[Fluxus] rubyk & fluxus

[David Griffiths]

Hi Gaspard,

On Thu, 2010-09-16 at 15:44 +0200, Gaspard Bucher wrote:
> With Swig, you have to create "mapping" files, for a project like
> OpenCV, this looked daunting to maintain (at least for me). On the
> other hand, the "dub" binder understands the C++ files (thanks to
> Doxygen) and creates bindings from trivial templates (one for each
> language). Since everything is automatic, the bindings very closely
> map the original C++ methods, objects and such.

Yes, this makes sense. Although one of the problems I've found with
automatic bindings is that they can sometimes give you strange
interfaces that don't really suit the language being targeted. 

> The fact that Swig looked overkill, huge and that reading the manual
> would have taken me as much time as writing Dub is another
> (irrational) reason.

Also makes sense - fluxus doesn't use swig either, fwiw - I just wonder
sometimes if it should :) Another option that Gabor has used more than
me is the ffi in racket, which is really comprehensive for accessing C
libs from scheme.

> My turn to be curious, what is the relation between (fluxus) and
> impromptu ?

A seething vi/emacs like religious war - no, nothing that exciting
really - I had a beer with Andrew Sorenson once...

I've seen performances with it by both of the Andrews and Michele Pasin,
it's facinating to watch as a schemer - as obviously it's much more
understandable (for me, anyway) than SC or chuck, so it's a different
experience as a member of the audience. 

We could probably do more in sharing approaches/ideas, but that's
probably a bit too practical for the livecoding scene :)

cheers,

dave