Re: [Octal-dev] is a "gearlib" wanted/needed/possible/trivial?

[Dave O'Toole]

> anybody thinks it would make sense to develop a "gearlib" for octal
> to hold all the low-level dsp routines common to many octal machines.
> i think this could lower the entry treshold for writing new gear
> resulting in more varied gear available to composers and potentially
> minimizing the number of installed shared libs needed to play a typical song.

Hi luka. IIRC we had a discussion about this on #octal. I think a
gearlib is a pretty good idea. If it's general enough (i.e. set to
process buffers of type "samp" (an octal-project typedef) it could be
used in apps besides the main Tracker--for instance, other apps in the
Octal Project Audio Suite, of which the tracker (not decided on a final
name yet) is one part. 
 
> and second question, anybody know what should end up in this lib?
> or actually has the knowledge and time to coordinate this effort?
> or at least, what is already written well and what should be developed
> from scratch?

Personally I'd like to focus on the design and implementation of the
core system; I think that the gearlib should be by and for machine
writers, who, as the intended users of the library, will know best about
what to put in and how to structure the code. 

About providing space, when+where, etc. I think this will be in some
kind of machine repository (likely built on top of GNU's CVS server). I
am in the process of talking to fellow GNU people about what's the best
way to set this up. One snag is that GNU CVS is only for the code of
actual official GNU projects. But I **don't** want to make machine
writers go through all the assignment/legals etc just to write and
distribute a machine. I'm thinking of setting up an "auxiliary arm" of
the OCTAL Project at SourceForge. 

> i guess its better to talk about this in advance than to wait until machines
> start appearing and then try to pull this routines out of them... ;)

Yeah; on the other hand if people don't write machines because there's
no agreement on what goes into gearlib we have a problem. :-) Not that I
really think there's so much danger of that happening--but high-level
tracking-oriented unit-generator systems are a relatively new field. We
might want to let practice lead theory on this one. 

Overall the gearlib is something that's definitely a good idea. 


-- 
@@@ david o'toole
@@@ address@hidden
@@@ www.gnu.org/software/octal