Re: [NTLK] Swedish on a US UMP2k

From: Paul Guyot (pguyot_at_kallisys.net)
Date: Sun Feb 03 2002 - 01:11:41 EST


Date: Sat, 02 Feb 2002 23:17:46 +0100
From: Robert Benschop <rbenschop_at_mac.com>

> > Does it provide all characters I'd need (assuming this
> > language has some special characters as most other languages do)?

I don't know Swedish neither, but I'd say it does (Rosetta at least)
as there is a Swedish locale and even a Swedish lexical order.

> > If it doesn't, is there something I could do about it? Is there
> > maybe even a (freeware) dictionary?
>
>Now the only dictionary builder that I know is Paul (Guyot) and he has
>French, Italian, Dutch, Finnish, Estonian, Danish and Brazilian Portuguese
>on his site, so sorry, doesn't look that there's a Swedish one.

Two other Newton users (I think unlike Frank that they actually speak
Swedish) asked me for a Swedish dictionary.
Let me take the occasion to remind what these dictionaries are and
how one can create or make me create one.

These dictionaries are word lists for the HWR (both recognizers) and
for the spell checker (and any other software that uses
dictionaries). (however, the spell checker will still do US guesses
and this will lead to very poor results).

Compared to other existing software (and what Apple described in the
Q&As), the technology used allows you to have the data compressed on
the Newton store (any store) and only loaded when required with the
NewtonOS virtual memory mechanism, hence these dictionaries don't
take much space in RAM or in storage and can be big.

The full docs about how this works has been published within the
Bowels project.
http://www.kallisys.com/newton/bowels/

If ever you want me to build a dictionary for you because all that is
Chinese in spite of the effort I made to make it clear (or you don't
know what "debugger" means, or where NTK can be downloaded), I might
build a dictionary for you.

For that you need to follow these rules:
a/ find a word list that can be used in such a dictionary without any
restriction but the fact that the dictionary is released for free and
anyone can copy/redistribute it including selling it on UNNA CDs.
This especially means that GPL'd word list won't work, so iSpell
lists is probably not a good choice. Some dictionaries are based on
such a list, but submitters got a special authorization from the list
authors.
b/ send me the list in a simple format: one word per line. I prefer
MacRoman encoding and MacOS (not OpenSTEP or mixed, please) new
lines. If you don't know what all that means, well, send me the list
and I'll ask you further questions to be sure of the format.
c/ don't bother me until I build the list. Building a dictionary
requires doing nothing else with the Newton and the mac for several
hours, so I'm not going to do that as soon as I get the list but
instead at a moment when it won't bother me. Additionally, it's not
very exciting.

Paul

-- 
Home page: http://www.kallisys.com/
Newton-powered WebServer: http://newt.dyndns.org:8080/

-- This is the Newtontalk mailinglist - http://www.newtontalk.net To unsubscribe or manage: visit the above link or mailto:newtontalk-request_at_newtontalk.net?Subject=unsubscribe



This archive was generated by hypermail 2.1.2 : Sat Mar 02 2002 - 10:01:54 EST