[BioLib-dev] BioLib

[Pjotr Prins]

Software projects often benefit from a rewrite. With BioLib I aim a
new attempt, revisiting some of the choices made in the beginning.

Two things I want to ask right now:

(1) What is a priority for you?

(2) Anyone object to using rake over cmake?

For (1), is anyone looking to use some C/FORTRAN/R library from their
favorite dynamic language right now? I may be able to help. 

For (2), BioLib has taught me that cmake is holding development back.
Biolib is more complex than most projects, as it needs to allow for
multiple languages, but also very flexible target building. I.e. I
want BioLib's Affio for Perl using SWIG with Debian include files and
the shared lib, or I want Affyio built from Biolib's source tree for
Ruby using ffi. CMake is great, in the sense that it is a powerful
iterative programming/scripting approach to generating Makefiles. But,
when it comes to exceptional cases, it is not enough of a proper
language, which makes adjustment complex and hairy to develop.  Rake
comes with Ruby these days, so a cmake dependency gets replaced by a
Ruby build dependency. I see two potential problems: (1) speed of
analysing complex dependency graphs - rake is known for that, and (2)
rake is not really strong on C/C++.  But, as a Ruby guy, I think I can
handle that. I spent way too much time working on CMake issues.

Pj.