----- Original Message ----- 
From: Stephen White <spwhite / chariot.net.au>
To: ruby-talk ML <ruby-talk / ruby-lang.org>
Sent: Sunday, February 11, 2001 10:12 AM
Subject: [ruby-talk:10708] Suggestion for threading model


[huge interesting snip]

Here's my (mostly useless) opinion.

In terms of modeling the universe, this might be a very
interesting, useful approach. After all, every object in
the real world has its own "world line" and there is no
inherent "sequence" in accessing them.

As long as we're dreaming, I'd assign a separate 
processor to each object instance. And the instances 
wouldn't share any code among them.

The universe implements True Concurrency in possibly the 
strictest sense of the word... probably without even a 
quantized timeslice... and there is no evidence that God
does timeslicing anyway... :)

So in a sense I like this idea (more long term than short).

However, in terms of practicality and ease of implementation,
I think there might be huge problems.

To exaggerate a notch or three, it's rather like the idea of 
building a giant expert system with millions of rules, giving it 
a natural language interface, and slowly dumping all of human 
knowledge into it. (Things like this have been proposed 
seriously.)

To all who propose these things, I say, "Fine... go do it."

I recall an interesting 2,000,000-rule expert system proposed 
right here in Austin at MCC... but I haven't heard from it in
so many years, I'd bet a nickel it's defunct.

Hal Fulton