Re: Internationalizing Linux

Konrad Rosenbaum (htw6966@htw-dresden.de)
Mon, 14 Dec 1998 13:25:00 +0100 (NFT)


Hi all,

On 14 Dec 1998, Matthias Urlichs wrote:

> Horst von Brand <vonbrand@sleipnir.valparaiso.cl> writes:
> > Whatfor? The de facto "Universal Network Language" (on the Internet, at
> > least) is English, and nobody will be able to change that for quite some
> > time to come. English has several virtues over artificial languages:
> >
> It has also the political disadvantage that some people have it as their
> _first_ language while others have not.
>
> Thus the development of Yet Another Language. If I would be doing it (which
> I'm not) I'd use Loglan/Lojban, at least that one is a language which is
> designed to be (a) computer parseable, (b) less suitable to doublespeak
> than English...
[cut]
> > - English has a rather simple grammar (spelling is quite another matter :-)
>
> English is rather difficult to parse without hard AI, its free-form grammar
> lends itself to some rather creative uses. Granted that those are more
> common in SF novels (I particularly like Douglas Adams when it comes to
> creative usage of English ;-) than in UN documents...

Ok, did I misunderstand all what I've learned about hackers? We are
creative people - aren't we? Creative humans prefer creative languages. I
guess if you would chose a formal language in a hacker-list you would see
this language turning around into a creative one - just by people adding
other meanings of words, extending some words, adding totally new words
beeing creative with grammar.

natural languages grew up with mankind - they are totally perfect for
human communication. IMHO a computer should _not_ understand human
language fully - A computer is a tool, not a partner!

By the way: human language is context sensitive. If you programmed the
full context into the computer and then let it parse the text it will
find the same meaning as another person (except you made the error that
the computer believes that every sentence makes sense). It's just like
patching unclean kernel-sources - it may succeed but it may scramble the
sources as well (these are the situations human beeings request more
info). The bad thing with them: if you have not a 100% copy of the
context the text was formulated in, you have the risk to misunderstand,
to make it worse: if you have this copy you don't need to transfer
information.

So the main difference between human and computer languages is this:
- computer language is non-ambigous and can be used to formulate commands
which don't need much interpretation or even guessing
- human languages are:
* interesting
* extensible
* creative enough to design new "microworlds" (=new technology,
literature etc.)
=> human language allows different views on the sentence

So, if you want to come closer to perfect descriptions: just write longer
and clearer designed sentences, this way you reduce the amount of
possible views. The "network" of persons who populate INet is a
(prestate of a ) new society - with all kinds of communication you can
think of (to name few: scientific drafts, smalltalk and even
love-letters). Please let they use a natural language which allows all
those kinds of communication.
I personally DON'T want to speak a diffent language for every different
task to describe (two natural and 15 computer languages are enough for
me).

happy hacking,
Konrad

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/