On Sat, 23 Apr 2011, Heinrich Apfelmus wrote:
David Terei wrote:
Good chance you've already read this but if not here is a good post by Linus about his take on the problems with darcs:
I always have to smile at the complaint that something is "academic". :D
You know, like purely functional programming, that's soo academic. It's centered around some academic ideas, like mathematical functions, higher-rank types, monads and zygohistomorphic prepromorphisms, that have absolutely no relevance in real life, and that just don't work in practice. You do *not* want to write whole programs that way. At some point, you need something that works at another level than pure functions. What the *hell* do you do?
I also found the introduction about 'darcs' being too academic quite silly. However at the end of his invited rant Linus proposes a requirement (or may we call it 'axiom'?), that would be nice to be have: An identifier (a 'version') that can be uniquely mapped to a set of files and their contents. In Darcs this is the darcs history and it is usually the largest part of submitted darcs patches.