[Wikipedia-l] Comments on Axel's beta/stable ideas

Sat May 25 08:42:13 UTC 2002

On Thursday 23 May 2002 12:01 pm, Axel wrote:
> That said, here's my minimalistic suggestion: everything works exactly
> as it does now, except that every page gets two additional links:
> "View last reviewed version of this article" and "I have reviewed this
> version of the article and I think it is ok". The history of every
> article would record who has reviewed which version of the article and
> when.
>
> The set of all "last reviewed versions" could then be seen as the
> "stable" Wikipedia and could be pressed on CD. This would at least
> guard against vandalism, stupid jokes and blatant propaganda and
> advertising that sometimes gets through.
>

Cool! What a fantastic idea!. I also like the idea of having two versions of 
an article; but I would simply call them Reviewed and Development. The 
reviewed one would be a static page that could be replaced with the 
Development (editable) version of an article whenever a certain number of 
reviewers give it the OK. It also would be most excellent to have an even 
higher level of review called "Peer Review" that would be performed by 
somebody with a related college degree. 

This level of review would be roughly analogous to what Nupedia had (has?) 
set-up but should done in a better way. But then the devil is in the details 
here and we may need to temporarily freeze an article while it is being peer 
reviewed so the expert can fix any glitches, submit the article as peer 
reviewed and reopen the development version back to the masses (the 
reviewer(s) should only have a very limited amount of time to review/fix the 
article while the development version is locked). This might be a messy thing 
to do in practice though and we should discuss this at length in order to 
work out the details if we decide to do this at all.

So potentially we might have three versions of an article; one which is world 
editable, one which was voted to be OK by users, and one that was tweaked and 
made to conform to higher Nupedia-like standards. 

> The only issue is: who is allowed to review articles? The pragmatic
> answer would be: all sysops. 

Somewhat disagree: I tend to agree with Chuck's previous comment here in that 
I think we should open this up to anybody who has been a user for over 3 
months. I would add though that exceptions should be made for anybody who has 
edited a reasonable set number of articles and who has the backing of at 
least one sysop. We could update the software to give these user's special 
additional rights automatically after three months and/or after they 
edit more than a set number of articles (thus making them then eligible for 
sysop promotion). We could then call these users "trusted hands" (which is 
already in the wikiware code -- but a "trusted hand" doesn't have any more 
rights than a regular user does in the current set up -- correct me if I am 
wrong Brion).

It would also be nice for a sysop to have the ability to "reset the clock" as 
Chuck proposed for anybody who doesn't follow established policy after being 
warned. The warning process could be started by a posting a message on a 
special page by at least one person with "trusted hand" or greater 
status. These warnings shouldn't do anything other than inform the user that 
"they have been warned" with a reason why (there could be an automatic 
temporary suspension of "trusted hand" status by being warned -- but I'm not 
sure if I like that idea). These warnings could also automatically expire 
after a set amount of time and/or number of edits unless somebody else renews 
the warning. In addition, a "trusted hand" or sysop should be able to remove 
the warnings if a mistake had been made. If the user doesn't get his or her 
act together during the warning period, a sysop could then reset the user's 
clock (which should have to be a different person than the one 
starting the warning process).

Blatant trolling or VANDALISM could also be logged on these warning pages. 
The warner could choose from a checklist of options when warning the warnee. 
Such a list might include these options: Introduced bias, Rambling 
contributions, Falsification of facts, Nonsensical/inappropriate 
contributions, Introduction of propaganda, Sloppiness, Violation of naming 
conventions or VANDALISM. If VANDALISM is selected then in addition to the 
warning page a VANDALISM in PROGRESS page can also be displayed on 
RecentChanges with the "contributions of" the warnee automatically placed on 
that page. 

There should, however, be some safeguards to help ensure that honest mistakes 
by contributors are not used as the basis to warn a person (perhaps have a 
silent first warning that only the warner and warnee know of... but then how 
would you make that work?) . I'm concerned that such a system might be used 
too much and contributors might think twice about hitting the save button. If 
misused, such a system could have negative effects on the number and extent 
of contributions wikipedia receives. So, if something like this is deemed 
necessary, then great care must be taken to minimize possible negative 
side-effects. But something similar to this will probably be needed in the 
coming years as the number of active contributors becomes larger than anyone 
(including users like me who contribute at least 3-4 hours a day on average) 
is able to keep track of.    

> A code of honor is probably in order,
> saying that no sysop should review an article that they themselves
> substantially contributed to.

In principle this might be a good idea but in practice I don't think this 
would work with the size of our current contributor-base (not to mention 
sysop-base).  As it is right now in many cases the people with the greatest 
knowledge and ability to review an article have already significantly 
contributed to it. 

The "honor system" set-up may work after a few more years when wikipedia has 
literally thousands of people contributing daily and has close to 100,000 
articles (which is an arbitrary goal BTW -- I doubt we will all stop 
submitting new articles when the project is "complete"). Then we could 
reasonably expect that there would be enough "trusted hands" with sufficient 
knowledge of the subject and who had not already contributed much to 
the article to be able to vote it up or down.

Given all that -- I do have reservations that the power to warn and reset the 
clocks of users may be abused and this could change the character of 
wikipedia and discourage contributions. However, given that IP banning hasn't 
<yet> been abused by any sysops I have hope that such a system might just 
work.

If we could somehow ensure that we have reasonably competent reviewers then 
Wikipedia could then (eventually) become what Nupedia was not able to be: an 
extensive, trusted and useful source of human knowledge. We are almost at the 
extensive stage -- Should we work on a framework for establishing reasonable 
trust and therefore usefullness of our articles? I vote for yes.

Just my 1.5 cents 

maveric149  