So many ways to skin that scm cat
Saturday, September 18th, 2004While pulling sources for tortoisesvn, I came across the better-scm site; from there, I found David Wheeler’s essay on OSS/FS scm tools, which primarily discusses CVS, subversion, gnu ARCH, and Monotone (the latter two were new to me).
Having recently moved to subversion, and being very happy with it, I was surprised to see a lukewarm review. Of course, my previous experience with scm has been primarily HP’s softbench, clearcase, and RCS (and, fortunately, the briefest of encounters with VSS), so I’m already lightheaded with being set free from the pessimistic-lock camp of version control. Liberal use of branching, and easy separation of commits and lands, has been a major major improvement in my sw development team lead experience. Funny just how long you can work in a cruddy setup without knowing how cruddy it is.
So Wheeler’s paper turned me on to the next scm things that I don’t even know that I don’t know. A basic difference looks to be centralized vs. decentralized models:
Some people believe SCM systems should primarily aid in controlling a centralized repository, and so they design their tool to support a centralized repository (such as CVS and Subversion). Others believe SCM systems should primarily aid in allowing independent developers to work asynchronously, and then synchronize and pull in changes from each others, so they develop tools to support a decentralized approach (like GNU arch, monotone, darcs, and Bitkeeper).
Though even centralization can mean different things; he quotes Bastiaan Veelo:
In short, one could say that Arch is centralized around a code integrator, and that Subversion (like CVS) is centralized around a repository.
Wheeler’s darc horse for scm leapfrog is darcs, something written in Haskell, a language I’ve never even heard of. His example of token replacement showing the handling inter-patch dependencies was pretty cool. But he doesn’t think it will happen: most scm tools are working well enough (yeah, CVS worked well enough for a long time), and Haskell’s functional programming is a serious barrier to contribution.
I had no idea what I was missing in scm. It seems wherever you go, someone has gone a lot, lot further.