User:Marnen/Pickypedia

Pickypedia: a proposal

I recently read The Great Failure of Wikipedia by Jason Scott, and it got me thinking. I don't agree with most of the criticism in Jason's article, but it seems to me that he got one thing right: he noticed Wikipedia's inherent lack of reliability. One of the wonderful things about Wikipedia is that our editors include all sorts of people with extremely diverse expertise, viewpoints, and writing skills. However, this is also our great shortcoming: since anyone can edit any article, anyone can post any sort of content they like, regardless of its relationship to the truth -- and the next person to read the article may well take it as gospel.

Now, don't get me wrong: I really love Wikipedia, and I'm not planning to leave it. I love its open spirit, and I love how excellent articles often emerge out of many people's incremental changes. Unfortunately, all it takes is one vandal or POV-pusher -- or one well-meaning, misinformed person -- to ruin an excellent article. All it takes is one ruined article to damage our credibility as a source of information. And if Wikipedia loses its credibility as a source of information, then it loses its reason for existing: an unreliable encyclopedia is perhaps worse than none at all.

The standard Wikipedia response to this, of course, is that we can always revert the damage, and this is certainly true. However, this does not address an important question: what happens to the people who read the article between the moment the damage is done and the time that it is reverted? Sure, people will recognize statements like "Italy is in Asia" or "TH|S ART1X0L IZ 0WN3D BY H/\X0RD00D!!!!!!111!!!!!" as out of place, but is the same true for statements like "J.S. Bach was born in Berlin in 1686"?

Unfortunately, this is not a trivial question. Even though IBM's famous study found that most vandalism on Wikipedia is repaired within five minutes, I have often run across vandalism that persisted for hours -- or, in one case, over a day. And that's only scratching the surface: I am only referring to obvious vandalism, not things like the Bach example above. Besides, Wikipedia is an extremely high-traffic site -- in September 2004 (the latest month for which I have found statistics), there were 6 million hits on the English Wikipedia alone, or over 138 hits a minute on average. With that much traffic, even a few minutes of unreverted vandalism is a potentially serious problem for our reliability.

Clearly, what is needed is a way to improve the reliability of Wikipedia, and thereby its utility as an encyclopedia, without losing its openness and community spirit. I am proposing an experiment along these lines, which I am tentatively calling Pickypedia (I'd love other name suggestions).

I am open to suggestions on how Pickypedia will work; at the moment, here's the model I'm thinking of (rather like the Gutenberg proofreading or Zeal models in some ways):

• There are three levels of users (or more if it's found necessary). For the moment, I'll call them Users, Editors, and Sysops, in ascending order of privilege. The software used is the MediaWiki software or something similar, allowing history and talk pages.
• Namespaces are very much like on Wikipedia (main:, talk:, user:, user_talk:, pickypedia:, pickypedia_talk:, special:), with the addition of a new pair of namespaces which I'll call unreviewed: and unreviewed_talk: (or x: and x_talk: for short).
• Everyone starts out as a User. Users can edit their own user page, any talk page, and also any page in the unreviewed: namespace, which would act basically the same as the main namespace in Wikipedia. (I'm not sure how pickypedia: should work; for the sake of this model, I'll simply disregard it from now on.)
• Once certain elementary criteria have been satisfied (perhaps having 100 changes approved by an Editor), a User may become an Editor if desired. Editors can edit in any namespace including main:, and can also approve Users' changes made in x:. Approved changes are applied to the version of the article in main:. This way, many articles exist in two versions: a bleeding-edge version in x: and a more stable Editor-approved version in main:, somewhat like the common programming practice of having both unstable and stable builds at any given time.
• Sysops perform systemwide administrative and maintenance tasks. I think I favor having them and only them being able to ban users, and perhaps even having them be the only ones who can edit in pickypedia:.
• It may be desirable to have Reviewers, "super-Editors" who make sure that the Editors don't do anything stupid. I'm not sure about this; I don't want to increase the bureaucracy more than necessary.

