Hello use perl journal. I'm still not dead, I'm just logging somewhere else (new.brians.org).
Linguana is stalled for lack of hosting space. I'm working on options.
I wish everyone a happy new year!
I've begun "blogging" again, at new.brians.org. I wrote a small XSLT content management system for this that I'll put on CPAN soon.
I haven't died, I'm just really busy. Advanced Perl 2ed is taking most of my time, and the rest is being spent managing layoff fun at Verio. This means I haven't had time to finish the Linguana data models, or many updates to my modules.
Realistically, I probably won't make much progress until just before OSCon 2002, as I'll have at least one Linguana paper there. If you've been looking for related information, please have patience. I will get busy there next year.
I am actually going to have a Linguana site and mailing list up within a week or so. I've had some domain problems, among other things. The revised lexicon web interface is SLICK.
I'm a Christian, and have always liked this scripture, but didn't actually think about it in this forced context until tonight:
I'm not sure what it means, but it must mean something! Or not.
The Wordnet-1.7-compliant Lingua::Wordnet is on CPAN, version 0.71. Now, the next version is going to be very different, since it (hopefully) will allow format-independent storage of the lexicon. This is important, I have come to realize, for many folks desiring to use the lexicon as a multi-purpose data store. And while you can do this with the current Lingua::Wordnet, it is not as clean as it could be, say, with an RDB schema. I have a few really good RDB folks anxious to help with the design, so I'm limited in how much I can botch it up this time around. The general excitement I have seen for this and the other Linguana stuff has been really rewarding for me, and I hope it will continue.
I'm working with Jon Bjornstad (the guy who gave the awesome lightning talk @ TPC5 about the Tk system he built for his quadriplegic friend) on a TPJ article and perl.com interview about what he accomplished. I'm going to praise his efforts as much as I can without sounding like an ass.
My slides (both tutorial and talks) are at brians.org/tpc5. I'm told the tutorial slides have "problems"; I'm checking into that right now.
The conference was utterly good, and if you didn't go you missed out. Not like you didn't know that. A big round (applause, beer, what-have-you) to gnat and the ORA cohort for a well-paced, content-rich week.
gnat, somehow, talked me into doing his XML tutorial for him. He told me it would be good for me, or something along those lines.
The Wordnet 1.7-compliant Lingua::Wordnet is finished, and being tested a bit before making it to CPAN. It will arrive this week.
The Linguana web site is also supposed to launch sometime this week, but I suspect it will take until at least next week. I haven't yet decided whether hosting this project is me shooting myself in the foot. It will be a very cool OS project, nonetheless. If you have any interest in NLP and large lexicons, I recommend attending my Linguana/Disambiguation presentations at TPC. I'm very excited about it, and am eager to find contributors. This is a project accessable by programmers and non-programmers alike.
The other site that I hoped would be up by now is the "Voices of Extremity" net-opinion mining survey, which will also be done in time for the Linguana project. This is the "What Languages Suck" o-meter, but implemented as "What Sucks". It regularly ranks the most oft-insulted topics on the net, as well as the most liked. Something of a followup to my Lingua::LinkParser TPJ article, taking into account such factors as "apostrophal smoothing" to lend highly useful reliability factors into the data samples. It rocks.
Everyone's doing it, so I will to. The journal.
I released a new XML::SimpleObject this week, and hope to release the next version of Lingua::LinkParser in the next week or so.
Wordnet 1.7 is "pending", and this will correspond well to the rewrite of Lingua::Wordnet for a 1.0 release. This release is going to largely genericize the data format storage, including optional RDBM schemes. This will make it much easier to use the lexicon data as an extension to other data sets, or embed lots of new information within a record that references a synset.
My own current project using all of the above is a massively multiplayer game world system, that uses the Wordnet hierarchy to model the world objects, as well as make use of the data for NLP within the system. This pertains to my Linguana paper/talk as the July TPC.