Re: [Swarm-Modelling] newbie question

[gepr]

Paul Johnson writes:
 > Swarm reflected a paradigm-shifting set of concepts that we did not (do 
 > not) understand fully.  The original Swarm idea was so big and 
 > paradigm-changing that I doubt anybody could have had the foresight to 
 > carefully structure the research agenda in order to cultivate funding.  
 > In fact, if you go back and look at some of the initial design and 
 > strategy, or comments about Swarm, it seems to be designed in the 
 > opposite direction.  Way back when, say 1996 or 1997, Swarm had the 
 > "giant picture" framework in place with a lot of methods that were just 
 > stubs and the source code said "not implemented."  It is very difficult 
 > to explain to a funding agency that you want money to fill in the "not 
 > implemented" part.  Its even more difficult to explain why you want 
 > money to replace a "working prototype" with a newer, better thing.
 > 
 > I hope this observation does not offend or cause frustration among the 
 > people who have worked so incredibly hard on Swarm itself.  I do not 
 > intend to denigrate either the grand vision of the project or the hard 
 > work of many programmers along the way. 

Not at all!  These observations are needed.  Thanks for providing them.

I agree completely on the reasons for a lack of funding.  There's no
compelling "vision" to incite an agency to fund the development of
Swarm.  Part of the reason for the lack of vision is because the goal
for Swarm was to be a _platform_ for the study of complex systems.
There are 2 problems with that: 1) "platform" is a specious word for
the most part and 2) "complex systems" is as ambiguous as any
buzz-word has ever been in the history of man. [grin]

I can point out case after case of failed attempts to build platforms.
It happens all the time, over and over again.  And it's likely that 
it will continue happening and it will continue failing.  Along these
same lines are the OS wars between things like FreeBSD and Linux, or 
Windows and OS/2.  It also includes formalism, database, programming
language, and text editor wars.

So, that part of the vision for Swarm was definitely ambitious.  Note
that I'm not claiming that we shouldn't continue to try and fail...
one learns alot by failing.

However, agent-based systems _are_ becoming more well established as
"tools for the study of complex systems".  And it seems fairly sensible
and feasible to carve out a piece of "tools" and work on that.  Many
of the other agent-based modeling tools are doing exactly that.

The question is:  What is Swarm particularly good at/for?  *That*
part of Swarm should be fostered and grown.

I think it's safe to say that the Activity library is interesting
and still useful (though mostly unused).  I also think it's safe to 
say that the "integration approach" that we've fostered in Swarm is
interesting and still useful.  So, my guess is that both of those
aspects of Swarm need to be pursued and stand a chance to incite 
an agency to fund their further development.

It's more debatable; but, I also believe that the "object lifecycle"
or defobj library is interesting and useful and should be pushed
either to completion or expanded and re-implemented.  With the various
attempts at things like run-time compilation, reflection, and
specification/description of adaptive/evolving systems, it's clear
that this domain is still subject to research, both basic and applied.
It's also my opinion that it is the defobj library that sets Swarm 
apart from the other ABM tool sets.

A fourth, more vague, thing Swarm had as a part of its original vision
is centered more around "exploratory" or "synthetic" modeling.  Chris,
Roger, or David can correct me if I'm wrong; but, "play" was an
inherent part of the original agenda.  Swarm was NOT intended as a
simulation tool.  (In fact, I maintain that it's not and never was a
simulation tool, for whatever my opinion is worth. ;-) It is this
aspect that I'm working on right now and that I think has a solid
future.

The idea, here, is that the computational constructs you are creating
when you build a swarm are not _models_ of something else so much as
they are things in and of themselves.  They are autotelic.  The 
goal is to do scientific experiments on the computational constructs.
In this vein, things like "validation" and "verification" are nonsense
except in the more traditional uses of the words that bear on 
things like "trust" and the ability to re-perform experiments with 
roughly the same methods and materials.

That is where the future lies and I believe that Swarm contributed
to that agenda.  The simulation side-track is analogous to the 
pork that is exploited by politicians. [grin]

-- 
glen e. p. ropella              =><=                           Hail Eris!
H: 503.630.4505                              http://www.ropella.net/~gepr
M: 971.219.3846                               http://www.tempusdictum.com