Hi,

When using some IDE with CVS-intergration, like Eclipse, CVS is indeed
*very* easy to use and very comfortable!

On the other hand, I've heard that especially the Subversion-plugin for
Eclipse is not very sophisticated and not very stable...

Ciao,
Ingo

If you scratch the surface of CVS you'll find that it's not a pretty piece  
of software, but it does the job quite well (unlike Visual SourceSafe,  
which makes the Windows Recycle Bin look like a quality configuration  
management solution).  A decent CVS client makes it almost enjoyable to  
use.  I've always recommended SmartCVS (http://www.smartcvs.com), and the  
same author is developing an equivalent Subversion client that is  
currently available in beta.

Dan.

I think both are pretty much the same on ease of use. Both have similar
command line tools and there good tools to make them both easier to use.
There are good support for both in Eclipse. Subversion has a real nice
integration with Windows with TortoiseSVN. CVS requires more techy sort
of knowledge than Subversion (sticky tags, pservers. So while Subversion
is a little easier to use that's not it's strongest selling point.

The big advantages over CVS are:

- directories are versioned: CVS doesn't really support moving,
renaming, or deleting files in a very good way. In subversion the entire
tree is versioned.
- commits are atomic: Either all the files being committed are committed
or none of them are. No one else will see a partial commit.
- Access to the server is through standard URLs: If you use the plug-in
for Apache as your means of accessing the repository then you get a web
view of the repository for free.
- Supports WebDAV: If you are using the Apache plugin a client can
actually mount the repository as a file system. This is really only
recommended for read-only access however.
- Better support for binary files
- Branching and tagging are constant time operations: tagging in CVS is
proportional to the number and size of the archives
- Supports more operations off-line. Subversion keeps a pristine copy of
what you check out. You can do a diff to see what you changed and even
revert back without actually being connected to the repository
- Versioned metadata properties for files and directories

Subversion does take a paradigm shift in how you think about the
repository. In CVS each file has its own version number. In subversion
there is only one version number for the entire repository.

There are no tags or labels like CVS. In CVS there is filesystem space
(where it is in the tree), branch space, and tag space. In subversion it
is all just file system space. Branching and tagging is simply making a
copy, but that is actually an eficient operation (think of it like links).

I suggest reading the Subversion book which is an easy read. If you are
familiar at all with CVS you might want to start with appendix A which
describes subversion for CVS users.

http://svnbook.red-bean.com/