Improving the Wikipedia Model - Reed Meyer

On Improving the Wikipedia Model

By expanding the bounds of collaborative content to enable practically anyone with an internet connection to make contributions or editorial changes, Wikipedia has cemented its place in the modern Web, the so-called “Web 2.0,” and deservedly so. My only substantial complaint with Wikipedia—and I imagine this is shared by many—is that, by being so permissive in allowing just about anyone off the street to make changes, they make themselves too readily susceptible to spam and vandalism. If I were setting up my own collaborative-content encyclopedia, I would make a couple of policy changes, which, although simple in definition, are hopefully sufficient to avoid much of the shortcomings of Wikipedia.

1) For articles in a given field, I would only allow experts in that field to collaborate on such articles. Experts would be given user accounts by invitation only; they cannot ask for such accounts. Experts would be found by culling the existing literature, offline and online, on the subject. For example, for the field of astronomy, invitations could be sent to every author of a research paper published in the last year or two. For sports, writers for every major sports media outlet, and contributors to the more established amateur sports websites, could be contacted.

2) Someone would have to be appointed to oversee the invitation process in the above step, for each subject. After the initial round of invitees, it is possible that the subject contributors themselves could periodically vote on a new subject administrator. This process would likely require little policing, but the site administrator could intervene if it does not run smoothly.

3) Whenever a user makes an edit, including the creation of a new document, the document is flagged to be reviewed by another user who also covers that field. The edit does not go live until it has been approved by the reviewer, whose identity is unknown to the first user. If the reviewer does not read it in a timely fashion, it is submitted anew, to another reviewer. If the revision is denied, the first user must either abandon the edit or resubmit the edit to the same reviewer, presumably with changes to address the reviewer's concerns. The writer can appeal to the subject administrator in extreme cases. This procedure is meant to mimic the peer review process which is the hallmark of prestigious research journals.

Notes:

1) Because user accounts are awarded by invitation only, this in principle eliminates circumvention tricks which rely on people having access to multiple email accounts and the like.

2) Because accounts are awarded by invitation, and therefore the true identity of each user is known, this identity is published with each article and is available in the article revision history. Experts in the field presumably have a stake in maintaining a good reputation, so this plus the review process should be sufficient to prevent most frivolous edits.

3) Given the potentially long interval to wait for approval, it'd be nice to come up with an article locking mechanism to alleviate merging issues when multiple writers are making changes.

Articles by Reed Meyer

All the way up to the Meyer Home Page