On 12/31/06, Rob Church <robchur(a)gmail.com> wrote:
On 30/12/06, Steve Bennett <stevagewp(a)gmail.com>
wrote:
This user refuses to do the cleanup. Hope this
helps.
Superb attitude.
With Wikipedia, and wikis in general, one of the basic principles is
that if some improvement can be broken down into several different
independent improvement work units, then it's fine to carry out one of
those units and not the others. There's no shame in adding {{stub}} to
an article, and leaving the job of recategorising it
{{brazil-music-stub}} to someone else. I improved a page by renaming
it. Someone else - preferably a bot - can do the mind-numbingly
tedious job of updating double redirects.
The problem we'd face if, for instance, 200 or
even 2000 redirects
needed to be corrected in an autonomous, atomic fashion, is that the
software has to do the same as the user - open the page for editing,
make the changes without disrupting other content on the page (e.g.
categories, or cute little templates), save the page, handle edit
conflicts, rinse and repeat.
How ever long it takes the server to do it, it takes a lot longer for
the user, and pisses them off a hell of a lot more. That's why we even
have computers, remember. Your paragraph might serve as an argument
against renaming heavily redirected-to pages. It might serve as an
argument for some improvement to MediaWiki. But how does it argue
against getting a bot to update double redirects instead of a human?
It might be much faster on the server side, but
it's still a
time-consuming operation, which could well time out for large sets -
leaving us with a worse mess to clean up. Then there's the problem
that it might be desirable to have some of the redirects left intact,
etc.
There are times when we want to leave double-redirects left intact? So
that we're on the same page, we have redirects:
A->B
Page B gets renamed to C, creating a redirect at B:
A->B->C
Some page x that used to link to A can still link to A, even after the
double-redirect is fixed:
x->A->C
Why would you want "some of the redirects left intact", ie, A still
pointing to B?
I'm not against some kind of automated assistance
for such operations,
but it will have to work in a robust fashion, taking into
consideration that we don't want to make the situation worse.
Worse than what? A bunch of double redirects?
Steve