[BioLib-dev] GSoC: Mapping JAVA libraries to Perl/Ruby/Python using Biolib+SWIG+JNI

Pjotr Prins pjotr.public14 at thebird.nl
Thu Apr 1 08:37:36 UTC 2010


On Wed, Mar 31, 2010 at 06:25:34PM -0300, Luiz Irber wrote:
> Challenging and exciting =D

Good :-). Remember GSoC is competitive. Really. It is important to
have a very clear plan - as detailed as possible - where you set
milestones. Since this project is about exploring possible scenarios I
suggest you spend the first 2/3 on possible solutions (part 1). The
last third on working out a working proof-of-concept that tries
different JAVA type arrangements and map them automatically against
Python (part 2). Perhaps, at the start of part 1, take the JAVA
toolkit of choice and write a bunch of Python unit tests you want to
see succeed when the conversion process works.

With part 1 and part 2 I would daily blog every thought and decision,
similar to my

  http://biolib.open-bio.org/wiki/Biolib_Documentation_1.0
  
so it is clear what you are trying and how. 

With part 2 set significant time apart to document online what the
proof-of-concept does and what works (and does not). This way we make
it easiest for someone to build on the new tools. 

It would be nice also to see it work with Jython - if you opt for the
reflection API. It is kinda satisfactory to have the Python bite in
its own tail.

Pj.



More information about the BioLib-dev mailing list