Spell checker problems

Author

Message

Jim Hawkins

Posted: 01/31/2005 21:45:48

Sorry if this is a long post :-(

I've tried various spell checker interfaces, and had some problems with all

of them. There are plainly problems with Unicode - unfortunately - my classroom

software editors are Unicode because they have to handle everything from

Chinese to Arabic to French to English. There don't seem to be any Unicode

spell checkers as such - but ASpell does imply that it can do it.

But there seems to be a problem with Aspell and LSSpell when it comes to

accented characters and code pages even the editor is NOT in Unicode mode.

This is that the current codepage and font don't seem to get through correctly.

Assuming that we're NOT in Unicode mode, when we call Check() the codepage

and font in the current item from which the word being checked is derived

should be available to the spell checker and the spell checker dialog form.

With Aspell, a correct word which contains a diacritical often returns a

match failure, followed by suggesting exactly the same word. If the word

ends with an accented letter usually the word is presented as wrong with

the accented letter missing. Is this known behaviour, or am I being stupid

(always possible!)?

Incidentally, the ASpell Delphi demo throws an exception early in the actual

dictionary iteration process, for no reason I can quite debug yet, whereas

your interface does work within the above limits.

Maybe it's time for an open-source Unicode spell checker project!

Jim