[lojban-beginners] Re: learning lojban for computer science?

[Michael van der Gulik <mikevdg@gulik.co.nz>]

ethan wrote:

> Hi-
>     On the main lojban site the software section doesn't seem to have 
> extensive software. I was under the impression that lojban and loglan 
> were partially created with an eye towards human computer interaction. 
> Is this true? If so, is there a reason there is not more software? Is 
> there some fundamental problem that prevents this from happening? Is 
> it worth learning lojban if my main interest is computer science?
Hi Ethan.

Lojban has a YACC grammer 
(http://www.lojban.org/en/publications/reference_grammar/chapter21.html). 
There are parsers out there to parse and syntax-check lojban.

Lojban is based on predicate logic, just like Prolog but a tad more 
complex and complete. Hypothetically, it would be possible for a 
computer to understand the semantics behind a lojban sentence and draw 
conclusions from it. There are some very very useful words/structures in 
lojban: "ko", "mo", "ma" and "xu". These would cue the computer to 
actually do something - e.g. "ko" is an imperitive, making that sentence 
a command.

I'm interested in Lojban because of those two facts. Prolog uses a 
simple backtracking algorithm to search and draw conclusions. By using 
Lojban instead of Prolog (the raw Lojban would need to be normalized, 
perhaps to horn clauses), and then improving the search algorithm to use 
perhaps an evolutionary algorithm combined with a pattern-matching 
neural network, I could create an artificial intelligence that would be 
able to evolve and learn continuously.

However, that's a long way off.

Mikevdg.