[darcs-users] GSoC: network optimisation vs cache vs library?

Eric Kow kowey at darcs.net
Wed Apr 14 09:32:16 UTC 2010


On Sun, Apr 11, 2010 at 19:39:58 +0200, Reinier Lamers wrote:
> I think a GUI is very important, almost just as important as performance. If 
> we were in a fairytale world where we could get two GSoC students, I'd rather 
> see one work on performance and another on a GUI than to see both work on 
> performance.
> 
> As for the original question of network vs. cache, I think the network is more 
> important, simply because it is the harder of the two.

I agree that GUI work is very important, but

(1) I tend to feel we're not really in a position to make much better
    progress -- although the current refactor proposal would put us in a
    better position to do so and

(2) It's probably good to avoid spreading our attention too thinly,
    to sort of focus on one thing and show the community that we can
    really deliver.  Working hard to get one thing and really get it,
    rather than trying to get two things and never quite deliver on
    either of them...

Then again, our actual experience with Darcs may be proving me wrong:
 - Ganesh did great things with the hunk-split work and now rebase.
 - Florent is doing a very useful SelectChanges refactor (which will make
   a future GUI much easier).
 - Reinier's fixing our UTF-8 support.

and this general bottom-up spirit of people working on what interests
them makes me happy :-)

In the meantime, to the extent that there are top-down elements to
the project, I'll continuing going with the stick-to-one-message
approach (and hope desperately that this is the right thing to do).
For now, I'll continue spinning the Darcs story as

  For a few years after 2008, the Darcs Team worked really hard on
  performance problems and made it much faster for daily use.  Having
  delivered the faster darcs they promised, they then set their sights
  on the Darcs library paving the way for a Cambrian explosion of tools
  in the Darcs ecosystem.  So the Darcs team grew and new developers
  took on the day-to-day maintenance of the Darcs tree, while the
  original team from 2008 retired to work on what would be Darcs 3.
  And so in two decades, Darcs took over the world.

Hmm, maybe that's going a little far.  :-)

-- 
Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow>
PGP Key ID: 08AC04F9
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: Digital signature
URL: <http://lists.osuosl.org/pipermail/darcs-users/attachments/20100414/d7bb29c4/attachment-0001.pgp>


More information about the darcs-users mailing list