The responses from everyone are improving my understanding... I think. Thanks all.
On Wed, Jul 6, 2016 at 2:59 AM, Christopher Arndt <ch...@chrisarndt.de> wrote: > It's not uncommon for OSC software, that you have to specify client (a) > address(es) in the server as well, since OSC, commonly using UDP as a > transport, is only uni-directional, so to get bi-directional communication, > every participant in an OSC communication has to be a server and a client. This part, if I understand you correctly, I'm already doing. I'm creating a poor emulator of the Symbolic Sound Paca DSP's OSC. The idea is to send Paca-specific OSC messages to the emulator, and receive Paca-specific OSC responses. I have very limited access to the Paca and want to be able to continue development when away from the machine. Paca, in addition to OSC, speaks Bonjour / Zeroconf / Avahi. I've set up a Raspberry Pi with a known hostname and registered a Zeroconf service. In my test code, I query the network for the OSC zeroconf service, explicitly specifying the permanent zeroconf name of the Pi, just as I do with the actual Paca. That works. But the real, physical Paca figures out who called to it and responds to the caller. The caller has to use the fixed, permanent zeroconf address of the Paca, but the Paca does not have a hardwired address for who to respond to. _______________________________________________ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev