From: "Hal E. Fulton" <hal9000@...>
Date: 2003-07-02T08:22:39+09:00
Subject: Re: Speaking of I18N...

----- Original Message ----- 
From: "Chad Fowler" <chadfowler@chadfowler.com>
To: "ruby-talk ML" <ruby-talk@ruby-lang.org>
Sent: Tuesday, July 01, 2003 5:53 PM
Subject: Re: Speaking of I18N...


> I did a screen-scraping one some time back.  It was truly a 
> several-minute hack.  I put it in RAA under the "Jokes" category, but I 
> don't think it works anymore (since I depended on the HTML output of 
> babelfish for it to work).  I should probably either fix it or remove 
> it, though I guess being under the Joke category, it's not going to have 
> a huge impact on RAA's credibility. :)
> 
> I also had a method called "stupidize", which would translate text to 
> and from a language, allowing it to suffer the often hilarious semantic 
> distortion that you get from Babelfish.

That hilarious semantic distortion is also interesting
in a theoretical sense (to me, at least). I find myself
sometimes wondering if there might be a way to measure
the drift in meaning -- I tend to think that a one-way
translation is probably only "half as garbled" (whatever
that means!) as a two-way one. 

But if that's true, it implies there must be some way
to quantify it -- or else my thinking is just nonsense.
And then that raises issues like: If I translate something
from English to German and then back again, is more meaning
lost on the first leg of that trip or the second leg? Is it
different translating lang X to lang Y for different values
of X and Y? It would have to be, I think.

But I've wandered offtopic, as I do so well.

I'll check out the SOAP4R thing.

Hal

--
Hal Fulton
hal9000@hypermetrics.com