Welcome! and call for introductions....

[Chris G. Langton]

Hello everybody!

  Since we have just added about 12 more sites to the Swarm beta-
program, I think it would be a good idea for everybody to introduce 
themselves to the other sites by posting (to swarm-support) a 
paragraph-or-two description of who y'all are and what projects you 
will be working on. 

As people start communicating about various problems/solutions/needs 
etc via swarm-support, this will help us all by providing the context of 
your project and establishing real people behind the cold-hard email 
addresses and frosty bug reports!

The folks who were in the original beta-release should introduce 
themselves to the new group as well.

The people at this end include Nelson Minar at SFI (chief-programmer), 
Roger Burkhart of John Deere (chief  Object-Oriented-Magic guru), JJ 
Merelo in Granada Spain (SGI version, neuro- and GA-libs), Manor 
Askenazi at SFI (implementing parallel kernel, and bringing MPI-based 
parallel batch-processing to Swarm in the near future), and me, Chris 
Langton (project manager, of sorts).

For the benefit of the new victims, er, beta testers, who have been 
dropped into the middle of a rough, and as yet not very user-friendly
version of Swarm, I will be posting some messages over the next few
days to try to give an overview of what our original goals for Swarm were, 
how and why we've arrived where we are today, what will be coming 
soon, and where we think this is all going over the next year or so. 

Much of what we've built to date is foundational machinery to support a 
wide variety of tools and library objects, which themselves aren't there 
yet. This may seem a bit confusing. 

Swarm now is mostly a "virtual computer" for managing the concurrent 
interactions of large numbers of agent-type objects. The high-level 
"software" that will run on that virtual computer is only there in 
preliminary form now, but the virtual computer itself is pretty solid. 

Think of it as the early days of Apple computers when you got the 
machine but very little software existed to run on it as yet, and you had 
to program the machine directly at the assembly level. With Swarm now, 
there is an "operating" system of sorts (the "activity" libraries), there are 
some utility libraries (e.g., simple neural-net and GA libraries), and there 
are the beginnings of the class-hierarchy of world- and instrumentation- 
building objects (spatial arrays, graphers, averagers, probe-tools...). Our 
primary effort in the immediate future is to rapidly build up the 
libraries of these basic simulation tools that researchers typically need.
"Assembly-level" coding for Swarm means simply coding in Objective-C.
We're hoping that users will contribute back to Swarm library objects
that they have written which might be useful to a wider community.

We strongly encourage the new users (if they have any hair left after 
installing Swarm and its attendant packages!) to first experiment with 
the included sample applications - play with them, look at the code, 
and begin branching out by tweaking that existing code in different 
directions. 

As mentioned in the documentation, the "heatbugs" application is 
probably the best place to start - it is the most-commented and the 
easiest to understand. 

There are still a bunch of code-fragments in heatbugs and the other apps 
which will migrate out as the proper library support objects come into 
existence, but for now the apps are built with kind of a low-level raw 
coding style. Coming soon will be support tools for constructing agent 
and space objects that will raise all this to a higher level, simplifying the 
process of implementing and interacting with experiments.

Maybe some of the early beta users can help initiate the new users to the 
rites of swarm, based on their experiences of hitting Swarm cold?

More soon...

Let's introduce ourselves to each other ....

Cheers!

Chris Langton