I've done the coding for the transition to the new framework architecture for Jabber, and I'm now testing it, trying to run down all the code paths, checking my TODO list, seeing what happens when I disconnect the network. Also got a bit of documentation to sift through and correct, which I'm doing right now.
I also want to once again look into the possibility of adding Jabber functionality to xml.rpc, since I don't need a connection reference any more. (I forgot to list this as another good reason to transition to this architecture.)
It's looking good; failure is still a bad thing, as some messages may get dropped on the floor, but I don't think I need to nursemaid the connection any more. Perhaps most importently, it feels significantly more reliable. So like I said, this should be the beginnings of production quality code. (This is a benefit to me being on the @Home network; I had to handle bad network connections early, I couldn't just brush the problem under the carpet.) I wish my IM client would restore connections so transparently for me!
The only bad news is that the conferencing protocols look to be a disasterous mess right now. A couple of the sample uses of the framework I was hoping to implement are looking mighty iffy. I had hoped to push your news to a group; it looks like you might only get it to a particular account. I was also hoping to implement something for Weblogs.com to easily push the updates out on Jabber, but without a highly configurable conference group (I'd want to make sure only the Weblogs.com thing could post, or it'll get deluged in spam), that won't happen. I can't even find a spec I can code to, let alone a sample implementation to test against. The right answer might even involve programming a module for the Jabber server. (I could do it, but that's a bit more then I'm willing to undertake.) I don't even know that the servers even have the capabilities I'd need right now.
It's interesting to look at your neighbors on the net when you have your own domain name. Obviously, I have "JERF.ORG", which you think would be kinda isolated, but that's not true.
I receive a lot of messages directed to INFO\@IERF.ORG. (All caps deliberate; I get stuff from IERF precisely because j's look like i's.) IERF seems to be a teacher re-certification group; if you have a foriegn country certification for teaching, and would like to translate that into a USA certification, they're the ones you talk to.
A new one I just received a message for is JDRF.org, which is the "Juvenile Diabetes Reasearch Foundation". The Internet makes for strange neighbors.
Also includes comments on the SSSCA.
'"It's the same information as the front of the license," said Frank Mandelbaum, chairman and chief executive of Intelli- Check, a manufacturer of license-scanning equipment based in Woodbury, N.Y. "If I were to go into a bar and they had a photocopier, they could photocopy the license or they could write it down. They are not giving us any information that violates privacy."'
I wish I could take this guy's computer off his desk. It does nothing that couldn't be done by hand. It doesn't give him any capability that enhances his ability to do business, as long as he has paper, a pencil, some envelopes, and some stamps. So why does need a multi-thousand dollar computer?
Meta-point: We often contrast "qualitative" and "quantitative" as if they were always opposites. But they are a continuum: Enough "quantitative" change makes a qualitative change, eventually. There is nothing a computer does that could not be done by hand, and there never will be; that's the point of a Turing Machine. But somewhere between seconds-per-operation and billions-of-operations per second, we get qualitative changes. And suddenly, things become possible that weren't before.
But does that method of figuring it out count? *chuckle* Been looking forward to this for a while, and it's done in style, too.
|<- Future Posts||Past Posts ->|