In the early days when I was programming on DOS we had no version control software, and PKZIP was typically used to archive releases (crude but functional). I knew from a brief exposure to Xenix about SCCS, and what real version control was like, so I created my own version control system for DOS. I called it VCS, and it used a basic folding technique to store multiple copies of a file in a single version controlled file. This worked really well for source code, as subsequent checkins were usually very similar to the previous version.

RCS

Over time I moved on to developing on Unix systems, and came across RCS, which I used for many years, on Unix, Linux and Windows.

CVS

CVS was a massive step forward, and although only really a front-end to RCS it changed your perspective by working on collections of files rather than single files.

Subversion

I have now settled on Subversion as my version control system of choice, and despite the rise of the distributed version control systems such as git, mercurial, darcs etc…, Subversion is still an excellent choice for most software projects.