Fnord

Random bits from a random nerd

Time to Git 'Er Done?

Source Code Management, or SCM, is a topic beloved by nerds and misunderstood (or ignored) by everyone else. Since we live and breathe the stuff, that kinda makes sense.

Me, I started with CVS at Fermilab, used it again at Argonne, and am now on Subversion at SDSC. I argued for Perforce, a commercial system that’s free for open-source, but the other developers hated it due mainly to a different conceptual model. (They were too used to how CVS works, I think)

Lately, I’ve been looking at the topic again since we may need to migrate repositories. The current best seems to be Git, a project Linus started after the BitKeeper fallout. Git is fully distributed and quite fast, and (finally!) seems to have killer merge support.

Since we humans are narrative-based, a personal anecdote: When I was at Argonne, the Globus group used CVS, with developers scattered around the world. Because CVS merge is horrible, they’d plan a ‘merge day’ where all the developers were locked into a room with a projector, just to do a merge.

No one enjoyed this. Needless to say, I try to avoid such pain in SCM myself. One of the reasons I converged on Perforce was its killer merge support. Subversion, while much better implemented than CVS, started out with the same merge algorithms and really hasn’t progressed much IMHO. If you want to do lightweight branches and allow developers to try risky (i.e. innovative) ideas, you need easy merging! From the Shuttleworth post:

The “time to branch” is far less important than the “time to merge”. Why? Because merging is the act of collaboration - it’s when one developer sets down to integrate someone else’s work with their own. We must keep the cost of merging as low as possible if we want to encourage people to collaborate as much as possible. If a merge is awkward, or slow, or results in lots of conflicts, or breaks when people have renamed files and directories, then I’m likely to avoid merging early and merging often. And that just makes it even harder to merge later.

The beauty of distributed version control comes in the form of spontaneous team formation, as people with a common interest in a bug or feature start to work on it, bouncing that work between them by publishing branches and merging from one another. These teams form more easily when the cost of branching and merging is lowered, and taking this to the extreme suggests that it’s very worthwhile investing in the merge experience for developers.

Anyway, the other day I found a nice essay on Stefano’s site about Git, complete with a link to a (typically incendiary) talk by Linus. He’s quite persuasive, and it does look cool. Time to play!

Links:

Update The title is for ghort, who introduced me to the phrase. His context lacked SCM, but wordplay is fair game as far as I’m concerned.

Update 10/12: A reader suggests I consider Accurev as well, commercial. See comments.

Comments