On 10/8/07, Aaron Schulz jschulz_4587@msn.com wrote:
I don't think that is a good idea. It removes much of the editing incentive of getting an account and makes editing more tiresome for those who edit the most (users).
I don't agree.
I think that instead we should improve the interface and operation so that it's not tiring for *anyone*. If there is a problem editing from the flagged version then it should be fixed, not just left to only impact people who will complain less.
One thing I've seen suggested is making anons always see the most recent revision on articles they have recently edited (a session length cookie could be used for tracking that). That idea could be extended to logged in users and also be applied to pages which the user has watchlisted. There should be no performance impact from these because editing sets a session cookie already, and page loads for logged in users already check watchlist status.
Doing this would do a lot to unify the view of Wikipedia for both logged in logged out users even when there was widespread use of flagging.
luca@soe.ucsc.edu wrote:
I think most of you know more than I do about the dynamics of user contributions to the Wikipedia, but I am seriously worried that showing stable revisions
I too am worried: I am worried that people with long term histories of categorical opposition to anything like stable versions might be playing too much of a role in our implementation choices and we may end up with a compromise system which combines all of the harms every proposal and none of the charms.
We had a prior system for revision quality markup, "mark as patrolled". The feature was considered a failure by many of our larger communities. Many people, myself included, believe it failed because marking a revision didn't accomplish anything useful so there was little incentive for anyone do anything with it.
If we do not use the ability to show the flagged version by default, I fear that the revision flagging will be little more than marked as patrolled and as much of a failure.
But at the end we don't need to worry about the worries: Instead we can simply use objective measurements. If we turn on users defaulting to the flagged revisions on ten thousand well distributed articles, we can then track the performance. We can measure the amount of editing before and after, we can automagically or manually measure the amount of vandalism, we can see how often readers are seeing a stale versions and how stale.
There is no reason to be afraid, we can use data to illuminate the dark.
The aggressive debating over the supposed risks of this feature are counter productive. None of us know what will happen because this is new ground. All any of us can do is guess. It would be really arrogant of us to think that we'll have it right at the first cut.
What we should be discussing is not how stable versions should be done but rather we should be discussing how to *study* stable versions.
My preference for a more aggressive implementation comes not because I think I have some magic understanding that proves everyone's worries wrong, instead it comes from two factors. (1) A more substantial change should produce results which are easier to measure. (2) I view the more aggressive implementation is closer to the ideal from the quality perspective and since we are trying to improve quality we should probably start testing from the ideal and back off until we remove the negative side effects.
So, how can we best study the results?