[cgiapp] darcs? Really? (was: YAPC::NA talk proposals due today)

[Mark Stosberg]

> 
> darcs? Really? Don't get me wrong, I love darcs and am still using it, 
> but my repo's have gotten to the point where simple darcs stuff takes 
> forever. I thought I saw that the Haskell team is even moving away from 
> darcs?
> 
> I've moved to git and it's pretty cool. I haven't totally left darcs 
> because I haven't figured out git can work into our company workflow 
> (it's complicated ;)), but darcs' days are numbered. ;) And I don't 
> think I'm the only one...

The darcs community has been more active lately, not less. It is
growth, not decline that happening now in the darcs community. You can
see all the progress and people involved in the last hackathon:

http://blog.darcs.net/2010/03/darcs-hacking-sprint-4-report.html

Darcs has been a useful tool since before 1.0, and it continues to
improve now with the 2.4 release.

I am not blind to git. I use it regularly and contribute to a number of
repos on Github. 

It is precisely because I use both tools that I believe people should
continue to be aware of darcs as a useful, viable tool. 

I'm not sure why your darcs repos are slow. Perhaps you are using an
older darcs storage format, or an older binary. Perhaps you need to use
"darcs optimize" or add some tags so that pulls are faster. Our large
project has over 100,000 lines of code, nearly 4,000 files and over 100
megs of data managed by darcs. Our _darcs directory is over 700 megs. It
continues to be sufficiently fast for us. 

Further, we regularly save minutes if not hours through the more
efficient workflows that are unique to darcs. 

If YAPC is going to have several more Git talks this year, I hope they
will consider allowing an alternative perspective. 

   Mark

-- 
 . . . . . . . . . . . . . . . . . . . . . . . . . . . 
   Mark Stosberg            Principal Developer  
   mark at summersault.com     Summersault, LLC     
   765-939-9301 ext 202     database driven websites
 . . . . . http://www.summersault.com/ . . . . . . . .