> I'd still ditch the C, and build a stateless variant of the port-504
>protocol hosted in citserver itself. That would be enough to break the >depedency on having an "engine" per se in webcit, which is necessary >because webcit is currently acting as a TCP concentrator for port 504. Revisiting this now. I'm thinking we're actually going to land somewhere in between. Thankfully I have the space to experiment a bit here before committing to something. The structure I'm experimenting with right now is to write the "primitives" in C. For example, when the UI initializes and it wants to learn things about the server (server name, whether it supports anonymous access, things like that) instead of transmitting an HTTP-ized version of the INFO command, it will make an ajax request for /ctdl/a/info which will return a JSON structure containing all of that data. Then the UI, written in JavaScript, can do what it wants with that. We'll see how it works. So far it "feels right"
