Universal Networking Language

[A. Vine]

Grant Barrett wrote:
>
> On Thursday, October 14, 1999, A. Vine <avine at ENG.SUN.COM>
> wrote:
> >
> >And history repeats itself.  Again.
>
> Do you mean Esperanto? Or do you mean we've seen this article
> here before?
>

No, I mean machine translation.  I mean machine translation based on a set of
internal primitives with generation engines for different languages.  I took a
couple of courses from a professor in college who had been working feverishly on
this sort of thing for years (Roger Schank, for those of you who know).  He was
in the discipline of Artificial Intelligence, coming from psychology.  This was
about 20 years ago.  And machine translation has been worked on since the 50's.

The problem is understanding natural language and converting it into primitives
in the first place.  In order to make a sensible machine translation, you have
to start out with a structure initial language.  This is fine for controlled
environments, but it constrains the folks producing the information, and it
won't work in a general context.  There's an English version of the structured
language concept called ???? English.

The point being that this is not a new concept.  I hope at least these folks
will look at what didn't work in the past.

The alchemy of data processing is natural language processing.

Andrea