Re: CHAT: programming langs

Boudewijn Rempt wrote:

> I don't want speech recognition! I want handwriting recognition,
> in any script!
That's a much better idea for a quieter office environment.
Handwriting OCR is more difficult than printed, or was when
I used to be involved in it. But if a training period is
allowed, it should be possible. Hmm. In 1980 I could easily
sign my name on a screen using a simple "joystick", but I've
never been able using a mouse. Now I'm using a "glidepoint";
all the laptops seem to use them now; I can *almost* do a
decent signature with it.

Whether input is from mouse or keyboard or speech or email,
it must be text at some point. I used to do GUIs as emitters
of text to an underlying text parser, so command files and
"macros" and rational testing were easy to do. That way,
there are many possible front-ends channeled into a single
uniform input stream. GUI, keyboard, speech, whatever:
nobody gets excluded, all get to the same functions.

But then what is the underlying, ultimate generic "language"?
The two modes I can think of are "imperative" as in Perl/Python/etc,
or "declarative" as in Prolog and backward-chaining logic systems.
I think we need a new "desiderative" mode, whatever that may mean.
It appears to me now that syntax and morphology are no problem,
rather it (the mystery) is all in the semantics.

--
As they say in Tepa: hike waipettu.