[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: PATCH: install hooks



> How does this Finnish dictionary and hyphenator differ from
> the one on the OpenOffice.org dictionary download site whose README states:

It's the same program. The reason I'm making a Debian package for Soikko, is 
that it requires a bit complicated installation which involves regcomp and 
has to be repeated every time openoffice.org-bin is upgraded.

> I assume the Soikko is a closed source spellchecker  (I saw no source for
> the Soikko spellchecker anyplace), ... is it?

Unfortunately so.

> If different, is the source to your spellchecker freely available under an
> opensource license and if so how does it compare to this one?  I would
> rather have a fully open source spellchecker available on the
> OpenOffice.org site instead of an open interface to a "closed" one.

Me too. 
I actually previously planned to develop one by myself. However, grammar rules 
of the Finnish language are rather complex.

They contain a lot of postfixes, word transformation for different purposes 
and, most notably, allow joining words that belong to same logical group. 
E.g. apple=omena, piirakka=pie, and thus apple pie = omenapiirakka. If the 
pie happens to be mine, I would say omenapiirakkani (my apple pie). As you 
can see, this basically renders simple dictionary schemes useless:

 katsoa = to look
 katsoin = I looked
 katsotutittekokin? = did you(plural) also make someone else look?

 You can pretty easily come up with monster words like:

   Epäjärjestelmällistyttämättömyydellisyydellänsäkäänköhän? =
     Not even with his or her own ability or property of not making other
     people to make things unorganized, I wonder? (approximately)

While hyphenation rules are simple in principle, alas, joined words break them 
by requiring that they should be hyphenated separately => you need a 
dictionary and morphological analysator even to make a hyphenator.

Pasi has studied this mess in a university and is working for some company 
that does such stuff; I can understand his decision to keep the source closed 
for now.

It doesn't mean, however, that I wouldn't like to try it someday. I happen to 
be in the same university as Pasi does, so maybe I could attend to the 
lectures on the subject as he did.. ;)

- Jarno



Reply to: