I find this discussion frustrating, because my impression is that Tom
wants to discuss the design of an Ideal SCM system, and somehow use
that as a basis for critiquing Subversion.
But there's a fundamental schism at the very root of this
conversation: two years ago, the Subversion designers/developers all
took this statement to heart:
CVS, despite its problems, is still a good version control system.
We like its model, and now we're going to make it better.
But you, Tom, have never believed that lemma, so you think Subversion
is therefore doomed and/or bad for the world. You seem annoyed that
we didn't spend careful months proving theorems at the beginning. But
it's too late to make us feel guilty for taking this path. We've
never been concerned with writing perfect software or discovering
great new concepts in SCM design. We want to scratch an itch, and
we've done a good job of that. Practical value is still that: value.
There's no shame here, nor are we attempting to bury criticisms
because of commerical or other "corrupting" motivations. We have a
different set of values and priorities than you, plain and simple.
I have a great admiration for projects like Bitkeeper and Arch; these
systems really have broken new design ground. But that's never what
Subversion was about. Yes, it's sometimes sad that the world takes
more notice of practical over theoretical value; but this shouldn't
surprise you either. Larry McVoy resolved this tension by funding
ground-breaking research via royalties, and we've resolved this
tension by simply placing research low in our priorities: "evolve as
you go". But it makes no sense for you to throw ire at both camps;
you need to find your own way of resolving the tension.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Aug 13 20:07:48 2002