Re: Any old Swarmers reading the list today?

[Nelson Minar]

Chris is really the right person to answer all these questions, maybe
he'll chime in. But memory is fun, and it's better to write this stuff
down sooner rather than later.

I was around from the beginning of the current Swarm, the one in
Objective C. But there was an earlier Swarm before then, written by
David Hiebeler and Chris, called pgas, for "process gas". The basic
agent-based simulation stuff was present in pgas, but no hierarchical
swarms and simple time-step scheduling. It was written in C with a
bunch of stuff that made it look vaguely object oriented, and every
agent just received a "step" message every tick of the clock. The
heatbugs y'all know was a clone of Dave's pgas version, the first
proof of ObjC Swarm.

>How did they decide to build the project under the GPL?

>From the beginning (pgas and before), Chris intended to give the code
away - it's a research too, afterall. I proposed GNU style licensing
(the LGPL, actually) as a way to formalize that. The only issue there
was understanding the damned details of licensing, pretty much
everyone agreed that free and open source was the right idea.

>How were the original programmers drawn into this project?

I don't know about David, although I think he started doing a CA
package on the CM-2 with Chris. I spent a summer as an undergraduate
intern (under the NSF REU program, bless them) working with Chris,
Dave, and Walter Fontana to implement a model of Walter's in
pgas/Swarm (it was a spatialized AlChemy, if you know hiw work). I had
a great time then and got hooked on complex systems and SFI.

About six months later I was back at SFI working on my senior thesis
right when Chris was looking to ramp up the Swarm project, hire
someone fulltime to develop the system for release. (Dave had gone off
to grad school). Right place, right time. Hey, if you were fresh out
of college and *Chris Langton* offered you a job at the *Santa Fe
Insitute*, what would you say?

The other programmer from this period was Roger Burkhart. He got
involved with the project the same time I did, the summer of 1994, but
always in the employ of John Deere. Roger is largely responsible for
the serious OO and scheduling models - his background in CS was very
influential. I think I was the one who proposed reimplementing pgas in
Objective C. We were originally going to start with Dave's codebase
but wanted it in an OO language, so we started designing our own OO
extensions to C before we realized that they looked an awful lot like
Objective C.

Well, that's some of the back story from my view. Chris has a much
broader conception of the project as a whole.

                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.
                  ==================================