assessment ratings. Two (GA and FA) have separate processes and can't be assigned by an individual assessor. A third (A) is rarely used, and some projects also confine its use to articles that have been reviewed by multiple people. That leaves an individual assessor with stub, start, and B. Since stubs are usually unambiguous, assessors sort articles into start or B. Surprise! We have an encyclopedia full of start and B-class articles. I'm sure it varies between wikiprojects, but I see much larger variations in quality within than between these two rating classes, and a large gulf in quality between B and A.

So if I were designing the rating system from scratch, it would have five levels available to assessors, of which the lowest (stub) is unambiguous, and the highest (A) is regulated by the corresponding wikiproject. (Obviously FA retains its own process.) That leaves an individual assessor with a practical choice of start, B, or B+/GA/whatever, which I suspect would help in distinguishing between 'usable' articles and 'raw material' articles. It also eliminates the problem of what to assign the current crop of GAs (and GA-quality articles) if the GA process is shut down or reformed. I would not agree that all current GAs are A-class, given the way the A rating has mostly been used. (It's possible, of course, that the problem is underuse of/overly variable standards for A ratings.) Opabinia regalis 18:29, 19 February 2007 (UTC)

Well, to be succinct, I believe that another rating is equal to more bureaucracy. If that's the case, then there's no point. I think you also underestimate both what a B article is and what an A is (and would be under this proposal). They should be clearly distinct. And yes, that may mean more clarity on the grading scheme definitions, but that is a lot easier than forcing yet another grade upon everyone. Please also think over my comments above about the process needing to be coarse-grained. I believe that those sentiments were reflected during the actual creation of the assessment schemata. Girolamo Savonarola 18:39, 19 February 2007 (UTC)

Keep in mind, also, that some projects are adopting more well-defined standards for B-Class (e.g. here). Kirill Lokshin 18:54, 19 February 2007 (UTC)

Yes, please! This will significantly simplify the overall structure of the various assessment schemes (since they won't be structured as internal → external → internal → external, but rather as internal → internal → external, going outside a project only for the final FAC) and get rid of the GA backlog/bureaucracy/etc. issues at the same time. Kirill Lokshin 18:54, 19 February 2007 (UTC)

Does an assessment from one Wikiproject apply for all? For instance, if WikiProject Biography assessed a military figure as A-class, would that be accepted by MilHist? Trebor 19:24, 19 February 2007 (UTC)

In general, it depends on the projects involved, and whether they have particular requirements for a particular assessment level. In your specific example, not necessarily, since MilHist has a formal review process for A-Class (it would, of course, be a fairly good indication that the article ought to be submitted to that review); conversely, an A-Class rating from MilHist is often copied by other projects that don't have any sort of formal review. Kirill Lokshin 19:30, 19 February 2007 (UTC)

Okay, that's what I would've thought. I'm just trying to think things through with regards to rating. One advantage to GA is that articles are centrally reviewed, so GAs are roughly the same quality across the project; considering the difference in activeness of Wikiprojects, A-class could end up being applied rather inconsistently. Although maybe that's not a problem, as you could take into account the processes of the project when considering the "authority" of the rating. (Excuse me thinking out loud here.) Trebor 19:39, 19 February 2007 (UTC)

That's true, I suppose. (But GAs do tend to vary quite a bit depending on who's doing the actual reviewing; so I suspect that the consistency is actually pretty similar across both processes.) Kirill Lokshin 19:44, 19 February 2007 (UTC)

With the over-reaching criteria thing, a big problem with that is that discussion is rather difficult to start. Not because there's some group of people forcing GA to go one way with things, but because many people involved don't contribute to the discussions unless something really nasty happens, there's just so many candidates on the list it takes up a bunch of time :/. I for one have some things i'd like to change with the rules so that they'd go back to older, simpler versions, but I dunno how to start the discussion when sometimes people don't pay attention, and often times large chunks of rules get changed based on the discussions of maybe 3 or 4 people. Not that anyone's trying to make things bad on purpose mind you, its just discussion of process isn't very good yet.... Homestarmy 20:24, 19 February 2007 (UTC)

I don't think that anyone is acting out of bad faith, but my whole point is that the GA process is massively inefficient and in some ways counterproductive to other larger processes. I believe that it has some merits, but it also had the bad luck to be created at about the same time, but with little true coordination with, the assessment project for 1.0. Had assessment been started a year earlier, I wonder if GA wouldn't have instead been the efforts of people to create more rigid A-class standards. And with regard to consistency of A-classification across different WikiProjects, surely we can all agree that this can be much more easily improved with tighter and clearer assessment guidelines? Girolamo Savonarola 20:34, 19 February 2007 (UTC)

The thing is, many of the things making it more inefficient resulted from the 3 to 4 people discussions i'm talking about, and when the changes happened, not many people really said much. If there were more people to talk about things then there's certainly several rules I for one would like to see reverted to earlier versions, when GA wasn't really as inefficient. But when sometimes I propose things and maybe one person responds, (Like when I got the MoS criteria changed) it makes me feel like it would be a bad idea to just do something like that and change rules back when not many people might notice, it already has confused things a good bit in the past when that kind of thing happens. Homestarmy 20:54, 19 February 2007 (UTC)

Yeah... everyone's going to want to get their hard-worked on article to be good, instead of A-class. It's an issue that's occurred to me before.--h i s s p a c e r e s e a r c h 17:07, 20 February 2007 (UTC)

This discussion seems to have lost momentum, which is a shame because I think this is worth trying to sort out. Do people have objections to merging GAs into A-class, and think GAs are worth keeping separate? Trebor 16:32, 21 February 2007 (UTC)

No objections to deprecating the GA rating in the assessment scheme, though I would oppose default mass migration of GA to A. That's something for wikiprojects to handle. If the goal is to disassemble the centralized GA bureaucracy, the proposal should be posted there, and probably at WT:FA too. Opabinia regalis 04:30, 22 February 2007 (UTC)

Sorry, I just saw this discussion for the first time! It arises from the fact that the GA level as in the assessment scheme overlaps with B and (mostly) A, leading to ambiguity and confusion. I am proposing to remove the GA-level from the WikiProject assessment scheme, but have WP 1.0 bot read GAs so that WikiProjects have all GAs listed in their "WikiProject Foobar Articles by Quality" worklists. I think that should keep everyone happy and resolve this problem. Please read the proposal and leave comments. Thanks, Walkerma 04:52, 22 February 2007 (UTC)

• I suppose merging A-class with GA-class for these purposes. Considering that "A" and "B" are the only letters that the system actually uses (it uses "Start" instead of "C", "Stub" instead of "D", and nonexistent articles would presumably be "F"), rendering their use kind of arbitrary and potentially opaque (especially to people unfamiliar with the A/B/C/D/F grading scale), it might be worthwhile to replace "A" with "GA" (or perhaps "Good", to mirror "Start" and be more descriptive to people unfamiliar with WP:GA), and to replace "B" with something that is also more descriptive. Perhaps "OK"-class. (Another, related problem that is probably more significant is that "Start" and "Stub" start with the same first two letters; this is obviously a very bad idea.) -Silence 21:22, 22 February 2007 (UTC)