So...that's my proposal. I'd love to hear your comments. --Marnen Laibow-Koser (talk) 15:32, 29 Apr 2005 (UTC)

• There's a proposal to make a DVD or CD-ROM of Wikipedia content. This process already requires fact-checking and confirming the contents of article. Jimbo Wales had a link to this proposal on his userpage the last time I checked it's called "pushing to 1.0". On another note; no source is infallible and people should double check their sources anyway and while we like Wikipedia to be perfect there's no reason why Wikipedia info shouldn't be verified when it's used. Mgm|^(talk) 20:59, Apr 29, 2005 (UTC)
• The proprosal is Wikipedia:Pushing to 1.0. German Wikipedia has already done something similar[1]. Samw 04:19, 30 Apr 2005 (UTC)
  • Nice as far as it goes, but this seems to be talking about a paper or CD-ROM edition. I'm still talking about an online encyclopedia, just one with a little more editorial control. H2g2 may be the closest thing out there at the moment, though I'm not certain. (And Mgm, I know no source is perfect. That's not really my point.)
  • Can we take further discussion to the talk page? --Marnen Laibow-Koser (talk) 04:52, 30 Apr 2005 (UTC)

• I don't really have any comment directly on this, but you might also want to have a look at Wikipedia:Researching with Wikipedia. -- Jmabel | Talk 05:20, Apr 30, 2005 (UTC)

• Now, now, Marnen; you can do better than that. If you come away with "there seems to be a lack of reliability with Wikipedia" from my essay, then you're really missing some other major points. Wikipedia is Pickypedia; you're just proposing making it a little more so and adding a layer of metadata which is going to encounter problems of its own, because anyone who doesn't play the rules of your little game ("I read it and verified it.... SURE I did...") will soon overflow the game-players. My contention is that this fanatical adherence to anyone-can-undo-anyone-else's-crap and you-don't-even-need-an-account-to-edit, in the name of "free access to all information" does nothing but ensure that, over time (and I'm talking months, thanks), the whole thing will be polished down to a dull nub. Value systems clash constantly on Wikipedia all the time, and while it's fun to focus on edit/revert-wars as "oh, that's what Jason is talking about", I mean the iceberg-like drop into crap. Believe me, I've gotten mail and am in communication with people who are, right now, gaming the system into self-destruction, and they are doing a fantastic job. My advice is do what a few dozen people have already done: fork. --Jscott 07:43, 2 May 2005 (UTC)

• • Jason, I'm happy to have you participating in this discussion. I think you've misunderstood some of my points, however:
    • Pickypedia is not "my little game". If it becomes that, it will not be viable. I hope it will be a collaborative effort
    • I am specifically not adhering to "anyone-can-undo-anyone-else's-crap and you-don't-even-need-an-account-to-edit"; that nonadherence is kind of the whole point of the proposal. I am surprised that this was not apparent from the proposal itself.
    • My proposed Pickypedia structure is just that -- a proposal. If you think I'm just "adding another layer of metadata" and not fixing the basic problem, I am open to suggestions on how to avoid that error and where the basic problem lies.
    • Of course this would be a fork. What did you think it would be?
  • I'm curious to hear what you think is needed for something like this to work. --Marnen Laibow-Koser (talk) 13:05, 2 May 2005 (UTC)

• • Seems to me that Jason and Dvorak are close cousins in their abilities to stir up hornets' nests over a bunch of nothing. All I've encountered in my 1 year + of being an editor is increasing excellence. While I've only briefly reviewed this proposal, it sounds like a lot of useful thought went into it, and I applaud you Marnen for working on something constructive, rather than Jason and other folk trying to scream bloody murder over a few destructive individuals who really aren't doing the purported damage being reported. — Stevie is the man! ^{Talk | Work} 18:37, May 2, 2005 (UTC)

• • • I work on constructive projects, elsewhere, Buttercup. I voted with my feet long ago; I just happened to notice the links to my website and stopped by to see what was going on. Don't you fucking lump me in with Dvorak. --Jscott 19:47, 2 May 2005 (UTC)

• • • • With language that, I won't lump you with Dvorak. I'll instead say that you are worse. It's clear you are a destructive force. — Stevie is the man! ^{Talk | Work} 21:57, May 2, 2005 (UTC)

• • • • • So it's the use of profanity that defines someone as a "destructive force"? That's pretty fuckin' pathetic. Go show me where I'm being "destructive".--Jscott 22:04, 2 May 2005 (UTC)