[Chicago] Sprint, Imagescape

Ian Bicking ianb at colorstudy.com
Thu Mar 16 21:47:17 CET 2006

Brian W. Fitzpatrick wrote:
> On 3/14/06, Ian Bicking <ianb at colorstudy.com> wrote:
>>Andrew Dudzik wrote:
>>>On 3/14/06, *Atul Varma* <varmaa at gmail.com <mailto:varmaa at gmail.com>> wrote:
>>>    CVS just plain sucks compared to its competition
>>> Allow me to introduce you to a little program called VSS...
>>I think the better statement would be that CVS compares to SVN, and
>>mostly sucks compared to the new generation of distributed VC systems --
>>but it's really easy to compare CVS to SVN, and not as easy to compare
>>to the other systems (the distributed-vs-centralized debate is seperate
>>from the quality-of-tool debate).
> Hmm.  I'd give Subversion more credit than that.  Subversion is a CVS
> replacement, but it's *light years* ahead of CVS in just about every
> aspect.  I think CVS is pretty much dead and Subversion will continue
> to happily swallow up CVS' market share.  See
> http://subversion.tigris.org/svn-dav-securityspace-survey.html

I don't know what I actually meant to write; reading over what I wrote, 
apparently I left out some words.  I meant to say that CVS sucks 
compared to SVN, and CVS sucks compared to the distributed systems too. 
  Except maybe Arch, which seems just as crufty as CVS.  Haha, we can 
pick on Arch because it is dead.  (Or is making fun of the dead gauche?)

>>But notably every new generation of systems works from the CVS *model*
>>as a basis.  SVN copies is very closely.  Distributed systems change it,
>>but use CVS as their point of reference for changes.  And all those
>>other proprietary systems are just dead ends.  (Except probably
>>BitKeeper, which has been very influential)
> The biggest problem with the myriad distributed VC systems out there
> today is that they don't really integrate with the developer tools out
> there (Eclipse, IntelliJ, Xcode, etc.).

I've never really noticed this, but I don't use any tools (I just 
generate WebDAV requests by hand).  My problem with the distributed 
systems is that while they solve some process problems, they don't give 
any guidance about how to do CVS/SVN-style centralized development.  I 
like the idea of an escape hatch, but the bulk of development fits with 
how SVN works now.  Accomodating *only* the exceptional cases is a 
little silly.  And svn has become one of my primary filesharing 
protocols, which seems a bit heavy but I guess it works.

Ian Bicking  /  ianb at colorstudy.com  /  http://blog.ianbicking.org

More information about the Chicago mailing list