Erik wrote:
>Done, but I'd prefer not to flush the page cache for
>performance reasons.
Fair enough, and thanks (I knew asking for developer access would scare one of
you guys into action ;).
-- Daniel Mayer (aka mav)
There seems to be general agreement to do this. Can somebody with developer
access either do it give me access to do it?
I'll figure it out.
-- Daniel Mayer (aka mav)
The dumps have been fixed, but Brion is obviously too busy fixing other
problems to rerun the stats (or spending quality time with his fiancee,
to make yet another unfounded accusation).
Also, the en: stats for November are missing since two weeks (I expect a
runtime error, or lack of memory, who knows, it puts quite some load on
the machine).
Without feedback from Brion who runs the job on his PC there is not much
I can do.
At least the log files will be uploaded from now on, so that I have some
clues where to look for.
Erik Zachte
I've been talking to Tim on the IRC channel and there is some desire to
get things moving on the dev branch.
He's got pretty much all the fixes in stable merged into dev, so what
I'd like to see is for us to commit to putting the dev branch onto the
live main servers in about a week. I'm going to say November 23 sounds
good... (Nov 22 is the SoCal Linux Expo so I'll be busy that day, but
free the next.)
In the next week we should make sure that everything in the dev branch
either _works_ and is ready for prime time, or is cleanly disabled
pending further work. Once we put it on the main server it'll get more
exposure and we'll find more bugs. :)
After a few more days of fixes we can branch it off as a new stable
point release and swear never to touch it again except for emergency
bug fixes. For real this time!
I'd like us to then commit to a fairly regular release schedule;
perhaps every 3 or 4 weeks we make sure that anything in the dev branch
is either cleanly working or can be cleanly taken out.
It'd also be great if we had more automated testing tools. Lee wrote
this whole big Java-based client-side test suite that takes up a huge
amount of space and, as far as I can tell, no one has ever used it or
attempted to maintain it. Heck, I couldn't even figure out how to
compile it... I don't know if it'd be worth someone trying to update
that, or if we should just toss it and concentrate on unit tests in PHP
that work with the actual code and test individual components in
isolation.
-- brion vibber (brion @ pobox.com)
Christopher Mahan wrote:
>I agree with Fred. I think it's important for people
>to realize right away that they can help the project
>financially. It also answers the question I get when
>I show the W to aquaintances: "Where's the catch,
>they can't be doing this for free."
That is why there is a donations link in the sidebar.
It is prominent yet it isn't obtrusive (the 'support'
message is under the article title and is thus part of
the article - that's obtrusive). We can afford to be
obtrusive only when we are having an official
fundraising drive. Otherwise we, at worst, annoy
people and, at best they start to ignore the message
(try saying the same word over and over and over again
- after a while you can't remember what the word
means).
Fundraising drives are supposed to be a bit annoying
and attention getting, but that's fine because they
are only held a handful of times a year. Last time I
checked we were not in such dire straights to require
a constant fundraising drive. Do you want us to give
the impression that we are desperate for cash?
Another way people can get to the donation page is
through a link to the foundation's homepage under the
logo of every project. It would say "A [Wikimedia]
project". [Wikimedia] would be a link to
http://wikimediafoundation.org (which in turn has
prominent statement asking people to consider donating
and a "Donate Now" link in the sidebar). Then there
would be an indirect link to the donation page at the
top of every sidebar and a direct one at the bottom of
every sidebar.
-- Daniel Mayer (aka mav)
__________________________________
Do you Yahoo!?
Protect your identity with Yahoo! Mail AddressGuard
http://antispam.yahoo.com/whatsnewfree
Anybody given thought to making a xml-rpc interface to the W?
=====
Christopher Mahan
chris_mahan(a)yahoo.com
818.943.1850 cell
http://www.christophermahan.com/
__________________________________
Do you Yahoo!?
Protect your identity with Yahoo! Mail AddressGuard
http://antispam.yahoo.com/whatsnewfree
This message:
Find out how you can help support Wikipedia's phenomenal growth.
should only be displayed when we are having a fundraising drive. Since we
received the money for the new database server before such a drive even
started, and since we are going to delay distribution of our next press
release until we hit 500,000, I see no need to keep this message at this
time.
A couple weeks before we distribute the next press release (sometime early
next year) the message can be re-enabled in order to give the page cache time
to update. Keeping the message up all the time will result in people tuning
it out and not paying attention to it when we having a fundraising drive.
But having it gone and then brought back will get people's attention.
-- Daniel Mayer (aka mav)
Jimbo wrote:
>This is really not a wikitech-l type of question, is it?
Well removing it is a technical issue... I didn't think its removal would be
contentious. Silly me, /everything/ is potentially contentious around here.
>I think this, or some variant of it, makes the most sense.
>If someone wants to donate, we shouldn't make it difficult
>for them to find, but it probably is better in the long run if
>we reserve that top slot for times when we really really
>need to shake the money tree. :-)
That is why there has been, and always will be, a "Donations" link in every
sidebar and also under "About the Project" on at least the English Main Page
(not to mention the link in the contributing FAQ). That is more than enough
for people who are spontaneously motivated to donate; a minimal amount of
looking will result in them finding a link to the Wikimedia fundraising page.
But hitting everybody over the head with it everytime they view a page would
tend, IMO, to dull the blow and result in less overall money generated than
if we only showed it during fundraising drives. It is also a bit annoying and
gives the impression that we are desperate for cash (thus we should be
judicious in such a prominant use).
-- Daniel Mayer (aka mav)
-----Original Message-----
From: Jimmy Wales [mailto:jwales@bomis.com]
>This is really not a wikitech-l type of question, is it? I don't
>personally have a firm opinion about it.
>1. On the one hand, we're always having a fundraising drive in a
>sense. We have enough funds right now, or nearly so, to buy a 4th
>machine to either be used exclusively as our mail server, or to be
>used exclusively as our load balancer, or to be used as our 3rd
>webserver. Even after the new db server is installed, all three of
>those will be legitimate needs. (We might combine some of those
>functions at first, of course. But there are good arguments for
>separating them all as soon as we can afford it.)
>
>2. On the other hand, Mav's argument is (perhaps) that we can
>maximize revenue by not having the request up there all the time,
>and I think there's something to that.
>
>--Jimbo
Re 2) Perhaps experimentation is in order. I think it is hard to tell
from here which method i.e. "continuous" or "periodic" style of
requesting will generate the most.
Re 1) The load-balancing recently introduced seems to have created an
almost miraculous speed compared with before. Obviously the monster
machine is going to have an effect as well. We should expect the upward
curve on bandwidth use to continue for a good while yet - regardless of
the purchase of a fourth machine and its purpose. Is cost of bandwidth
likely to become a great issue than now, where we are extremely grateful
to one person that it hardly seems to be an issue at all. It would be
terrible if we had lots of machines at some point in the medium-term but
not able to utilise them due to bandwidth costs.
Pete, [[User:Pcb21]],
Wikarma: [[Wikipedia:WikiProject Cetaceans]] over half of all whale
species now have an article!