I'm sitting down in front of a piece of code which was an idea that I
now don't know what to do or how to implement anymore. It was an image
notification bot in which would go through several of the densely populated
no source / no copyright categories and go off to notify the user.
Unfortunately, the upload logs are not consistant, and a database needs to be
rebuilt if I were not to query the information directly from the database.
- Jason Lee
On Tue, Oct 25, 2005 at 05:47:32PM -0400, Gregory Maxwell wrote:
On 10/25/05, Jason Y. Lee <jylee(a)cs.ucr.edu>
wrote:
I know this is likely a stupid question, and it
is likely a gross
violation of policy and accountability that I know of. And I'm already
guessing what the answer is, and why I bother to ask is currently beyond
me other than the reason that it's 5 in the morning where I am and I'm
likely just being stupid. For which this point of this email, and for
what I'm writing for has been a complete waste of time...
Anyway, would it be possible or stupidly possible to set up something,
and I'm not specifically saying the Toolserver, however, this is a passing
thought... where all the bots are just dumped into a fault tolerant server
and have access by several trusted programmers and people so that if a
useful and trustful bot goes awry that someone has immediate access to try
to fix it or otherwise kick the damn thing to stop it? This is of course,
an alternative from shooting at it by blocking the user account, or the IP
address for that matter.
What might be good is to amend the bot policy to require the bot
notice when it has new talk page messages, and if someone writes
"STOP" on the talk page that the bot pauses it's operation until the
stop is gone.
This would make it easier for any user to stop a broken bot without
having to track down an admin.
What might also be useful in the future is running certain kinds of
bots on the toolserver simply because there are a number of operations
that are much cheaper if you have access to a copy of the database....
For example, I am currently adding a category to all "Fair use images"
which are not inlined from the main namespace. (On en it's a CSD if
they've been tagged for 7days, because it's not fair use if we're not
using them!). It's about 2,200 images. If my bot didn't talk to my
local copy of the Wikipedia database it would have to load the 20,700
image pages in the 'fair use images' category just to find the 2,200
images I care about. The database query takes about 45 seconds on my
*laptop*, hot cache, or about 2 minutes if I just started mysql.
_______________________________________________
Toolserver-l mailing list
Toolserver-l(a)Wikipedia.org
http://mail.wikipedia.org/mailman/listinfo/toolserver-l