[WikiEN-l] Featured article deterioration

[maru dubshinki]

On 4/12/06, Fastfission <fastfission at gmail.com> wrote:
> I've wondered what the best way is to deal with this sort of thing.
> One might be to make FA status awarded to specific versions of the
> article in a more direct way -- i.e. the FA template on the talk page
> could link directly to the version which was voted to be featured, to
> make comparisons with the current article easier.
>
> Another is a better way of detecting bad edits. I've seen a lot of
> really bad content sneak into high-visibility and highly-watched
> articles because of the following pattern:
>
> 1. Editor X vandalizes an article once
> 2. Editor Y vandalizes the article three times
> 3. Admin Z reverts Editor Y, does not look at the edit of Editor X
>
> This is especially prevalent when there is a small spate of vandalism.
> I try to make sure I always check the current version against
> something further down the page,  but it's hard to remember to do in
> all cases. Perhaps if it was easier (i.e., with a single click and
> without any looking at the history) to see the diff for the edits done
> by the most recent five users (sorting it by user rather than edit
> would make it easier to avoid skipping single bad edits)? Maybe if
> there was a regularized system of evaluating many edits at a time and
> screening them for problematic content? Maybe we need to find ways to
> get around seeing histories as collections of single edits and rather
> as clusters of edits? I don't exactly know. I'm just speculating
> wildly here and not proposing anything concrete.
>
> I don't think anyone thinks the wiki system is flawless in any
> respect, or that infinite edits means an infinitely good article.
> There are benefits and detriments, and a good working-philosophy is to
> try and focus on ways in which can -- utilizing social, technological,
> and other approaches -- try and boost the strengths of the positive
> aspects and get around the negative ones. It's no more destined to
> succeed than it is doomed to fail.
>
> FF

That would work well with patrolled edits- one could then check back
to the last patrolled edit, and then if all the subsequent changes
were good, mark them all as patrolled.
Of course, we'd need patrolled edits in the first place.

~maru