Re: OT: CXS chart and machine-readable Unicode->CXS mappings

I know little about phonology and IPA, but after I had
a look at
http://cassowary.free.fr/Linguistics/cxschart.png
I came up to a rather simple idea. Let's consider only
the Consonants (Pulmonic) to begin with:

1/ There is an array with 9 (or 11, depending how you
consider it) columns, and 8 rows

2/ Each cell may contain 0, 1, or 2 values

3/ The codes available on every browser, email server,
text editor or whatever are at least:

ABCDEFGHIJKLMNOPQRSTUVWXYZ
abcdefghijklmnopqrstuvwxyz
0123456789
plus some general-purpose signs like: / # &, etc.

So you associate a letter to each row, and a letter to
each column. You use mnemonic codes as far as
possible.
Ex: column 1 = Bilabial, let's call it B. If confusion
is possible, use B # b, if there are more than 2
possibilities, use another, non-mnenonic sign.
Ex: row 7 = Approximant, let's call it A.

So to code a phoneme, we decide that we use the column
letter, followed by the row letter, followed by a
number (1-9 is more than needed).
Ex: Retroflex Fricative, 2nd possibility = RF2 (it
looks like a z with some strange ornament; I have no
idea how it should be pronounced, but I'm sure that if
I code it RF2, everybody can translate it into his own
system).
Ex: Alveolar Trill = AT1 (because only one possibility
mentioned)

If somebody decides that there should be a 2nd
Alveolar Trill, because he found an old man in
Papuasia New Guinea who used it, well it will be AT2,
and so on.

We can do the same for vowels.

So we can use an array containing up to 62 (26+26+10,
if only letters and numbers) columns x 62 rows, each
cell containing up to 10 possibilities. This I guess
is more than we need.

If we have to add diacritics, we use a 4th, and 5th
code if necessary.

Groups of codes (each group being one phoneme) will be
separated by blanks.

It may look a little long, but it will be very much
clearer to everybody (especially if using mnemonic
letters), and will be universal, everybody being able
to read, write and transmit the codes used. And any
computer program will be able to recode it any
customized way.

Good ! No more problem. I just invented the Esperanto
for Phonetics. Now I'll try to fix my bath tap, which
is much more hazardous.


--- Henrik Theiling <theiling@...> wrote:
> Hi!
>
> Mark wrote:
> > Cool idea, and thanks, but why not supply an XML
> version of the
> > mapping, with associated DTD or Schema? Then it
> would be
> > machine-readable by software written in any
> language whatsoever,
> > without you having to write more code.
>
> Well, I would not say so.  You need an XML library
> for this to be true
> in that specific language.  I've never used one (I
> never understood
> the XML hype and the point about XML -- my general
> syntax for data was
> always Lisp).
>
> If I succeed in using an XML library for C++ on the
> machine I write
> this on (I don't see why not :-)), I'll provide the
> data in XML -- the
> syntax as such is easy, but I will need the check
> it, of course.
>
> Bye,
>   Henrik

=====
Philippe Caquant

"Le langage est source de malentendus."
(Antoine de Saint-Exupery)

__________________________________
Do you Yahoo!?
Yahoo! Search - Find what you’re looking for faster
http://search.yahoo.com