Originally Posted by mr.cello
Excuse my ignorance, but why is it necessary to have a PC running with the TiVo server on it to use the app (which sounds super cool by the way)? I would think since TiVo has a networking stack (and they have all those Internet enable services running on the box stand alone) that it should be possible to do the same thing with this app. Is there something in the HME architecture that stops stand alone network applications?
Great question. HME supports two different models for applications (like mine, and like the Yahoo! apps) to be hosted -- either on a PC on the local network, or centrally, on some big server farm in the sky.
So, I could (in theory anyway) host my applications on my website (http://bitrazor.com
), but this means that every instance of the app is hosted there. This means that in order to support lots of people, I'd have to have a reasonably scalable server farm. And, in my case, I pay for bandwidth -- http://bitrazor.com
will do only 35GB of traffic per month for the rate I'm paying (and I already use half of that today in a typical month). So, if all 4.5 million TiVo subscribers decided to use my app, it'd kinda blow the budget quick.
So, technically it's possible, but financially it doesn't make sense for freebie-app writers like myself to centrally host. Yahoo can do it, but they've got more than enough server capacity and bandwidth to handle it.
Edit: I should probably clarify -- the TiVo box itself is not powerful enough to run the applications directly. The HME protocol is a low-level protocol that exchanges views, keystrokes, and events with some other host -- that host does all the hard work. This is the case with the Yahoo apps -- they might appear to be running on the TiVo, but they're actually not -- they run on Yahoo's servers, and your TiVo simply receives the "views" from those remote applications.