Reflecting on reflective computing

[Christopher Barry]

I agree with everything David Jeske had to say, particularely this
point:

David Jeske wrote:
[...]
> What happens if we spend 50 years to write a product which allows
> programs to be built in seconds, but programs are no longer useful in
> 50 years, because we've replaced computers with some suffuciently
> different technology? :)

I think this will indeed be the case. In 50 years I bet we'll have
something *superior* to the HAL 9000 series. But will we in 20 years, or
10? I really, really doubt it. Ultimately the concept of writing
programs will be obsolete. But Tunes would provide a great system and
framework with which to write these programs that will make programming
obsolete. But does Tunes need to run on the bare hardware to do this?

We need better programming languages, better compilers, and better ways
to verify these programs for correctness. I'm sure at least some of you
have stumbled across John McCarthy's page (principal creator of the
original Lisp waaaaaay back when):
http://www-formal.stanford.edu/jmc/

I've spent much time reading through his material and I think this page
contains more information relevant to what you guys want to do than 90%
of all your other links combined.

Non-monotonic reasoning, methods of formalizing common sense (in many
senses, mind you), methods of proving software correctness and compiler
correctness, a proposition for a new language called Elephant 2000 (not
so sure I'm particularely crazy about this one, but it looks
interesting), and other information can be found here. Most of what my
idea of the future of computing and software is has been influenced by
this, and I think AI is the way to go, not nessecarily cool object
oriented frameworks that allow *extremely* rapid development and low
maintainability effort, but don't fundamentally by themselves improve
the usefulness and intelligence of applications, since the algorithm of
usefulness and intelligence itself is still left to the human to
develop, just the 95% annoying details associated with this have been
eliminated.

Again, at any rate, to develop this framework (Tunes), your gonna need
some pretty intelligent compiler technology that doesn't exist yet and
some algorithms and methods that nothing too similar like has ever been
implemented.

Christopher