[2816] in tlhIngan-Hol
Re: Word processing
daemon@ATHENA.MIT.EDU (daemon@ATHENA.MIT.EDU)
Wed Jan 26 17:42:10 1994
Reply-To: "Klingon Language List" <tlhIngan-Hol@klingon.East.Sun.COM>
From: Will Martin <whm2m@uva.pcmail.virginia.edu>
To: "Klingon Language List" <tlhIngan-Hol@klingon.East.Sun.COM>
Date: Wed, 26 Jan 94 17:29:55 EST
On Jan 26, 4:52pm, "John C. Kirk" wrote:
> Subject: Re: Word processing
...
> OK, I've done some digging around in DOS, and I've found that any custom
> dictionaries are stored in the directory windows\msapps\proof.
Now, this is true, though if people have upgraded from earlier versions
of Word, they may have a separate set of custom dictionaries in the \winword
subdirectory, since Word once used its own spell checker and now it uses the
one shared among all Microsoft Office apps. Since newer main dictionaries
contain words not in older ones, it is a good idea to read your old custom
dictionaries as ASCII files and spell check them into new custom
dictionaries, instead of just using the old custom dictionaries. (little tip)
> They are
> straight ASCII files, and so they can be edited in the normal way, which
> should be a lot faster than using the ADD feature via the spell checker.
That's not the problem. The problem is that with affixes, the
less-than-two-thousand Klingon words in TKD becomes potentially millions of
meaningful word-affix combinations that your spell checker cannot split up
and check the way a parser can.
> When I was using a demo for languages, it said something about making
> sure your .lex files were correct. I found 4 of these, 2 in the windows
> tree, and 2 in the winword tree. These were all in gibberish, buit one of
> them had some English words at the end of it. I don't know where words
> are stored for the standard dictionaries of the languages it supports,
> but maybe they're coded in there.
They are DEFINITELY coded in there. It's not just to keep us out. They
HAVE to code them in order to compress them for faster reads and lookups. If
they didn't, it would take FOREVER to spell check even a small letter. My
understanding is that most dictionaries of this type use an algorithm that
states a root word, then encodes the differences between it and the next
alphabetic entry without repeating the root word. The next code builds on the
already altered word to describe a third word with a minimum of characters.
The whole scheme depends upon a LOT of preprocessing of the original list. In
short, you can't do it yourself.
They don't have to do it for your custom dictionaries because they
rarely accumulate more than a couple hundred words max (instead of the
100,000+ entries in the main dictionaries).
> I suspect that you need to modify the
> .lex files, to add Klingon to the list of recognised languages.
> Unfortunately, I have no idea how to do this! Any ideas, anyone?
I'm very nearly positive that you can't build a primary .lex file. If
Word 6.0 allows us to define a range as "language X" WITHOUT a primary
dictionary, then it might let us use one of the plain ASCII custom
dictionaries with it instead. It's a small likelihood, but not unreasonable,
given Microsoft's tendency to listen to user complaints. Hey, they added
kerning, didn't they? I was bummed when I hand kerned every conceivable
character pair for Dr. Schoen's Klingon font AND THEN FOUND OUT THAT WORD 2.0
DID NOT SUPPORT KERNING INFORMATION IN THE FONT!
> John
charghwI'
