Sheeesh. I've been trying to ignore this thread. My two bits worth:


> Dan Ingalls
> 
> Our work has
> followed a two- to four-year cycle that can be seen to parallel the
> scientific method:
> 
>           Build an application program within the current system (make
>           an observation)
> 
>           Based on that experience, redesign the language (formulate a
>           theory)
> 
>           Build a new system based on the new design (make a
>           prediction that can be tested) "

This is pretty much how some philosphers of science have described the
scientific method. I've yet to meet a scientist who was quite prepared
to nail his/her flag to that mast! 

Popper, ever in search of rigour, held that if the test led to the
theory being falsified, the theory should be rejected. That this idea
was absurd has been pretty well demonstrated by Paul Feyerabend in
Against Method, using some rather good case studies to demonstrate the
complexity of the scientific process (but Feyerabend himself sometimes
cheats, just like the scientists he admires). Indeed, the whole edifice
of the scientific method has been pretty much blown out of the water,
stamped on, sprayed with gasoline and burnt to a cinder by Kuhn,
Feyerabend, Lakatos et al. Which has, I suspect, impacted on scientists
not a jot, but then, why should it?

If we see OSI as a "get it right first time" design (and I'm not so sure
I'm prepared to see it as anything other than a reference model) then
surely it can be seen as the networking equivalent of a Popperian
scientific method! So, it's hardly surprising that a "real"
scientist/network engineer goes and puts together something which works
which touches the method/model/design where it fits (or where it needs
to to secure funding).

Cheers

John


-- 
Reputed to be the reason Windows 2000 was nearly a year late, (paid in
shares M$Ds needed the cash and kudos) Netproject's Eddie Bleasdale has
renewed his challenge to virus writers. The first person to infect his
Linux box wins 10,000 pounds.

http://www.silicon.com/bin/bladerunner?30REQEVENT=&REQAUTH=21046&14001REQSUB=REQINT1=48211