> CVS would have been *just fine*.  Subversion would have been *just
> fine*.  However, if he really wanted patchsets (which is what git does,
> but badly) then darcs or arch would have been *just fine*.

How do we know those would have been just fine?  Perhaps with the
rate of patch submissions Linus has to deal with he NEEDs
a customized system with no extra distractions.

How do we know we can extrapolate our assumptions/experiences with
our group sizes up to the entire world community of the
Linux kernel and know for sure the same holds true?

> darcs, in particular, seems to be garnering an unusual amount of
> attention lately.

I haven't used darcs so can't comment.  However, compared to CVS,
Subversion is divine.  I love it to death and would never go back.

I wonder if Linus at least considered for a nanosecond whether it
would have been faster to mutate the Subversion code base into what
he wanted rather than start from scratch.  I'm sure the Subversion
developers would have jumped at the chance and worked round the clock
to win the kernel guys to their camp.  Your suspicions seem valid.
The more I think about it, the more odd it seems that Git exists---
unless there is some obvious reason you couldn't convert Subversion
into Git that I'm not seeing.

Chris


-- 
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list

Reply via email to