- page 6

Preface

I was first introduced to version control (and CVS) in college, about the same time I was introduced to Linux. At that time though, most of the projects I worked on were small and generally involved only a couple of developers. So, although version control would have been useful, I never took the time to really use it; my knowledge of CVS remained limited to what little I needed to know to check out the occasional bleeding-edge project on Linux (which seemed necessary a little more often in those days). As my college career progressed, the projects I worked on became more involved, and I began to learn about "software engineering." The instruction I received on software engineering never really covered version control in any depth though, and despite the increased size of the software projects I was working on, I never delved into using a version control system to keep track of things. I wanted to; I thought CVS was a neat idea. I just never invested the time necessary to learn how to set it up and use it. Then came my first major team project. It was a real-world project, with real-world clients , and its completion was required for graduation. Finally, I had an excuse to really give version control a try. I presented the case for CVS to my teammates and (although there was some small resistance) convinced them that we needed to use it. It was a success. By the end of the project, I was fully sold on the necessity of version control in any future projects, however big or small. I loved CVS.

After school came the real world, and the love affair with CVS didn't last long. As I learned (mostly through trial and error) how version control systems should be used, CVS steadily became more and more inadequate. I could see its potential, but it didn't measure up. Code was lost, fits were thrown, and hair was pulled. Still, CVS was the best free, open source version control system out there, and as an entrepreneur trying to keep a start-up company going, free was a required feature. Then someone told me about a new version control project called Subversion, so I went to its site and took a look. It seemed intriguing, but it wasn't quite up to the point where I could trust it for my codeand I barely had time to eat back then, so getting involved in the project's development was out of the question. Instead, Subversion went on my back burner and I moved on to other things.

Several months down the road, I saw that Subversion had become self-hosting. "Well," I thought, "If they trust it with their own code, maybe it's time to take another look." Rolling up my sleeves, I sat down to play around with it. Once again, I had fallen in love. Subversion was everything CVS could have been. It was stable, it was flexible, and it didn't eat my code. Thus, after a suitable period of testing, CVS was unceremoniously chucked and replaced by Subversion. I've never regretted the change. In fact, the only thing regrettable is the hours of my life wasted fighting with CVS.