At the talk on Coda I gave yesterday, someone asked what the advantage
was of using Coda instead of just making local copies of files.
I began describing the differences, and then I realized: Coda's
functionality closely mirrors that of CVS.
Coda provides, but CVS doesn't:
- fine-grained access control with ACLs
- creation and deletion of directories
- application-specific merge processes (although CVS can use such things)
- replicated servers
- transparency
CVS provides, but Coda doesn't:
- access to full version histories of everything
- a huge support base
- a guaranteed-stable code base for storing things
- user-controlled merge times and user-visible merge status
It seems that the obvious thing to do would be to add Coda's features to
CVS, perhaps as a layer over the top (the way CVS is a layer over the
top of the RCS code), particularly since Coda is so huge and difficult
to build.
