User:Rick Block/Keeping sewage out of the wine

Wikipedia's stated goal is to be an encyclopedia. Its format as a wiki has allowed it to grow and improve to the point where it is now the largest reference website on the internet. Some users think Wikipedia_is_failing, while others assert it is not. Failing or not, Wikipedia cannot currently be considered a reliable source. Indeed, Wikipedia's own policies prohibit using an openly editable wiki as a source. Jimmy Wales, Wikipedia's founder, has said there needs to be a focus on quality rather than growth. Schopenhauer's law of entropy is ''If you put a spoonful of wine in a barrel full of sewage, you get sewage. If you put a spoonful of sewage in a barrel full of wine, you get sewage.'' Wikipedia currently has over a million barrels in which anyone with a web browser can put as much sewage as they'd like. In the opinion of this essayist, one essential ingredient to further the quality of Wikipedia is to implement a mechanism restricting the ability to add sewage to our wine.

Wikipedia currently has three types of editing restrictions: Since there is no restriction on the ability to create accounts, the only one of these that can possibly keep sewage out of our wine is full protection. Various proposals have been made to protect articles that reach a certain state of "completeness", for example Stable versions and Static version. Rumors of work toward some mechanism possibly related to one of these have been floating around for more than a year. None of these adequately address the fundamental issue of scaling to a project the size of Wikipedia.
 * 1) Article creation is restricted to only logged in editors
 * 2) Moving (renaming) articles is restricted to logged in editors whose account has sufficiently aged
 * 3) Editing of articles can be restricted to logged in editors (semi-protection) or administrators (full protection).

With the current implementation of MediaWiki, the act of editing an article and the act of making an edit visible to others is not separable. Edits don't "pend", they're immediately "live". This means to restrict the ability to add sewage to our wine we must restrict the ability to edit. Restricting this ability hampers the very factor which has led to Wikipedia's phenomenal growth, i.e. the ability of large numbers of anonymous editors to contribute. Protection is currently used only as a last resort in cases where articles are being frequently vandalized, typically not as a mechanism to control what edits are made to articles but as a mechanism to stop any changes from being made. A new mechanism is needed, specifically a mechanism not to stop edits but to control them. The proposal to require approval for anonymous edits would apparently be a step in this direction, requiring edits from anonymous users to be approved by a logged in user before the edit is displayed in the "live" version of the article. The inherent problem with this implementation is that anyone is able to open an account, so editing has not really been restricted in any meaningful way. In particular, the distinction between an anonymous user and a logged in user has nothing to do with how trustworthy that user's edits may be. For example on the English Wikipedia, one of the most prolific vandal fighters is an anonymous user, User:68.39.174.238.

The solution is to separate the permission to edit from the permission to approve edits, per the anonymous edit approval proposal, and also to introduce a new class of user between logged in user and administrator able to approve edits. Articles that reach a certain level of completeness would be marked as "requiring approval" which would do two things. What this does is close the lid on the barrel, keeping random people from adding whatever they'd like. There would still be a container outside the barrel in which anyone could place whatever they'd like added, but actually adding this could only be done by someone a little more trustworthy than anyone with a web browser. With over a million barrels, requiring administrators to approve additions does not scale. Most people are trustworthy, so what is needed is a lightweight mechanism to allow administrators to mark (and unmark) accounts as edit approvers, with the idea that most active editors would be marked as approvers. Approvers, and we'd need lots of them, would be tasked with making sure what is being added to the wine is not sewage but something that might plausibly improve it.
 * 1) The version displayed to the world would be the last approved version.
 * 2) Edits by anyone without "approval permission" would pend.

Separating the act of editing from the act of approving edits would allow levels of approval to be defined as well. Just as today it is possible to allow anyone, or only logged in users, or only administrators to edit an article, an article marked as requiring approval could be marked as requiring an approval from a certain level of approver. For example, there could be two levels, one easily obtainable for controlling changes to a large number of articles and a higher level not so easily obtainable for controlling changes to featured articles.