I got to talk to an old acquaintance, last night. It was good to see him, as we hadn't spoken in a while and he is an older developer who also experienced the days I so frequently romanticize. This means that he can both understand me and offer useful insights.

In talking about some WebRTC ideas I was toying with a few years ago, and had abandoned as not being possible within the design of its browser-level interface, he described a way that he had used lossy user-defined data channels to pass around audio data in a more flexible way than is provided by the native audio channels, while still preserving its light-weight UDP (or equivalent) nature.

The idea was clever and I realized that if I adapt my system design to use a wire protocol like he described, it might actually work.

The old, "how do I validate the authenticity of video data if I can't access the video channel raw data?" could be re-written as, "I can capture video manually, package it however I want, pass the entire payload through user-defined data, then manually stuff the data into a video source object on the receiver's side." Not sure how well it will work without any of the signalling data the native video channel may send, but it is technically the data I need with the network behaviour I require.

Hmm, I wonder if that would work.
Jeff.