[Foundation-l] Expert board members - a suggestion

Anthony wikimail at inbox.org
Wed Sep 16 14:24:30 UTC 2009


On Wed, Sep 16, 2009 at 9:51 AM, Chad <innocentkiller at gmail.com> wrote:

> On Wed, Sep 16, 2009 at 8:04 AM, Anthony <wikimail at inbox.org> wrote:
> > On Wed, Sep 16, 2009 at 4:46 AM, Samuel Klein <meta.sj at gmail.com> wrote:
> >>
> >> Why are we revisiting something from 2007-08 financial planning two
> >> years after it happened and 15 months after the final report?
> >>
> >
> > Because there isn't enough data on the mistakes that are being made
> today.
> >  15 months ago, there was "oh, the dumps will be fixed real soon now" and
> > "that money which wasn't spent will be spent in 2008-09".  But today we
> know
> > "no, 15 months later they still aren't fixed" and "no, that money will
> get
> > rolled into the general budget where it won't even be spent".
> >
>
> Have you not read people's replies to this? All dumps are working except
> for the full enwiki history. That's certainly a lot better than
> before, when pretty
> much every dump was failing. They've got all but one dump for one wiki
> working, and that's still being worked on too. What more would you ask?
>

I don't think you've been following the dump situation for the past three
years.  Maybe there are others on this list who also don't understand the
situation.  During the last three years or so, all dumps have been produced,
albeit intermittently, except the full history en.wikipedia.  In order to
"get all but one dump for one wiki working", the "solution" was primarily to
1) throw more hardware at the problem and 2) stop even trying to produce the
en.wikipedia full history dump.

What more I would ask for is to fix the actual problem.  That means
redesigning the dump system, which was not designed for such large amounts
of data, and needed to be rewritten three years ago (when the WMF plan was
to simply throw more hardware at the situation, which they didn't even do).
One or more people are apparently working on this.  I haven't seen any
redesign plans or progress reports though, so I have my doubts, not that one
or more people aren't actually working on this, but as to whether or not
it's going to get done.

Maybe if we could get a report on the status of the redesign, the plans for
the redesign, etc., at least those doubts might be allayed, and this would
become an example of a past mistake.  But it still would be worth talking
about.


More information about the foundation-l mailing list