latest blog entry
I broke the record but I broke my back going over the waterfall
March 19th, 2005, By Duncan Gough
// TOC, AIM and GrokItBot
I think that I always knew this would happen. The TOC protocol that I use to piggy-back messages over the AIM network, which, to my understanding, is there for robots to access the more refined OSCAR protocol, has become far to unreliable.
Hunting around for information, I found this quote from the developer of Py-TOC, *the* library for writing AIM bots with Python:
“My understanding is that AOL has chose to introduce intentional service interruptions and limitations to TOC to encourage the use of clients using OSCAR, such as their official AOL Instant Messenger client. I’m afraid there is not much I can do about this.”
http://www.jamwt.com/Py-TOC/
I can understand the decision but it doesn’t mean I have to like it. For a while I thought about scrapping a few projects like GrokItBot and another version of that same code, SuttreeBot, that I use for updating this website. The only real alternative was to port all of my AIML and Bayesian code over to OSCAR. The only real Python implementation of OSCAR is part of Twisted. And what do I know about Twisted?
It’s taken the best part of a day when I should have been out in the garden playing catch with Ryan but somehow I managed it. I say somehow not because the Twisted implementation is hard to work with, far from it. The biggest headache was porting my AIML and Bayes code over to work with OSCAR.
Once I tidy it all up into something much more presentable I’ll release the code. In the meantime, wouldn’t it be nice if AIM sent out a warning message to anyone using the old TOC protocol, rather than just dropping connections and suspending three of my accounts?
– http://www.mudcat.org/thread.cfm?threadid=1690&messages=7