IO protocol is non existent, Bot just decide they want to appear at another machine. So the only real protocol is sending and receiving complete Bson subgraphs. Each subgraph wont't hang around long, the engine runs it through the triple machine and it disposes of itself, often by requesting to be a static storage, and often by requesting that a whole or subgraph of the current bot go somewhere else.
So, the engine assumes it is talking to other engines, and each engine listens for outgoing and incoming, and responds with one send or one receive. I will get right to work on the send part, seems simple if the wait for connection routine waits for both outgoing and incoming.
The goal is to get engines making ACID sub graph transmission to other engines on the same port. Hmm...., seems easy, but I think I have to wake up the listener first? Mystery abounds! I will try it out, just up and send on the same port that net input is listening on.
Send it to whom? Hmm... I need another complete computer so they can all use the same ports, or else jury rig a test within the sngle computer.
Here at Imagisoft, we don't keep extra computers around.
Solution, I'll go pester the local computer science department at Fresno state. I fired up the marketing staff here and we send out a mass mailer to the two local colleges, enticing them to engage in Web Bot testing with Bson/Json and Sqlson.
No comments:
Post a Comment