Re: [NTLK] Newton Spanish text recognition.

From: Paul Guyot (pguyot_at_kallisys.net)
Date: Wed Sep 22 2004 - 00:08:41 PDT


On Tue, 21 Sep 2004, Larry Yaeger wrote:

> At 12:33 AM +0200 9/22/04, Roman Pixell wrote:
> >larry, is 70k really a good choice considered the small amount of heap?
>
> That may be a reasonable concern, since the new dictionary obviously
> is not in the ROMs. I was assuming the word list would be compressed
> into the standard dictionary format, which would reduce its
> footprint, but heap space might still be an issue. And I'm now
> wondering how/if the word list actually gets converted to the
> standard dictionary format; I don't remember that tool being built
> into the system (though that doesn't mean it wasn't).

The documented way to build custom dictionaries suck not NS Heap (aka
Heap) but C++ heap and you'd better not have a very big word list.

However, I figured out how to build a package with a 'dict' part. Such
packages do not suck any memory (actually, they do require a very small
amount of memory which does not depend on their size).

I did document the format and how to build dictionaries here:
http://www.kallisys.com/files/newton/Bowels/Dictionary_Parts.txt

I have built quite a large amount of dictionaries, but I prefer no longer
doing so because it's a lengthy process blocking both a Mac and a Newton
(plus it requires Classic and I currently don't have my own PowerBook with
its Classic setup).

Paul

-- 
This is the NewtonTalk list - http://www.newtontalk.net/ for all inquiries
Official Newton FAQ: http://www.chuma.org/newton/faq/
WikiWikiNewt for all kinds of articles: http://tools.unna.org/wikiwikinewt/


This archive was generated by hypermail 2.1.5 : Wed Sep 22 2004 - 10:00:02 PDT