We've used RCS, and then CVS, and now "ClearCase" at cisco to manage
development, as we gre larger.

As we used it, each file was individually lockable/check-outable by
a particular user, IIRC, but the directory hierarchy itself was shared.

CVS uses RCS underneath and is much more complex.  We needed to have our
unix wizards build additional scripts to make using it in our environment
easier, although *I* never understood it well enough to know the
differences.  Periodically someone would mistype a command and require
CVS-wizard services to set things right.  As we used it, each engineer
would get a full copy of the directory structure, insulating them from
changes made by others (till merge time, heh heh.)

I'll assume ClearCase is out of reach.  It hacks into the file sytem to
provide an "virtual" directory structure that gives you a particular master
file or your own private copy, which makes it scale better.  The way we
implemented it (large amounts of NFS) seems to create performance issues
(all you need to do is buy much faster hardware :-)

Historically, things work well when people work in separate areas of the
system.  When there are multiple changes to the same file that must splat
together correctly, THATs when there are problems.

BillW
cisco