I also wanted to pick RCS (I ended up picking CVS, because that's what we use at the lab).

RCS is nice because it's simple. You don't have to worry about pservers or creating a repository somewhere with proper permissions... just say mkdir RCS and you're there. Checkin/checkout is simple, with no extra commit stage. It's quick and easy and doesn't require any setup or administration.

I like using RCS during OS installs (a trick taught to me by a sysadmin I worked with, the kind who had plenty of time on his hands to teach newbies because his network rarely broke). Before messing with any config files, make an RCS directory and check them in. We had /etc/RCS, /usr/local/etc/RCS, and a lot of /var/www/foobar/RCSes. Handy when you want to back out a change, but haven't installed CVS yet.

I don't think I'd want to try using RCS for big projects, and using RCS for big distributed projects would be a nightmare. CVS is better for those (and I'd like to try Subversion one of these days), but that doesn't mean I like it any better.