The assorted finds of Artefact Publishing
Last week I decided that I wanted to release the first version of IPA Zounds this Friday. I don't know whether I’ll manage that, but in the words of a former co-worker, “progress is being made”. In fact, there are only two major bits of functionality that are currently unimplemented, and they were scheduled for the second release anyway. The trouble really lies with the data, not the code (though the GUI could be made more pretty). I am meeting with a professor of linguistics this week to discuss phonetic binary features, and I am hoping that will resolve all remaining issues with the data the application uses to make the correct alterations to the words it is given.
Apparently, not much work has been done in this area. I find this disheartening, as I’ve mentioned before, because it’s obvious that there are sets of features which map to characters in the IPA (voicing being one). It seems to me that there are three requirements for a set of binary phonetic features suitable for mapping the IPA:
The last rule points to the grouping that lies at the heart of the system, with a feature for voicing rather than a feature for each and every character.
In related news, I am getting into the swing of writing code for all of the little jobs this application involves. The documentation incudes a feature table generated from the data the application uses, and another script checks that the first two rules presented above hold true for that data. It is nice to knock such scripts off quickly and move on to real challenges.
Posted by jamie on June 17, 2003 15:18+12:00