At 10:50 AM 7/7/98 -0400, you wrote:

1. SourceSafe, if set up wrong, eats data. At a previous company I called it "sourceUNsafe". 2. Making a file appear in multiple places ("branch") is easy in sourcesafe, but it's much harder to keep related-but-not-the-same-file branches in SourceSafe. 3. The GUI interface is nice enough, but the command interface (which is what you'll use for scripting and for data mining) is terrible. 4. It's hardly fast. 5. It doesn't run on very many platforms, and you need to solve a number of Unix<->PC network issues to get it reliably up. 6. It's supported by Microsoft.

Also,

7. Creating labels is incredibly slow since it touches every file to create the label. 8. Retrieving files from the repository is slow. 9. There are no provisions for people working off-line (at least, there weren't when I used it a couple of years ago). 10. The standard response when it has data corruption problems getting to any file is to refuse to get *any* files from that part of the repository. In other words, if dev/src/include/x.h is busted somehow, you might lose access to all of dev/src. I hope you have good backups.

And if you think I'm being nasty, you should get me talking about ClearCase.

-Jeff Bowles