shit nobody but me cares about
Feb. 15th, 2004 01:26 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
maradyddMy reannotator, what there is of it so far, works. At the moment, that's just the bit that reads the reannotation information into an STL map (and a very complicated map it is; typedefs are my New Best Friend), but hey, it's crucial, and it's functional. I also grabbed myself a stemmer -- in case anyone else needs one, it is worth noting that PC-KIMMO is a morphological-analysis C library in its own right, but the site's weirdly silent on licensing information -- as I anticipate I'll need one when I start learning subcat frames off the Penn Treebank. (That's not the definition of "learning" where the knowledge learned goes into my brain, by the way.)
Tiny bites at a time seem to be the way to go on this thing. Next step: write the bit that reads in Brown-tagged text, splits it into word-tag pairs, and does the first reannotation pass off the map. Then: write the bit that disambiguates tags like TO and the being/having verbs. Then: work out how to automatically build a frequency table for subcategorisation frames that appear in Penn.
And then the real fun begins.
But now I have to go decorate my stairs.
On a wholly unrelated note, I should have found those Harman/Kardon speakers I wasn't using and ziptied them to my canopy bed long ago.
Tiny bites at a time seem to be the way to go on this thing. Next step: write the bit that reads in Brown-tagged text, splits it into word-tag pairs, and does the first reannotation pass off the map. Then: write the bit that disambiguates tags like TO and the being/having verbs. Then: work out how to automatically build a frequency table for subcategorisation frames that appear in Penn.
And then the real fun begins.
But now I have to go decorate my stairs.
On a wholly unrelated note, I should have found those Harman/Kardon speakers I wasn't using and ziptied them to my canopy bed long ago.
