The major reason for their failings is the presupposition that our
division and definition of words is correct. Current translating
programs are largely literally based and for this reason they produce
such poor translations.
I agree that the project is complex but I have full faith in its possibility.
The very fact that children can learn a language without having to depend on
translations shows that the algorithms exist and the program will just as much
a pyscholgical one as a linguistic one.
The secret is in finding the algorithms which I believe lie more the minds of
children than in the prowess of 'scholars'.
Archive powered by MHonArc 2.6.24.