Re: [fluid-dev] OSC support

[David Henningsson]

On 07/09/2012 03:17 AM, Element Green wrote:
> Having said
> that though, I don't personally see an issue with implementing OSC SYN
> support in FluidSynth, provided there are actually applications which
> could interface with it in a usable fashion.  It could become a
> compile time option, for those who don't want to build that
> functionality.
>
> If someone came up with a working and clean implementation of an OSC
> SYN enabled FluidSynth, I doubt anyone would object to merging the
> support into the main project.

I don't know much about OSC or OSC/SYN, and have not used it myself. And 
that's part of the problem, really, because I don't know what it can do 
for us, how common it is, how standardised it is, and so on.

But we are a library, and as such, for us to be successful, we have to 
be useful for other people and applications - the more the better.

OTOH, we don't want to merge something, see the original contributor run 
off, and after a few years with some refactorings inside and so on, 
maybe the OSC support breaks and we're left with bug reports for stuff 
people don't know how to fix (because none of the original developers 
use OSC), and FluidSynth losing reputation as a result.

Not unlike suggested by others, I think we could start as something on 
top, i e, as a binding layer that calls into the fluid_synth_* methods. 
The OSC layer would be calling these methods, and we would be open to 
take patches that makes support for OSC/SYN easier/possible, if that is 
necessary. As both Element and Pedro has pointed out, FluidSynth API can 
do more than what is possible to express with MIDI already, but there 
might be gaps which we discover now - or along the way.

If it then proves to be reasonably mature and bug free, useful and used 
by more than one person, and we have people on this mailing list that 
can support user's questions and problems in the area, I think we should 
merge that code.

Does that sound reasonable?

// David