Simulation "runs" are capturing when the simulation is running. This data can be "replayed" in some fashion. There are two scenarios:

• capture just the agent and input state per frame, not simulation
• rerun the simulation with just the inputs replayed, and hope that it plays out the same way. This may work.
• to run simulation in reverse, we need to restore the previous simulation state from agent state and input set. We also have to "roll back" the PRG so our simulated events run the same if we want forward simulation to also work.
• we want to be able to run a separate annotation stream over the playback data too.

don't remember what it was

Currently a server has:

• expressJS app server
• PTrack UDP listener, forwarder
• URSYS socket server
• FakeTrack TCP listener, forwarder

We want to assign different ports for each appserver, which then will provide through the ur/netinfo webservice what the addresses for everything else are.

This runs on the presentation machine that is the main display. It has to read a WebRTC stream (which we have code for) and render it into the background of the stage.

Required to use webcam and microphones on a host computer. The WebApp has to be served from a secure host with HTTPS. However, we can not make a lan host "secure" because there is no universal certificate for it. It has to be handled by proxy.

One possibility is to have the webapp served from the central host, and IT can have the HTTPS stuff in it. So our LAN host would also contact that machine to get its configuration information.