RE: project COSA

From: Ben Goertzel (
Date: Sat Aug 10 2002 - 07:32:51 MDT

> It probably is indeed inefficient to apply visual programming to heavily
> algorithmic environments, but it might be vastly better suited to other
> programming paradigms. Can this idea be dismissed so quickly?

People have been trying to make CASE tools and such work for a long time.

The visual-arrangement metaphor is just not as well-suited as language for
presenting extremely complex combinations of concepts. I think this is
shown by experience as well as by common sense.

Visual programming is one of those ideas that sounds better than it works.

Sorta like the idea of recording music by humming the notes instead of using
your fingers on an instrument. Sure, it seems easier at first, if you have
the proper tech. But ultimately, humming isn't going to give you the
ability to play Shostakovich or Yngwie Malmsteen within a reasonable amount
of effort...

> I again urge readers to ignore the emotive statements and arguments on
> the website and focus on the idea of a temporal-signal-based approach to
> software and the technical merits of the proposed implementation (COSA).

I tried to do that.

> I agree that pure signal-based systems would not be the best tools for
> most software tasks. I think the challenge comes in figuring out where
> they could best be applied (which may not always be obvious) and then
> figuring out how to integrate them with existing algorithmic systems.

I think that signal-based systems are more naturally *evolved* than

I think the best use for such systems is in carrying out real-world
categorizations and actions, as in the "training neural nets" paradigm.

However, the theme of the website is that signal-driven systems are more
reliable, and that, I just don't believe holds up except in very narrow and
simple domains, and even there only against straw-man algorithmic systems.
(I.e., simple signal-driven systems may seem robust compared to C programs,
but not against mathematically rigorously developed algorithmic systems.)

> A new question for thought: It seems that a complex signal-based system
> that would be difficult for humans to conceive could be conceived by
> sufficiently specialised AI software (e.g. a codic cortex).

Since humans have a hard time understanding each others' code anyway, I
think that ANY complex code written by an intelligent nonhuman is going to
be essentially impossible for us to understand.

-- Ben

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:40 MDT