[darcs-users] 'Darcs' Theory of Patches Isn't Useful'

[Mauricio]

> "The "theory of patches" proposed in the Darcs documentation is a nice
> slick theory, but not particularly useful to programmers. It is
> generally impossible to programmaticly determine whether one patch
> must precede another in the revision tree of source code for any
> reasonably complex program. (...)

I like that. Darcs works badly with too complex or caotic
patches, and works great with well organized patches. So
I try to plan and write my patches as clean and easy to monitor
as I can. That is good for darcs and good for me.

My personal opinion (based on my personal experience, of course,
so it may not be usefull to others) is that complex software
should not be done. Sometimes, when things start to grow, I
separate pieces of code and make them independent libraries
(and write a patch to remove then from main code). But, of
course, I imagine there are situations where that can't be
done. But then, I think, no sofware can (and neither should)
be the ultimate solution for everyone.

Best,
Maurício