Mike Mason wrote:
> I've been making some noises about how great I think Subversion is, but
> now I'm actually trying to convince some decision makers that we should
> use it for Real Projects(tm). So far I've waxed lyrical about the
> technical advantages of Subversion (atomic commit, efficient use of
> network being the main ones for distributed projects), but have to
> convince people that it's as good as a commercial product where you can
> scream at the vendor.
> - pedigree. Of the main developers I see floating around, I definitely
> know who Karl Fogel is, because his CVS book is sat on my desk at work.
> Anyone else mind name dropping themselves and why they know what they're
> talking about? (I know you know what you're talking about, but maybe you
> see what I mean...)
> - support. CollabNet are funding some of the development, will they be
> offering support? Anyone else?
> - maturity. We're not at 1.0 yet, and a whole bunch of folks won't even
> look at something until it's 1.0. I've used it since 0.30 and not had
> any problems, but that's (unfortunately) not much of an argument. Are we
> in beta, even?
Can't comment on the rest of this, but Subversion has been in Beta since
December 19, 2003. That was the 0.35 milestone release, at which point
1.0-stabilization was branched. If you have a look at the status page,
you'll see that 1.0 release is coming up within a month and a half.
> - tools. Another kinda-1.0 thing. I stuck with the command line client
> because I'm a l33t h4x0r d00d, but there are some less elite folk out
> there who like/need GUIs. TortoiseSVN, great. Subclipse, great idea, is
> that actively being maintained? Needs recompiling when a new Subversion
> comes out. IntelliJ plugin? Experience reports welcomed (although
> Tortoise was fine when we used it for CVS).
Can't really comment, as I use the command line client. ViewCVS-dev has
integrated support for Subversion, and it can be seen in action in
Subversion's own repository:
> The basic problem I have is that we have working source control, and
> even if it is painful, it works. I reckon it wastes vast amounts of
> developer time dealing with it (and the broken builds that result), but
> it hasn't eaten our source code. Well, not all of it. So it's much
> easier for someone to say "no, don't do anything, too risky" than to
> actually make a migration plan and figure out if it's doable and how
> much risk we're talking about.
Not too much I can offer there, unless you're using CVS, in which case it
makes sense to point out cvs2svn.py, the converter tool which appears to be
To unsubscribe, e-mail: email@example.com
For additional commands, e-mail: firstname.lastname@example.org
Received on Mon Jan 12 22:12:58 2004