On 2011-03-03 23:12, William "Chops" Westfield wrote:
> Some of these modern systems (git, svn, mercurial) may have important
> features for web-based collaboration and large, distributed,
> development teams, but the startup effort seems to be considerable (do
> I really need a web server?)

I have found the distributed versioning systems actually make fewer=20
assumptions about the infrastructure. Git and mercurial are explicitly=20
de-centralized, with no need for a server of any kind.  Everyone has one=20
or more repositories (being simply a directory in some project root=20
folder), and you commit, push, and pull changes between them at will. =20
That's all there is.  There is no checkout/checkin, server/client,=20
etc...  The collaboration website just happens to be a repository that=20
lots of people push their changes to.  And to me (for git at least) it's=20
both completely clear and completely magic how every node in the=20
revision tree is identified by a UID.... It's globally unique!  That=20
makes so many things easier and eliminates the need for central naming=20
and coordination points.

> I'm not terribly happy about having to use so many different tools
> just to look at open source source, either.  Sigh.
I agree with you a bit here.

> I've now used rcs, cvs, clearcase, acme (which might be entirely
> internal), svn, git, and mercurial.  rcs is the only one I really feel
> like I understood to any degree; for the rest I just follow
> instructions...
>
Check out http://hginit.com/ for a good Mercurial primer.  It helped me=20
rethink and understand what was going on after having come from=20
RCS/CVS/SVN.  After that I felt more ready for Git .


Joe Koberg


--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.