I'm still out. Sorry folks - I just have more on my plate right now than I can handle. I want to answer some of my e-mail - which I haven't read in like a week - and get back to Clement in the comments (who I'd love to meet in person, but just can't right now).
But in the mean time this has come to my attention. Titled Generación automática de hojas de estilo XSLT mediante Programación Genética, it looks like it was printed up in Actas del II Congreso español sobre Metaheurísticas, Algoritmos Evolutivos y bioinspirados in February 2003. What is remarkable about this bit of Spanish prose is that it discusses using genetic algorithms to build XSLT filters for document conversion, and, that it cites "el único trabajo publicado relacionado con el tema", which, as it turns out, was written by me.
It was not my proudest work. I did it for the grade in John Koza's genetic algorithms class at Stanford. I knew that it would, like all the other class papers, be bound and indexed, and I assumed it would spend the rest of its days resisting parasites on some rarely visited shelf in the Stanford Engineering library. Basically, since I was doing XSLT at Sun at that particular moment, I thought, hey, why not combine my job with my class? So, I did this project where I tried to use genetic algorithms to inductively build an XSLT filter that met certain specifications. It worked, kinda, for a simple problem, but it was hardly an efficient solution to designing document conversion software.
I figured I was the only guy out there nuts enough to even try such a dumbass thing, so I never figured there'd be any follow up. I was wrong.
Alas, my hispanophone colleagues have come to that same conclusion I did:
Algo tan aparentemente sencillo como lo que se ha planteado aquí, la extracción de contenido de un documento XML, ha supuesto un esfuerzo considerable, y aún así los resultados obtenidos no han sido todo lo satisfactorios que hubiéramos deseado.
Even something so apparently simple as what we considered here, extracting the contents of an XML document, has taken considerable effort and still the results we have obtained have not been as satisfactory as we would have wished.
Voila: my first cited contribution to the body of human technical knowledge. Someday, I'll be able to write about what I'm doing now, and hopefully get cited for writing something that doesn't cause me mild embarassment.Posted 2004/12/07 17:17 (Tue) | TrackBack