on 9/1/03 12:56 AM, David Levinson at dlevinson@mn.rr.com wrote:

Proposal for a Wikipedia Article "Approval" Process

Assuming there is support for a stable version of wikipedia, some mechanism for stability is required.

Allow logged in users to approve any version of an article. Logged in users could also disapprove any or every version of an article. Articles would be scored based on number of users who approved (weighted?) - number of users who disapproved (weighted?). The article with the highest approval rating would be the released (approved) article. This would be displayed somewhere distinct from the working wikipedia (e.g. sifter.wikipedia.org, or something similar).

New users would be presented with Wikipedia Release edition. They could still edit the article, but it would not be released until the approvals on the new article exceed the approvals on the previously released article. The under-edit version from which users are working may differ from the release version if approvals have not yet been found. The differences would be highlighted on the edit screen.

Edits would of course appear in Recent Changes. In addition, there would be Recent Approvals tracking articles which recently were voted on for approval.

An edit would of course count as a vote for approval (if preferences so set).

Under the article (in view mode) (for logged in users) would be "vote to approve" and "vote to disapprove" buttons. A vote to approve would transfer a users vote to this version for approval and remove it from previous versions. A vote to disapprove would remain even if another version was disapproved.

Possible weight = 1 if under 100 edits, 2 if under 200 edits, 3 if under 300 edits, 4 otherwise. Alternative weighting schemes are possible as well, including weight=1 for all logged in users.

Comments?

-- DavidLevinson

This is a recipe for Pablum. Often the better article is supported by only a few people, or only one person. But feel free to do something really dumb. After all, I'm the competition.

Fred Bauder

http://www.internet-encyclopedia.org

David Levinson wrote:

Allow logged in users to approve any version of an article. Logged in users could also disapprove any or every version of an article. Articles would be scored based on number of users who approved (weighted?) - number of users who disapproved (weighted?). The article with the highest approval rating would be the released (approved) article. This would be displayed somewhere distinct from the working wikipedia (e.g. sifter.wikipedia.org, or something similar).

Great, but IMHO we don't need a counting/voting. After all, articles on wikipedia are not three-versions-in-parallel. There's one article, and lots of old versions. I say: Any logged-in user can approve any article with a single click, which is then copied to a sifter site. Software's already in place (more or less, and at the wrong place, but...)

New users would be presented with Wikipedia Release edition. They could still edit the article, but it would not be released until the approvals on the new article exceed the approvals on the previously released article. The under-edit version from which users are working may differ from the release version if approvals have not yet been found. The differences would be highlighted on the edit screen.

I'd prefer having the Wikipedia Release edition (I like that!) with a link "view current revision" to wikipedia; *there*, you can edit if you want, or just read...

Edits would of course appear in Recent Changes. In addition, there would be Recent Approvals tracking articles which recently were voted on for approval.

I'm afraid that would end in a how-many-voted-and-edited-what mess.

An edit would of course count as a vote for approval (if preferences so set).

Under the article (in view mode) (for logged in users) would be "vote to approve" and "vote to disapprove" buttons. A vote to approve would transfer a users vote to this version for approval and remove it from previous versions. A vote to disapprove would remain even if another version was disapproved.

"Approve" should do just fine. Not approving means disapproving or not caring. Approvals will show on the sifter RecentChanges only. That won't clutter the wikipedia RC, and keep the sifter RC clean. Also, no more code to write ;-)

Possible weight = 1 if under 100 edits, 2 if under 200 edits, 3 if under 300 edits, 4 otherwise. Alternative weighting schemes are possible as well, including weight=1 for all logged in users.

What happened to "all wikipedians are equal"? ;-)

Magnus

David Levinson wrote:

Allow logged in users to approve any version of an article. Logged in users could also disapprove any or every version of an article. Articles would be scored based on number of users who approved (weighted?) - number of users who disapproved (weighted?). The article with the highest approval rating would be the released (approved) article. This would be displayed somewhere distinct from the working wikipedia (e.g. sifter.wikipedia.org, or something similar).

New users would be presented with Wikipedia Release edition. They could still edit the article, but it would not be released until the approvals on the new article exceed the approvals on the previously released article. The under-edit version from which users are working may differ from the release version if approvals have not yet been found. The differences would be highlighted on the edit screen.

Under the article (in view mode) (for logged in users) would be "vote to approve" and "vote to disapprove" buttons. A vote to approve would transfer a users vote to this version for approval and remove it from previous versions. A vote to disapprove would remain even if another version was disapproved.

Comments?

The problem with this is that democracy is often at odds with NPOV and truth. The proposal would too easily lead to a "Tyranny of the Majority."

Ec