Re: [NTLK] English dictionary with phonetic spelling / phonetic

From: Paul Guyot (pguyot_at_kallisys.net)
Date: Tue Nov 20 2001 - 01:21:24 EST


>From: Victor Rehorst <victor_at_newtontalk.net>
>
>I guess what would have to be done is to build a "DictEN" and then use
>DictMgr to use it...

DictMgr is not here to let you use custom package dictionaries (such
as DictNL, BR, EE, DK, FR, IT). These packages don't need anything to
be used.
DictMgr is just there to (a) play with ROM dictionaries (mostly
deactivate them) (b) create user dictionaries (but unlike package
dictionaries which use virtual memory, these will be loaded in C++
heap so you're very limited on MP2000s).

>From: "Markus Rasch" <m.rasch_at_gmx.net>
>
>But I would prefer an additional installed dictionary, not to
>disable the internal. So
>that my german MP will understand german AND english.

This is the default behavior. And if you only want the ROM
dictionary, just freeze the dictionary package.

DictMgr will let you have only the package dictionary (i.e. disable
the ROM dictionaries). It's *not* required to use package
dictionaries. In fact, I would suggest not to install DictMgr if you
don't want to have an improved recognition by disabling the ROM
dictionaries.

>From: Robert Benschop <rbenschop_at_mac.com>
>
>Than Paul's alt..rec came along (personally I think it's a terrible name for
>a great piece of software ;-)

Sorry, I'm definitely not good for marketing. And a lot of folks
spell it "alt.rec" or "alt..rec" or similar variations.
Indeed, alt.rec.... is part of all these localization stuff. I
wouldn't be able to really take notes on my Newton without it. The
drawback is that now that I'm able to take notes with it, I no longer
work on this software, although there is definitely room for
improvement.
I'm happy that it works for you as well, Robert.

>All I did in DictMngr was uncheck the checkbox in "commons (attributed)
>(forward)" and that did the job.

Basically how I use it for.

>Never understood how and why but somehow it's now really good at both
>English and Dutch, though it ads English words to the "Recently Written
>Words" list (nice Winnie the Poeh touch BTW, all these capitals ;-)
>
>So though I disabled the internal (English) word list, at least that's what
>it looks like to me,
>(Paul might have another opinion about this altogether ;-)

Indeed, once disabled, the English words are not considered as in a
dictionary and therefore added (by Paragraph or alt.rec....) in both
your user dictionary and the add dictionary. I could fix this (in
alt.rec.... only, though).
In fact, I thought that a good way would be to not deactivate a
dictionary but instead give it a factor so English words will be less
frequently recognized but recognized nevertheless. Well, it's just a
thought.

>So either it would probably work like this right away or you might have to
>check and uncheck the "commons (attributed) (forward)" when you switch
>languages, I'm pretty sure that somebody might be able to make a little
>script out of that too, if the need arises.

I did write such a script for DashBoard.
It's something like:
// Disable US
foreach index_dict in [0, 8, 34, 48] do
GetRoot().|DictionaryMgr:Paul's Software|:?Deactivate(index_dict);

// Enable US
foreach index_dict in [0, 8, 34, 48] do
GetRoot().|DictionaryMgr:Paul's Software|:?Activate(index_dict);

You can also disable or enable your package dictionary (by
freezing/thawing it) or DictMgr user dictionaries (with the same
method for DictionaryMgr).
Personally, I work in only two modes:
* French
* French + US
but it's because I no longer have lectures in English.

>BTW, if a word list to create your DictEN wouldn't be handy, my guess would
>be that you would eventually build your own word list using the "Recently
>Written Words", not sure about this though, Paul....?

If you have a word list (preferably on the desktop computer),
*anyone* can build such a dictionary. Docs are on my website.
Now, from a user dictionary, one could make a word list (with small
code snippet) then pack it in dictionary data format. Or you could
just directly get the data out of the user dictionary and stuff it
directly into a package.

OTOH, it's true that I should include some package-making feature to
DictMgr. I should also do so many things such as fix this heap
corruption bug in ATA Support.

>From: oliver.brose_at_t-online.de (Oliver Brose)
>
>Maybe Paul could say something about this... it would make more sense *to
>me* if one could use the dictionarys side by side, but switching them would
>be tolerable as well, if it does not mean much hassle.

It's the default behavior. If you have a German MP and you install
the Italian dictionary, then your Newton will recognize both German
and Italian words.

>From: "Markus Rasch" <m.rasch_at_gmx.net>
>
>Maybe it's possible to copy the internal worlist of an english MP
>and make a DictEN out of it!

It's totally possible. Normally, if you ask the NewtonOS to export
the data of a ROM dictionary, it will throw you away. But with the
docs I've written, you can, for example using Hammer, export any
dictionary data directly in binary form and then include it in a
package.
You can also "walk" thru a dictionary and output every word of it to
NTK's inspector.

However, don't count me in. NewtonOS English word lists are the
property of InfoSoft International, Inc.

HTH,

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 Dec 01 2001 - 20:03:18 EST