User talk:Noclip/Assessment overhaul

Not too sure about "outstanding" level
The GA process, by its nature, is really not all that objective. I think it would be really hard to make it fairly assign two different levels of assessment, since different people will have different ideas of what makes it outstanding instead of just good. -Amarkov moo! 04:28, 16 May 2007 (UTC)
 * Outstanding could be applied if a featured-quality article had been nominated for GA. It probably wouldn't be a common occurrence. Noclip 15:25, 16 May 2007 (UTC)
 * But that doesn't address my point. The process that determines if an article is featured takes lots of discussion and lasts over a month. How can a process that involves no discussion fairly determine if it's featured quality or not? -Amarkov moo! 21:59, 20 May 2007 (UTC)
 * I've clarified the description on the project page. The Outstanding/Great Article classification is not meant to be a substitute for the featured article process at all. It is meant to be a peer-assessed middle ground between the fairly relaxed standards of the GA process and the stringent demands of near-perfection imposed by the FA process. Noclip 22:15, 20 May 2007 (UTC)
 * I understand that; what I'm saying is that I don't think there should be another GA-like classification. -Amarkov moo! 22:17, 20 May 2007 (UTC)
 * I can see your reasoning but since the current Good Article standards are so broad I feel they could benefit from being narrowed if a second positive outcome isn't added. As for concerns regarding objectivity, they could be addressed by making a Great Article classification act as an endorsement of FA nomination by the reviewer, which would serve to curb excessive promotion into the category. Noclip 22:26, 20 May 2007 (UTC)

Comments
I remain unconvinced that the problems you highlight are really serious issues, worth trying to dismantle the system currently used for 630,000 article assessments, as well as the French system, and the Hungarian, etc. This doesn't mean we shouldn't change, just that we need an incredibly compelling reason to change.

Also, is everyone really just working to get an article to scrape through as a B? I'm not convinced. You criticise the current system as too rigid, yet this new system looks to me to be much more rigid. Am I not allowed to improve the English until the article is complete? Of course I am, but that does seem to be the implication. The current system can reward an article that is well written but incomplete.

I have the impression that you have misunderstood how the current system works. For example, you state, "I've seen at least two instances of the A-class rating being inappropriately applied after an article was not listed as a GA." What was inappropriate? The GA system uses one set of parameters, the A-Class is added by WikiProjects, so it is perfectly possible for A-Class articles to fail GA, and vice-versa (both do in fact happen). If a project rates it as A, then it is A. In contrast a project has no say in whether an article is awarded GA. This GA/A confusion is why many of us have favoured eliminating GA from the scale, but a recent debate didn't give clear support for the change.

This confusion seems to be also present in this new proposal. It seems to assume that WikiProjects will simply hand over all assessments for their articles to outsiders. These outsiders will not be experts on the subjects being reviewed, but they will have even more work to do than at present - and bear in mind the current backlog at GA. Many projects have lots of articles that are solid articles, but they don't care to spend the time trying to please an outside reviewer in order to get a GA or FA tag. The project will focus on the quality and completeness of the content, whereas a GA/FA reviewer often focuses on things like how many citations are included and formatted in the approved manner. So I would say that the GA or FA process is actually quite poor at judging whether or not an article is complete, compared to the A-Class review by a subject expert.

You do highlight one truism: In the early development of an article, most effort goes into adding content, to make it comprehensive. Only once it comes close to being comprehensive (B), does the emphasis switch from quantity to quality of content, though some content may still need to be added. People begin to look at citation, writing level, layout, etc. The separation of these two aspects is a very nice feature of this proposed system, although (as outlined above) I have reservations about the actual process proposed. In fact, some of the earlier systems proposed at the 1.0 project used multiple parameters like this; some systems used four or five (see this example).

