(June 7, 2012 at 6:44 pm)Moros Synackaon Wrote: That wastes bandwidth.
Better to build a program that uses a schema to intepret youtube's syntax and hash several identifiers.
When the hash changes, we know youtube (or the data) is changed/bad and signals a fetch from Darwinian's servers for an updated schema or "update the world" message.
I'm not sure this is necessary at the moment.
Provided that Darwinian is comfortable in the assumption that get_video_info will reliably return the URL of the video stream in the url element, it shouldn't be necessary to care about the other return elements.