Re: The Swarm space (was RePast (another agent toolkit) released)

[Alex Lancaster]

>>>>> "CS" == Stumpo VIII, Cody (C M ) <address@hidden> writes:

[...]

CS> If we in general allow it to overcome its stigma as "that thing
CS> elementary school children use," it could be pretty valuable for
CS> a)developing simple physical sims, and b)providing swarm core
CS> developers an example of what easy-to-use-for-the-non-programmer
CS> really means. (Nothin' but model)

The goal of `easy-to-use-for-the-non-programmer is definitely part of
the Swarm vision, and StarLogo, in part, shows the way.  So we're
quite aware of what "easy to use" "really means".  However, our road to
get there is a little different to StarLogo.  The strategy that Swarm
has adopted is to build a core set of powerful features which can
ultimately be targeted by any number of interactive/higher-level
tools.  The XML/Java layer Marcus prototyped is one example of this,
MAML was another.

This way Swarm builds an extensible, powerful infrastructure that can
support a whole range of agent-based simulations.  The cost is that
using it (in it's current form) requires a modicum of programming
skill (at least until higher-level tools become available).  In this
sense the Swarm kernel is the minimal approach to developing
technologies to model complex systems (find the lowest common
denominator across all kinds of agent simulation, factor out
everything else, and implement that, start building higher-level tools
on top of the kernel).

The alternative was dumbing-down the feature-space, which would have
the (undesirable) effect of a built-in restriction to the space of
possible models the user could implement.

To use a YALOSA (Yet Another Lame Operating System Analogy - thanks
Miles): to jump straight in with a shrink-wrapped environment for ABM
from scratch, would be rather like trying to write GNOME or KDE
without the Linux kernel.  You'd probably end up with writing
something like a terminal emulation program.

CS> And hey, it runs on the connection machine.  How cool is that?

No objection to StarLogo at all, I think it's great - no stigma at
all.  In fact at one point we talked to the developers about an
`export' feature in StarLogo to output a Swarm program, so it could be
used to prototype simulations which would ultimately target Swarm as a
production environment.

Alex

-- 
Alex Lancaster * address@hidden * www.santafe.edu/~alex * 505 984-8800 x242
Santa Fe Institute (www.santafe.edu) & Swarm Development Group (www.swarm.org)


                  ==================================
   Swarm-Modelling is for discussion of Simulation and Modelling techniques
   esp. using Swarm.  For list administration needs (esp. [un]subscribing),
   please send a message to <address@hidden> with "help" in the
   body of the message.
                  ==================================