Spent the last few days writing a blog post about a core aspect of the architecture of Praxis and the influence of the actor model of concurrency. Even if you’re not interested in Praxis itself, there’s hopefully some good general pointers to working with multiple media, and in particular working with low-latency audio.
The Influence of the Actor Model (Praxis architecture 101) … otherwise known as Good reasons to banish Synchronized from your code. 
I’d welcome any comments with thoughts, insights, corrections, stupid mistakes, similar models in use elsewhere, profound disagreement etc., etc.
Thanks, Neil