On Tuesday 17 September 2002 07:43 pm, Craig R. McClanahan wrote:
> On Tue, 17 Sep 2002, Steve Downey wrote:
> > Date: Tue, 17 Sep 2002 19:13:59 -0400
> > From: Steve Downey <[EMAIL PROTECTED]>
> > Reply-To: Jakarta Commons Developers List
> > <[EMAIL PROTECTED]> To: Jakarta Commons Developers List
> > <[EMAIL PROTECTED]> Subject: Re: [Proposal] Scamp: Source
> > Control Abstraction
> >
> > On Tuesday 17 September 2002 06:15 pm, Daniel Rall wrote:
> > > In what ways does this differ from DAV?
> >
> > A SCM system can sit on top of DAV, or DAV + deltaV. Subversion, for
> > example, is using deltaV as the communication protocol between its client
> > and server. But DAV, in and of itself, isn't a generic SCM protocol.
> >
> > Certainly CVS doesn't speak DAV.
>
> But Subversion does, doesn't it?

Yes. Delta-V is the versioning protocol for DAV. DAV by itself is just 
checkout - edit -replace, with no real versioning. Delta-V adds the SCM 
notions of workspaces, versions and configurations. Its fundemental goal is 
to avoid the lost update problem, where A gets, B gets, B puts, A puts, and 
B's work is lost.

But, even with Subversion, a random DAV/Delta-V client can't communicate 
correctly with the Subversion server. At least according to their docs. I 
don't know why not, but I suspect that it has a little to do with the fact 
that DAV amd delta-v treats the Web like a filesystem. Work is supposed to be 
done 'in-place' on the server. 

It looks to me as though the model is much closer to that of ClearCase, rather 
than that of CVS. 

In terms of SCaMp, I think the key questions are if this is to be a general 
purpose SCM interface, or a targeted subset of use to automated tools? I 
think the latter will have a much narrower interface than what is shown in 
WVCM.




--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to