Is there a way to separate unicode letters from punctuation and the like? The reason is simple: I would like to separate words in a text for a spell-checker.
I see there are tables which list unicode categories http://www.sql-und-xml.de/unicode-database/#kategorien Is there already something for freepascal to get such information? Is there a better way to do what I need?

Thanks
Theo

In hebrew (at least) the punctuation is a different char that comes after the letter, but painted like it was part of the letter, so you can parse each word and ignore non letter value (it arrives in different range in the unicode table).

Ido

Yes, the question is: which are non letter values in the upper unicode ranges?
I've now written some code myself to do this and separate the words. If you're interested, I can post a download link here.

Thank you
Theo

Yes, I've done this today. http://www.theo.ch/lazarus/woprsplitwords.pas It seems to work. Do you see any problems?

Thanks
Theo

theo wrote

Is there a better way to do what I need?

I think that writing such a routine which compares the char to a table will be the best solution.