It is my belief that the current system has one great strength (other than the bot) that has caused it to become popular - it is very simple. Much of what we do at the 1.0 project is Byzantine and confusing to all but the most seasoned reviewer, but the assessment scheme is easy for a newcomer to grasp. (I agree that the GA/A confusion is a flaw, though.) And I think there is a de facto separation into adding content (Stub &rarr; Start &rarr; B/A) and adding quality (B &rarr; A &rarr; FA), and this meets the need you describe. So on balance I don't see a compelling reason to switch to a more inflexible system that takes all quality assessment out of the hands of the projects. Thanks for a stimulating essay! Walkerma 20:22, 28 May 2007 (UTC)


 * I can appreciate many of your points but must argue that this proposal is less a dismantling of the entire system than a set of reforms that may seem daunting as a whole but are quite manageable if looked at individually. The changes under this proposal would at most consist of:


 * Bot tagging of B-class articles with a start template while preserving existing B classification during an extended transition period to allow for individual examination of which articles warrant comprehensive status.
 * Grandfathering in A-class articles as comprehensive.
 * Changes to GA review instructions to allow for "Great Article" outcome (or the Great Article classification could be eliminated entirely and this step could be ignored)


 * I recognize that project-assigned assessments are valuable in theory, but they are currently being widely ignored in favor of umbrella assessment. Individual projects could of course retain the right to use their own classifications to supplement the main ones, it's just that many (with a few significant exceptions) don't take advantage of the ability now. Still, articles considered to be of high quality by their parent project shouldn't be presented as better than those considered to be of similarly high quality by the community (which is currently the case with GA and A). As a compromise, the addition of multiple project attributes to the comprehensiveness assessment template would allow for each project to rate the same article differently, and for the overall rating to be different still (note that this is just a proposed compromise, not part of the original essay). I embrace the idea of each relevant project assigning its own comprehensiveness assessment, but this shouldn't apply to "remarkable" assessments since the criteria for GA and FA uniformly apply to articles under any WikiProject.


 * I must disagree with you on the point of rewarding well-written but uncomprehensive articles, as a stub with impeccable prose is nonetheless a stub and thus has minimal encyclopedic value. Noclip 21:06, 28 May 2007 (UTC)


 * Well, no. Fundamentally, the major change here is not the trivial fiddling with how levels are defined, but the essentially total removal of WikiProjects from the system.  The idea that a WikiProject's role must be limited to determining whether an article is "comprehensive"—never mind that it be unreadable garbage—is absurd; there is far more involved in creating a proper article than merely shoving all the relevant facts onto a page.  The system would therefore be useless to said WikiProjects—so why would they participate in it?  And, if not them, who exactly is going to run this vast effort?  (You do realize just how much behind-the-scenes work is needed to keep an assessment system running for nearly a million articles, yes?)  Quite honestly, I get the sense that this proposal is not founded on a thorough understanding of how the existing system works and what its goals are. Kirill Lokshin 22:14, 28 May 2007 (UTC)


 * I think you may have misunderstood something. WikiProjects would still responsible for the Stub, Start, and Comprehensive classifications. A-class is eliminated, so nobody needs to maintain it. As for GA and FA, WikiProjects never had anything to do with them and there's no need to change that. The net change in WikiProject workload would be nil. Noclip 02:42, 29 May 2007 (UTC)


 * I'm not talking about workload; obviously, your proposal involves less workload—at the cost of less fine-grained control. The entire point of an A-Class rating is to give projects something to use that would be independent of the lowest-common-denominator politics that go on with GAs; removing it takes that away from the projects.
 * (This aside from the fact that B-Class and "comprehensive" are not necessarily the same thing, and, in some cases, quite obviously different things.) Kirill Lokshin 07:12, 29 May 2007 (UTC)


 * Look at the current usage of A-class, many A-class ratings are inappropriately applied by individual users who are not part of the articles' associated WikiProjects. A-class is a good idea in theory but it needs to at least be renamed to something more concise to encourage correct usage of it. Noclip 13:04, 29 May 2007 (UTC)


 * That's the problem of the respective WikiProjects whose ratings are being fiddled with; take it up with them. Certainly there are working models of how A-Class ratings can be done (c.f. WP:MHR), and they shouldn't go to the chopping block simply because other projects are incapable of keeping their house in order. Kirill Lokshin 16:01, 29 May 2007 (UTC)


 * The elimination of A-class as a core classification wouldn't affect individual projects' ability to use it, much the same way that even though there is "officially" no future-class it is still used by WikiProject Films. Noclip 19:33, 29 May 2007 (UTC)

Sub-Standard Articles
This point may fall outside the scope of the proposed Assessment overhaul, but I wish to propose the creation of a new category to cover what I think are sub-standard articles which could be catagorized as possible candidates for deletion.

