On Mar 5, 2005, at 2:23 PM, Max Bowsher wrote:
> Exactly what is our goal w.r.t. DAV and DeltaV ?
>
> I've never really thought about this closely before, but the
> impression I've got from snippets of conversation over my time here is
> that we began with the goal of "be a DeltaV server", but seem to have
> put that aside, mainly because it would be a lot of work for little
> practical benefit at the moment, and now seem to be settling for the
> slightly more modest goal of being a good autoversioning DAV server.
>
Your analysis is basically correct. When Greg Stein wrote the first
draft of mod_dav_svn, the original goal was to map DeltaV concepts to
subversion ones. Unfortunately, the semantics don't match exactly. So
the plan was:
a. implement the parts of the DeltaV spec that svn needs
b. use custom REPORT methods for things DeltaV can't provide to
svn
c. someday in the future, implement the "rest" of the DeltaV
spec.
(a) and (b) have been accomplished. (c) is still a dream.
The problem is that over the course of 5 years, there's been
essentially zero motivation to do (c), because there are no DeltaV
clients out there in widespread use. In the meantime, though, plain
old DAV clients *have* gotten popular, and there's an enormous benefit
to to interoperating via "autoversioning".
So we now have this new goal of
d. be a good autoversioning DAV server
At this point, I don't think anyone is thinking about (c) other than
Greg Stein. There's very little motivation.
On the flip side, we've never had the goal of making the svn client
into a "good DeltaV client". It depends on too many concepts to ever
work against a generic DeltaV client, and besides, we've never found a
DeltaV server to test against.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Mar 5 21:39:20 2005