[WikiEN-l] So what does Flagged Revs feel like?

[Gregory Maxwell]

On Sat, Sep 26, 2009 at 6:20 AM, David Gerard <dgerard at gmail.com> wrote:
>
> If you want to know how Flagged Revisions feels from an unprivileged
> position, go to Wikinews and fix typos. I just did this on
>
> http://en.wikinews.org/wiki/Geelong_win_2009_Australian_Football_League_Grand_Final
> - check the history. I'm not an admin or reviewer on en:wn.
>
> What did it feel like? Curiously unsatisfying. The fix not going live
> immediately left me wondering just when it would - five minutes/? An
> hour? A day? It felt nothing like editing a wiki - it felt like I'd
> submitted a form to a completely opaque bureaucracy for review at
> their leisure.


UI fail.

There is no reason for you to know or care that your edit isn't being
displayed to the general public.  It's being displayed to you, it's
being displayed to all the other editors, it's being displayed to
anons who click a link to see the latest.

It's our own damn fault for making the UI say the equivalent of "NOW
YOU MUST WAIT WHILE OUR TRIBE OF ELDERS SCRUTINIZES YOUR PATHETIC
EDIT" …  we don't have to do it this way, and we shouldn't do it this
way.

The process can and should be made mostly invisible to casual editors.


On Tue, Sep 29, 2009 at 7:15 AM, Surreptitiousness
<surreptitious.wikipedian at googlemail.com> wrote:
> edits, and they don't get checked in greater detail.  Looking at it,
> it's entirely plausible we're going to have people from all over the
> world examining edits outside their context. That's going to mean things
> will get missed, isn't it?  Not saying it isn't any better than the
> current model, but at least with the current model someone will not
> assume something is good since they will know it hasn't been checked.

The way I see it — What this is about is two fold:
Right now an edit to an article can often go hours before someone
experienced with editing takes a look at it. During that time the
completely unscrutinized edit is displayed to the world. The flagging
changes the failure mode: We display an older edition when review gets
missed.

And the tool also provides a way to decrease the time that a review is
missed by providing visible tracking of the review status. The tool
also avoids wasted effort on redundant reviewing (i.e. how many people
look at every good edit to George W. Bush just to reach the redundant
determination that it doesn't need to be reverted).

Your concern that the review-collaboration may prevent changes from
being noticed by qualified contributors but this has been addressed in
multiple ways:  There are multiple review levels which can be applied,
i.e. the "checked for obvious tripe" vs "blessed by geniuses" and we
haven't seen complaints of that with things like new page patrol
(which is under-utilized, but used).

I think what we'll find is that subject matter expert contributors who
are watching articles are going to want to look at all the changes
changes regardless of the flagging status, since they'll have concerns
about presentation and style which go even beyond simple accuracy
concerns.  It is easier to do this kind of reviewing as a subject
matter expert with flagged revisions because you can limit yourself to
diffing between the flagged versions, allowing you to skip
intermediate bad edits which have be resolved by recent change
patrollers.

This is another area where the UI can have a real impact: It's
important the it not overstate the level of review that is occurring.
Right now  flaggedrevs.labs.wikimedia.org is calling the levels
"Draft" "Checked" and "quality", but this is under active discussion.