March 13, 2006
Program D 4.6 final
The final/"official" release of Program D 4.6 is now available. With almost exclusively very minor tweaks, this is the same as release candidate 3. The one exception to this is the introduction of some memory optimization techniques in the graph building process. (I'm still working on trying to pack as many bots into a single running instance as possible.)
I'll address bugs and worthy feature requests as they come up, but otherwise, Program D development will now be focusing on a new 5.0 stream that will see some significant architectural adjustments, with an aim toward expansibility.
Posted by Noel at 01:05 AM | Comments (0)
March 06, 2006
New IRC Listener
Wow, can't stop it today. Found a great library (PircBot) for IRC connectivity from Java, and threw out all that yucky old code from the IRC listener, replacing it with a few methods that neatly start up the PircBot and let it do all the work.
There's plenty that could be done with this listener, given the rich API provided by PircBot, but for the moment I'm satisfied just to get this out there to those (apparently many) users who use Program D with IRC. Hope this will help!
Posted by Noel at 12:27 AM | Comments (0)
March 05, 2006
New AIM Listener
Since the daim library, which was used by the ICQ/AIM Listener, seemed to be no longer able to work with either the ICQ or AIM networks, and since it appeared that the author wasn't really interested in supporting it, it became necessary to find another library.
There was some interesting discussion on the Program D mailing list about whether or not to take a completely different direction, but in the end I was able to locate what looks like a good library, JAIMBot, and quickly adapt that to the needs of a Program D listener.
Weirdly enough, in a recent message on the alicebot mailing list, Richard Wallace seemed to be advocating against open source when he warned against downloading "third-party" free AIML software, suggesting that some of it "may be infected with the spyware". The message was, of course, a rather obvious FUD-based advertisement for Pandorabots. I can't help thinking that Rich is not really the best spokesperson for Pandorabots--he's always positioning them against the rest of the AIML interpreter development community, advancing these dubious notions that it's better to trust a hidden-source web app than free code you can download and inspect yourself. I just don't think there are that many people who are still so undereducated about free and open source software that they'd fall for that, so it tends to backfire and make Pandorabots look bad.
Pandorabots, just for the record, sounds like a cool piece of software, and I'd love to be able to look at the code. I can't call myself a Lisp hacker by any means, but I do know enough to write a simple program, and certainly to look through code, and I imagine that if their claims about capacity are true, there have got to be some very interesting things that can be learned from that codebase. And certainly we'd all benefit from that application joining the ranks of the other free/open source AIML implementations that have been made freely available for everyone to dissect, learn from, criticize, rearrange...(not to mention deploy).
Anyway, here's another example of free & open source software in action: Program D benefits from JAIMBot, so the whole AIML community benefits, and JAIMBot will benefit as well if we discover any bugs or issues, because we'll report them, and be able to interact with the developers in dealing with them.
Posted by Noel at 11:16 PM | Comments (0)
New release candidate, and new listener
Today I updated the Yahoo Listener to version 1.1.1, adding a small but important feature that filters out HTML and formatting codes from inputs sent by Yahoo clients.
Also uploaded rc3 of Program D 4.6, which has important fixes in URL resolution techniques that should address problems some people were having running Program D on Tomcat on Windows.
Looking forward to declaring a final release of 4.6 soon, once the ratio of confirmed open bugs to daily downloads stabilizes somewhere close enough to zero.
Posted by Noel at 03:56 PM | Comments (0)