From: John Joyce <dangerwillrobinsondanger@...>
Date: 2007-07-13T09:08:07+09:00
Subject: Re: language transliteration help..


On Jul 12, 2007, at 6:40 PM, alex_land wrote:

> On Jul 12, 5:11 pm, Aaron Smith <beingthexempl...@gmail.com> wrote:
>> What is a good place to read about language transliteration? I have a
>> piece of software that I want to provide language transliteration
>> options.. Is there something in Ruby to do this.. (probnot). Just
>> looking for a push in the right direction.. thanks all
>>
>> --
>> Posted viahttp://www.ruby-forum.com/.
>
> Hi Aaron:
>
> Here are a few sources (i can't say i've looked at all Sites of this
> type, but the one's i'm giving you, i have looked at and used):
>
> 1.  ForeignWord (http://www.foreignword.com/default_old.htm) list and
> discusses many translation Apps (i'm using that term to include both
> translation and transliteration, as they do, until they begin
> discussing particular Apps);
>
> 2.  Not sure if you wanted actual translit tables for particular
> langauges, but if you do, then i would suggest the Princeton Library
> (what i'm giving you the link to is entirely public access)-e.g.,
> here's an English-Azerbaijani translit table:
> http://library.princeton.edu/departments/tsd/katmandu/sgman/traze.html
>
> 3.  Lastly, i wrote some translit scripts about two years ago (for
> about two dozen languages, Romance languages, Germanic, and Slavic);
> they are in C++ (absolutely no reason i did not code in Ruby other
> than i didn't know it well enough at the time); I put them in one of
> the major C++ repositories.  If you have SVN and want these, let me
> know.
>
> HTH,
>
> -alex
>
>
>
>
Hey this should be relatively painless, depending on the languages  
involved. There are many ways to describe phonetics and phonemes.  
Most languages view themselves in their own way of interpreting  
phonemes. So you should aim for transliterating for the target  
audience's language's phonetics.
Some languages have writing that is highly non-phonetic. English is a  
great example of a terrible system of writing. As native speakers, we  
believe it is phonetic writing, and it arguably is, but it's a mess  
of contradictory patterns of symbols that represent multiple phonemes  
and multiple patterns for the same phonemes.
Japanese has highly regular phonemic structure, but its own  
Romanization system(systems) which are close to most of the  
continental European languages that use Roman Alphabetic characters,  
but is not consistent with English. (nothing is)
This is why linguists have a fairly well-established set of formal  
academic phonemic characters that don't really make much sense to non- 
academic linguistics people.
English alone has many phonetic description systems in use in various  
dictionaries.
It is always a great project to do something like this. So, I do  
encourage you to develop it, but might find more complexity than you  
expect.
You might find unicode.org to have some useful insights into some of  
this. You might also want to look into unicode for this purpose.  
Their recent GIANT book Unicode 5, is an excellent resource and a  
fascinating book, but it is really BIG and not cheap. That said,  
almost any university with a computer science or linguistics program  
should have at least one copy of it in their library!

John Joyce