>Despite that I strongly think that an audio server that not permit in >native way the traditional approach (what you call blocking approach) >will never achieve the driving role we'd need.
if linux developers continue to work with this traditional model, then yes, i think you are right and its a deep problem. but i see a brighter future because i think that slowly more and more developers will choose to adopt the same models they would be required to use on windows or macos. this will happen, even if its just because the developers will be coming from windows/macos :) besides, i think that the take-home message from so many of the discussions about audio servers is that there is no truly universal solution. what's needed are acceptable ways to move data between "worlds". the "world" (the non-MOTU) MAS represents is very different from the one JACK represents, but its quite possible to build a gateway between them. this does perpetuate the different programming models, but i think we just have to accept this. mac os x is the only system to have gotten this right so far, and i see no way we can do what they've done because nobody has enough control to impose a single solution. i suspect that even if someone stepped forward with a full linux implementation of the AU/CoreAudio API, there would be just as much resistance to it as there is to any other proposed "universal" solution. developers for mac os x don't have the choice. alas, we do. --p