Soundex [Re: [DGD] LPC algorithm question]

[Noah Gibbs]

On Sun, 24 Feb 2002, Christopher Allen wrote:
> I'd completely forgot about the Soundex algorithm, though when I was
> using it 15 years ago or so I was told that it only works well for
> English, not other languages. Do you have something different?

  Well, it's pretty specific.  It just skips vowels and groups consonants
that sound similar (M sounds like N, D sounds like T).  It works
surprisingly well.

  I currently just don't do any phonetic matching for other
languages.  Spanish, though, is pretty easy.  Soundex skips "y" already so
the n with the tilde over it (sorry, not using intl keyboard) is the same
as n.  H is already skipped so "ch" is grouped with "c", which works
decently -- not out of line with Soundex's existing accuracy.  You might
well want to make the double-l be skipped rather than grouping with l,
which could be done in a preprocessing step.
  I don't have any locales other than Spanish with any real progress
going.  I just stuck a simple structure in place for phonetic matching
figuring I'd work it out when/if I came up with good phonetic matching for
other languages.  It could be different for, say, Navajo.  Luckily (?) I
don't know Navajo :-)

> Sounds like a great idea for making help searching easier -- if sometime
> you'd like to share it, I'd consider using it ;-)

  My MUDLib's already available at
"http://www.angelbob.com/projects/mudlib".  The soundex stuff is in both
"/usr/common/sys/soundexd.c" and "/usr/common/sys/helpd.c", plus a little
in "/usr/System/lib/wiztoollib" for the actual help command.
  The help system's pretty well developed, and Soundex works great.  The
Spanish localization is the closest to being there, but it's still far
from done.  Feel free to comment!

<shameless plug>
  And when version 0.004 is released, I'll have portables working.  Look
for it later this evening.
</shameless plug>

-- 
angelbob at monkeyspeak.com
See my page of DGD documentation at
"http://www.angelbob.com/projects/DGD_Page.html"
If you post to the DGD list, you may see yourself there!

_________________________________________________________________
List config page:  http://list.imaginary.com/mailman/listinfo/dgd