Re: Languages for AGI [WAS Re: Syllabus for Seed Developer Qualifications...]

From: Simon Belak (simon@owca.info)
Date: Thu Feb 02 2006 - 09:58:05 MST


Quoting Richard Loosemore <rpwl@lightlink.com>:

> 1) Your requirment #1 (inasmuch as I understand what you are getting at)
> seems to be too much of a low level concern. Smacks of worrying about
> optimization prematurely....

Probably I was not clear enough (that, or we ortogonal views on programming,
considering I belive this to be the most abstract, or high-level if you will,
concept among my propositions).
To elaborate. I perceive the ability of a program to change (parts of) itself
during run-time the shortest path to RESP.

> 2) "Ability to create domain specific languages within itself": I can
> see why you say this, but I rank it a low level concern.

DSL vastly gain in power when the program can change itself since they
allow one
to create a more or less unified interface to any data/code (a trivial example
being having a single protocol to access a "node of data" anywhere, be it a
network package, memory page, XML file or a web-page).
True, omission is by no means a show-stopper but if nothing else usage of DSLs
can dramatically reduce code bulk, something you yourself, considering 6), put
great importance on.

Simon



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:55 MDT