[Aspell-user] aspell-import unruly behavior

[Stephen Harris]

Hello,

I first searched the archives and the only instances of
"aspell-import" mentioned are under making Aspell,
so I guess it is not a typical problem.

I 've read the documentation and it seems sketchy to
me, perhaps my syntax is off. I 'm using Windows XP
and have the new 1.3.6.17 native windows lyx and an
older Cygwin LyX version and have tested on both.

I wanted to make a combined English/French wordlist,
give it a personal dictionary filename, use aspell-import
to make it a dictionary (.rws?) and use it as the default
spell-checking dictionary for LyX.

I read how this was automatically done with a list of
English and technical/anacronymic words. It seemed
to me the method should work for English/French
words, though I read something about only Ispell
using the extended Ascii character set. However, I am
able to use both an English or a French dictionary
from the cmd line and it will correct francais with the
correct mark under the c in Francais.

The url, http://www.uvsoftware.ca/wordjobs.htm#1H0
explains how to extract a list of words from an already
compiled/binary dictionary in order to make an updated
ordinary/technical wordlist-->dictionary.

Only one of the utilities needed, uvhd, is free. But I'm
starting with the wordlist from Moby, plus a French
wordlist I found on the marketplace, merging them in
Xemacs, then saving the file as .aspell.en.pws, which I
think is the format for a personal dictionary, and finally
converting/aspell-import to Aspell en_us.rws dictionary.

I have Perl and Aspell in my Path. So from within the
C:\Aspell\dict directory, I tried various combinations
to produce an .rws file:  perl aspell-import .aspell.en.pws
I tried various combinations, letting the output file be a
default, specifying the filename output, using a > redirect
and so on like typing full pathnames to the executables.
I also tried various combination from a Cygwin prompt.
Also tried add-extra-dict = fr* in en.multi which failed.

I included information about the lang and num at the top
of the combined wordlist file created in Xemacs since a
doc mentioned Header, but didn't understand this really.

There seems to be a hole in my theory and a gap in my
practice. :-) In my research I came across a multilingual
screenshot which I found interesting. Spell-checking all
those languages from keyboard input, an AI challenge!
Imagine the complexity of the rules to recurse that doc.
http://www.linguatech.co.uk/multi.htm

À la prochaine,
Stephen