Every couple weeks or so, I get up the courage to fight with the ispell
code some more to fix this. Unfortunately, I usually leave it more
confused and frustrated than before.
As I've said before, it seems to have accented characters stored as 'X, ie
ñ is 'N. There seems to be some kind of mapping table in the dictionary
file, but might not be loaded in our version. However, I tried building my
own mapping table, so that the incoming words matched the stored words
exactly, and it still failed to recognize the word. There, it was hitting
the wrong hash bucket. Exploration since then has been unproductive. I'm
pretty sure it's just toying with me at this point ;)
I imagine that I'll give it another shot over Thanksgiving.
Justin