NTLK Re: alt.rec... and dictionaries

From: Paul Guyot (pguyot@pnm-consulting.com)
Date: Sun Dec 12 1999 - 07:05:40 EST


Dakkar <ceccarel@cibs.sns.it> wrote:

>Hi all!
>I've just downloaded and installed alt.rec... and it seems a great
>package... was it not that I write in Italian! So the recognition
>actually is a bit worse than before, and I think I will not be able to put
>an entire Italian dictionary into the 'new word'-list.
>Since there seems to be (on P&M site) a French dictionary (and a French
>locale?), would it be possible to create an Italian one? Paul?

Sure.
Two things about dictionaries:
Usual way to have custom dictionary does not allow it to be used but in
custom views (why they did that, I ignore it).
To make it available for recognizers & spell checker, DictMgr does register
them differently. To make system dictionary unavailable (disabling them) it
sets the domain to 0.
The problem is that the dictionary data is loaded into heap (not NS heap,
so I haven't seen it first). And DictFr is too big (c. 180 kb) for MP2000
heap, but DictNL (which also uses our technology) is fine (it is 80-90 kb).
There are 27000+ words in DictFr, and I can't remember how many they are in
DictNL. There are 80000+ words in the system US dict.

This problem with system heap is probably the reason why the user dict's
size is limited to 1000. This can be changed with DictMgr.

The advantage of switching from DictMgr to a separate package is that then
the dict is compressed a little bit.

The technology is not completely stable yet.

>THe Newton manuals seem a bit obscure on the matter of user-defined
>dictionaries. Is it so difficult, or it's just knowing the right things?
>(I'm not frightened of NewtonScript, or of the task of making a
>dictionary, just don't know where to start!)
My idea is to change the technology from stuffs I learnt by hacking. My
idea is to have the dict data to remain in storage (i.e. by using virtual
memory). Sure, it could increase the access time, but it will allow huge
dicts, as long as you have enough storage, of course.

>Also, a locale would be great. But on this the manuals are a bit
>clearer...

We also do locales. We just need to update our technology, and I haven't
found time for that (hey, I cannot develop everything at once). If you
provide proper information, I could release it.

Regards,

Paul

----
P&M Consulting Newton Program
http://www.pnm-consulting.com/newton/

****************************************** This month's NewtonTalk brought to you by:

EVOTE.COM, the ultimate Political Junkie site on the 'Net.

The Clinton Administration, the George Bush 2000 Campaign, and almost every other major U.S. politician has said something nasty about us at some time. Find out why at: http://www.evote.com

******************************************

Need Subscribe/Unsubscribe info? Visit http://www.planetnewton.com



This archive was generated by hypermail 2b29 : Tue Feb 01 2000 - 00:00:49 EST