Hi Marco,
as I said before I really like the idea of introducing a just in time compiler, and I like the ideas you presented here. the only thing I'd say is that I'd avoid introducing in Pd a pseudocode (max/gen~ style). in my opinion using (if possible) an existing scripting language (Lua, Python, etc ) would be a better option.
then, about having the possibility to write sample level patch in an optimized way sounds amazing. I'm just wondering if it wouldn't be a better idea trying to better integrate Faust. my only concern is about maintainability of this new object family, JIT (in case we're talking about a specific object family). Faust is already out there and works pretty fine plus has an active and big community that constantly works to improve and update it.
last but not the least, multithreading is obviously a good thing, but honestly do we really need to have something similar to what you described? don't get me wrong, the idea is cool and I only want to discuss these ideas, not criticize them. but personally, and of course I know everyone has different approaches and different needs, I think there's still a lot to explore before "moving to a multithreading dimension". also if it's a totally different situation (different hardware and programming language), I keep thinking about most of the commercial products, musical instruments (also the most complex) that are single thread based.
mine of course are just thoughts.
cheers,
Mario
2018-04-01 5:46 GMT-03:00 mario buoninfante mario.buoninfante@gmail.com:
as I said before I really like the idea of introducing a just in time compiler, and I like the ideas you presented here. the only thing I'd say is that I'd avoid introducing in Pd a pseudocode (max/gen~ style).
but besides pseudocode, Gen~ also deals with object boxes, for people who can't write text. It would be cool if we could take all of vanilla objects and compile it as an object for optimization, that would be like using Pd itself as a language and compile it like libpd. Am I dreaming? does something like this make sense at all?
then, about having the possibility to write sample level patch in an
optimized way sounds amazing. I'm just wondering if it wouldn't be a better idea trying to better integrate Faust. my only concern is about maintainability of this new object family, JIT (in case we're talking about a specific object family). Faust is already out there and works pretty fine plus has an active and big community that constantly works to improve and update it.
yeah, for what I hear, Faust would be nice, huh?