Neil Harris wrote:
>How about the much simpler approach of restricting page creation to
>logged-in users?
>This would act as a minimal speedbump which is probably just enough to
>discourage many (most?) drive-by junk page creators, whilst still
>allowing anons to edit any existing page as freely as before.
>This could be applied only to the main article space, thus allowing
>anons to create talk pages as before.
>This would be similar to the current restriction of image uploads to
>logged-in users.
Sounds plausible. What are the numbers on the problem? What percentage
of anon page creations are shoot-on-sight?
(Adding a new restriction is, of course, NOT something to be done lightly.)
- d.
Martin Richards wrote:
>I have been looking at this for a while now, something that is clear is that
>virtually all (as in ~95%) of the deleted articles (especially speedies) are
>from anonymous users, and a lot of which aren't tests, but plain old
>vandalism.
I think that's unduly negative; it's more "sandboxing". They see "Edit
this page" and can't quite believe it.
> I really think we need some kind of "votes for creation" or
>"approved creation" or something, maybe not soon as there are still vast
>quantities of topics we don't cover (see [[Wikipedia:Wikiproject Missing
>encyclopedic articles]] for ~150,000 examples), but maybe a in year or so
>from now I dont think we will have much choice.
*shudder* I doubt I'll be bothering to create many articles under such a scheme.
- d.
Alphax wrote:
>So, a website which sells widgets and is unknown outside of the widget
>community should not have an article
I must ask: why on earth not, if the info is third-party verifiable?
How is one to research an area one is new to without cues to such
things? If you're interested in an area, why *wouldn't* you be
interested in things of interest within that area?
I see no reason to deliberately reduce our possible usefulness.
The issue of webcomic articles mostly being pretty crappy is another
matter, and an important one. But your proposal seems an
overgeneralisation of a way to use deleting the whole article as a
tool to solve *editorial* problems. The problem there being that many
people think an article being deleted means another article on that
topic can never be created, ever. (Some even think an article being
deleted means its content shouldn't be allowed in new articles
anywhere else on the encyclopedia, which I can't make sense of.)
- d.
In my experience, this bug manifests itself only when one has multiple
tabs or windows open.
Then, it is necessary to press whatever button initiates a post action
on two or more windows or tabs, without waiting for the prior post to
complete. This is only possible when the site is responding slowly. I
don't believe I've had the initial problem occur with any pages other
than the delete confirmation page, though I could be mistaken.
Even then, the action will work much of the time, perhaps because the
posts happened to be processed in the proper sequence.
Once the first error occurs, other forms will not post properly unless
they are reloaded first.
Since the cookies are shared across browser windows, with multiple
apaches being round robinned by the squids, there is no way to
guarantee that the posts will arrive in the order in which they were
sent, and there is no way to guarantee that an updated session cookie
takes effect prior to the subsequent post. Therefore, any attempt at
session control is bound to be a fruitiful source of bugs which cannot
then be reproduced in a typical test environment where there is only
one apache and no lag.
My MW test bench is now badly out of date but I suppose I could get the
latest from CVS and look for it if there is interest.
uc