[Foundation-l] Stable versions live on de.wp
Magnus Manske
magnusmanske at googlemail.com
Thu May 8 12:46:10 UTC 2008
On Thu, May 8, 2008 at 11:15 AM, Andrew Gray <shimgray at gmail.com> wrote:
> 2008/5/8 Magnus Manske <magnusmanske at googlemail.com>:
>
>
> > I think Lars calls FlaggedRevs "soft locking" and doesn't want to
> > apply it on a large scale.
> >
> > IMHO FlaggedRevs are much different from protection/locking. The
> > purpose of FlaggedRevs is to present the uninitiated audience with a
> > vandalism-free Wikipedia. Applying FlaggedRevs to only a few pages
> > will not achieve that.
>
> I think Lars' worry is (partly) a matter of update rates leading to
> locking-through-inertia.
>
> Yes, we can purge through all existing pages and set a flag on them -
> indeed, it's an excellent opportunity to do so, and it means that
> hopefully all pages will get eyeballed to make them clean.
>
> But! What happens next? If I go off and update a short article (one
> which no-one has watchlisted, etc) on de.wp, what mechanism is in
> place to flag *that* revision? Is it possible that on non-high-traffic
> pages, an "old" revision could remain the newest flagged one for
> weeks, months, despite having been superceded?
>
> I can see how this would have the effect of soft-locking - I'd really
> like to know how we plan to get round it.
A flagged page that is edited by a "trusted" user (that is, one who
can flag pages) is flagged by default. Given the (potentially) huge
number of such users, many pages will have flags updated as a
by-product of normal editing.
> Automatically generated reports of all pages with the most recent edit
> "unflagged", sorted by age, perhaps? This'd allow the permitted users
> to knock off a few each a day, keep it churning over. The sort of
> thing that would appeal to inveterate RC patrollers ;-)
There are a number of special pages for these purposes. There's one
that lists pages with no flagged version; you can even filter by
category!
I think I saw a "sort-by-oldest-flag" page as well.
Of course, the potential for views on this would increase
exponentially once such info is available through the API. Maybe I
should write a toolserver interface until then. Are the respective
tables/fields replicated to the toolserver?
Magnus
More information about the foundation-l
mailing list