En/na Simon Brouwer ha escrit:
Hi Joan, all,
Joan Moratinos schreef:
Hello,
I'm Joan Moratinos, author of the Catalan dictionary for
myspell/hunspell.
The Catalan language has three main dialects, roughly corresponding to
the territories where it's spoken (Catalonia, Valencia and Balearic
Islands). There are some variations in morphology, and consequently in
ortography. If we put all these variations together, the overall
quality of the speller decreases (because each user wants the
"alternate" forms be marked as "incorrect"). We don't want to do full
different versions of the dictionary nor the rules, as the main part
is common.
So, we are considering a system where more than a dictionary (and
perhaps a file of rules) can be loaded. It would be useful for
languages having substandards (probably every language), and also for
adding speciality jargon. A simple approach could be allowing more
than a line for each language, or more than a dictionary name in each
line, etc.
Do you think it would be possible and useful?
Do you mean separate dictionaries for variants of a language? I suppose
the ISO language name should be extended if there is no separate name
for each variant.
No. I was simply thinking in cumulative lists. The overall result should
have to be the same as having a big list. The advantage would be that
each user could easily include/exclude a full set of words (as in
speciality languages). Also the maintenance would be simplified. Think
for example about a chemist maintaining a list of terms in her field
without worrying about other lists.
Or do you mean that a base dictionary could be loaded (containing the
words common to all variants) and an additional dictionary with
variant-specific words?
The latter is already supported (to a degree). For example, you can
install both the main Dutch dictionary and the Dutch medical dictionary,
and then Dutch text will be checked against the combination of the words
in each of them. This does not work perfectly however; since Myspell
loads these dictionaries as separate objects, suggestions are generated
separately, and lower-quality suggestions will result from the smaller
dictionary. A better (but more complicated) solution would be to
generate a single spelling object from multiple dictionaries.
I have also considered this "more complicated" solution. I think it
would have the advantage of being more portable between spelling engines
(for example, in Catalan, the same lists can be used for OpenOffice,
Abiword, ispell, etc.). Most engines expect a simple list, that could be
compiled from several dictionaries following the user preferences.
Joan Moratinos
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
