On Mon, Oct 25, 2010 at 4:24 PM, Aryeh Gregor Simetrical+wikilist@gmail.com wrote:
On Mon, Oct 25, 2010 at 7:15 PM, George Herbert george.herbert@gmail.com wrote:
(suppressing grumbles about diff engine for edit conflicts, probably an algorithm rather than speed issue)
Diffs are fast enough if you use wikidiff2, no?
Yeah; what I really want is a diff algorithm that properly deconflicts more current edit conflict problems. It's not speed, it's smarts in the algorithm.
I suspect that the MySQL engine is the one place where parallelism is most applicable, but I am not a MySQL internals guru. However, there is one associated with the project, and I believe on the list. Domas?
Yes, parallelism is very applicable to databases, which is why they've already had tons of effort invested by people who know a lot more about the topic than me, so I probably can't do much there.
To clarify, the subject needs to 1) be reasonably doable in a short timeframe, 2) not build on top of something that's already too optimized. It should probably either be a new project; or an effort to parallelize something that already exists, isn't parallel yet, and isn't too complicated. So far I have the password-cracking thing, maybe dbzip2, and maybe some unspecified thing involving dumps.
I understand the context of the class and scope. I just couldn't think of a MW related topic other than that that would seem to need serious parallelism work.
Good luck.