[fluxus] under the shade of abstraction

[Dave Griffiths]

>> Shriram Krishnamurthi is Dr Scheme? :) I'm quite close to writing an
>> experimental MzScheme version of fluxus btw.
>
> if you need a prod, ive been close to doing this a few times, but in
> practice never got much past section 1.2 of 'Inside PLT MzScheme'. the
> two major motivations for me, were the dificulties encountered during
> the guile 1.6 -> 1.8 transition (mainly on osX), and seeing how stable
> and fast the 3m gc has become compared with the conservative collector
> for fluxuslike data mulching. i think it would worthwhile branch.

It doesn't look like the api is very different from guile, so I don't
think it will be too hard. I'm thinking of plt mainly due to practical
reasons - it's hard getting support on the guile mailing lists, and the
plt lists seem to be much busier.

Whatever the interpreter, I'm thinking of making most of the procedures
extensions, and the fluxus app just an editor/gl window with a scheme
interpreter, into which you load the fluxus extensions. I've already split
up most of the functions into groups ready for this. I think you would be
able to do offline geometry processing, or use the osc code seperately
from the fluxus app in this way - there seems to be a glut extension for
MzScheme, which you could use to create windows in scheme, and render the
fluxus scenegraph wherever you want etc etc.

I'm not sure on the timescale for this as of yet though.

The priority for me in the short term (other than *using* fluxus a bit
more than hacking it) is to write a scheme script that parses the binding
source comments in a doxygen like manner and creates another scheme script
containing a list of documentation strings to be installed centrally for
an online help system. It's fairly easy I think - and should be
independant of the interpreter used - unless plt scheme has a killer
documentation system (the guile one seems to be broken).

cheers,

dave