Re: project COSA

From: David Hart (
Date: Sat Aug 10 2002 - 08:47:52 MDT

Ben Goertzel wrote:

>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.
I think an eventual fusion of the two is inevitable, as interaction with
virtual or augmented environments *is* programming. I might give a
complex Flare object to your agent with a combination of verbal, written
and visual (possibly embedded in the surface of the object) arguments.
I'd be programming (with your permission and in a sandbox) your agent. I
certainly don't want to see a bazillion XML strings fly by -- I want to
see some physical representation of my choosing.

* * *

I think that Savain's point about robustness is intimately tied to the
complexity of a system. It seems to require exponentially more
programming effort to retain robustness in a system of bolted together
software with linearly growing complexity (like military, aerospace or
AI tech). Is there any way to test Savain's claim other than to build
equivalent signal-based systems? If signal-based programming had the
same amount of human experience behind it as traditional programming
does today, would Savain's claim of 10-fold robustness and ease of
programming seem less ludicrous?

* * *

We humans might find the notion of programming signal-based software
unappealing, but what about an SI? Signal-based systems have the benefit
that the same system could easily have both hardware and software
implementations, and it can be made to easily integrate with wetware. In
our limited experience, we've seen them only from evolved origins, but
that doesn't mean they might not be vastly more powerful when engineered
in the hands of an SI.

Who knows, an SI might discard our semantic-language way of programming,
using it only when necessary to interface with 'legacy' human systems.
Can we predict even possible programming paradigms of an SI?


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