I am currently assessing articles as part of the Business and Economics Wikiproject, where there is a high incidence of spam. For instance, there are many companies seeing to increase their web profiles by creating articles about the companies themselves or their products, particularly amoungst IT companies such as software vendors.

As I am assessing quite a few spammy articles, it would be useful to tag them as being sub-standard. I know that the various deletion routes for deletion offer various ways to tag such articles, but with over 1,000 articles unassessed, it would be useful not to go through the (relative lengthy) deletion process while I am in assessment mode. However, once taged as "Sub-standard" or "Candidate for deletion", either myself or my collegues can drop back on this category at our leisure.--Gavin Collins 23:22, 2 June 2007 (UTC)
 * Just PROD them. That's not lengthy at all. &spades;P M C&spades; 04:01, 3 June 2007 (UTC)


 * Actually, PROD can is lengthy compared with the (relative quick) assessment process. PROD has to be substituted with subst:prod|reason and the creator of the article has to be warned of the proposed deletion. Then there is a five day wait for deletion, unless of course the article deletion is contested, in which case it goes on a lot further(AFD). But the point I am trying to make is how do you assess articles that are so bad that you don't want to give them so much as stub status? Also, would it not be better to have fellow Wikipedia Project members to review your proposed deletions before going throught the deletion process?


 * I can see why you might want to get rid of A and B class articles, but I think there should be a category for sub-standard articles. I should warn you though, there is some history regarding the categorisation of sub-stubs; but not as a negative category, but as another type of stub category. --Gavin Collins 19:57, 3 June 2007 (UTC)


 * The creator doesn't have to be notified; that's a courtesy, not an obligation. As for assessing articles that are that bad...speedy them if they're nonsense, blatant ads, or not notable. There's very little that's faster than a speedy delete. I just can't see the point of tagging something as "sub-standard" when you could be tagging it for deletion via PROD or speedy. After all, no matter when you try to get the article deleted - after assessment tagging or not - you still have "a five day wait for deletion". &spades;P M C&spades; 19:58, 4 June 2007 (UTC)
 * watch out about those speedies: non notable applies only to people, groups, bands, web content. It cannot be used for an article on anything else, though people do try. Furthermore, it can only be used if the article makes no assertion of importance. Apparently even an assertion that will clearly not be accepted at AfD still counts to prevent a speedy--I've tried to se it when people claim without any reason given that their college club is notable, and had the tag removed. What this amounts to, is articles about the subjects listed which are undeniable un-notable.
 * as for nonsense, the bar seems rather high. I've tried to do it for pretty stupid articles, and had the tags removed by the more experienced admins. It basically seems to apply only to gibberish and be useful for newpages.
 * Personally, I regard anyone placing a speedy or a prod and not notifying the author as a WP:DICK. How would one like it if it were done to oneself? Many of us have a great many articles watchlisted for various reasons, and this sort of a change is, to say the least, more important than most. DGG 19:24, 24 June 2007 (UTC)

Too corase
The leap from "start" to "comprehensive" is too big. You have to expect that few readers will assign the very highest rating, because they will assume it means perfection, so there has to be an intermediate rating (currently B-class) to cover articles that are better than "start" but not perfect. &mdash; Carl (CBM · talk) 18:12, 3 June 2007 (UTC)


 * It depends on how you define "start". The definition in use right now is essentially "long stub". Under this definition the vast majority of articles are B-class, but the proposed "start" is meant to be a cover a much broader range. Noclip 18:44, 3 June 2007 (UTC)