Network Computingwww.networkcomputing.com

To identify which version-control system we'd like to use, we first deployed each system on a supported platform (Microsoft Windows NT 4.0 SP6 or Red Hat Linux 7.0). During installation, we were able to examine closely the security mechanisms each system provides to control access to the repository. The process also gave us a good look at licensing issues (for those systems requiring licenses).

After setting up the server, we installed a client so we could test the actual function of the system. For CVS, we used the command line as well as an open-source, Windows-based client. All other systems included a client that we installed and used during testing. Once we had a client running, we created projects and did what developers do: check out files, change the source code, check source in, create branches of projects and source files, and then merge different source-file revisions together.

We also snuck a peek at the inner storage mechanisms of each system -- digging into the repository and examining how data is stored by the server. This gave us a good idea of the size to which the repository would grow when used by copious numbers of developers working on large projects.