RFC: startup time - again

[Matthew D. Fuller]

On Wed, Sep 10, 2008 at 03:20:55PM +0100 I heard the voice of
Adrian Wilkins, and lo! it spake thus:
> 
> Competing with git on raw performance is daft. The design of git was
> optimized to be good at merging a very large number of kernel
> patches, because that's what Linus does. Other considerations in
> it's design were secondary, and it shows. My userbase would never
> have grokked the default porcelain in git, so for them, git would
> have been a lot slower than Bazaar.

I don't buy this for a second.  git isn't fast because it has a wacky
UI; it's fast because it's _fast_.  bzr isn't slow because it has a
nice UI; it's slow because it's _slow_.

The essential differences between git and bzr are mostly in the user
model.  Very few of them necessitate any performance difference, and
those that do don't necessitate near the difference that exists.  bzr
is slower because it's doing gobs more work, but most of that work
isn't gaining anything.

I use bzr instead of git because I like the user model better, and the
speed difference doesn't, in my usage, eliminate bzr from
consideration.  That doesn't mean I'm thrilled with the speed, though
(nor that all of the twitching comes from operations dominated by
startup time, but hey, one thread at a time  :)


-- 
Matthew Fuller     (MF4839)   |  fullermd at over-yonder.net
Systems/Network Administrator |  http://www.over-yonder.net/~fullermd/
           On the Internet, nobody can hear you scream.