Swarm announcements/plans for the next year

[Chris G. Langton]

Hi folks!

  The Swarm Hive here at the SFI has re-consolidated
after a month during which we were all pretty dispersed.

Two bits of news:

  1) Glen Ropella has joined the Swarm group to replace
     Nelson, who has gone off to graduate school at MIT.
     Glen's email address is <address@hidden> and 
     Nelson will still be reachable at <address@hidden>
     for the forseeable future.

  2) Speaking of the forseeable future, over the next two 
weeks, we will be working out out our plans for the next 
years' worth of work on Swarm. 

In the near term, we will be finishing up basic 
building blocks in the Beta version, with an eye
towards releasing V1.0 later this year. So, we'll
be occasionally distributing upgrades to the Beta
version over the next few months.

In the longer term, our overall orientation will
involve a shift from fundamental Swarm implementation
machinery to fleshing out the higher-level
libraries of tools and objects that run on the Swarm 
machinery. 

  For example, the recent queries about getting 
random objects out of lists or collections, or
about getting random permutations of lists each
time they are traversed, are features we've been
planning on for a long time, but had to wait
to implement until the basic list, collection,
and other "fundamental" building blocks were pretty
stable. Now we'll be adding the diversity of functionality
on top of these basic-objects that will provide a wide variety 
of access and traversal methods. So, lists and collections
will, e.g.,  have subclasses that will first permute the
list and then traverse it, or will return randomly
selected elements according to default or user-supplied
RNG's etc.

  There are also plenty of user-requested capabilities
that we will be reviewing and fitting in to the year's
plans.

  The above examples of random selection and traversal
methods are just hints at the kinds of things in the
works. We'll post a more extensive list as we plan
over the next couple of weeks, and we'll be posting queries 
to the user community on swarm-support about a lot
of different things. We will post a "call-for-input"
in the next week. If there is something that needs
immediate attention, post it to swarm-support as usual,
but y'all should make some lists of suggestions/comments
etc. for things that you'd like but aren't immediate 
needs and hold on to them 'till we make an explicit call
(...because we'll probably set up another mailing list to send
them to in order to stay "organized" here...)

So the next year will see the addition of a lot more 
functionality and useful building blocks at a higher
level of representation in Swarm. We hope to distribute
"conventions" for user-supplied objects and libraries,
and will provide templates and suggestions for how to
assist the user-community by coding via re-usable objects 
"ala Swarm", so that generic stuff one user creates can 
be contributed to the Swarm libraries and used by other
people.


We are also planning a Swarm-Users meeting here in Santa Fe
for early 1997 (January or February) to provide an opportunity
for different "hives" of swarm users to get together to
compare notes, demonstrate their models, and also for us 
to present some tutorials on building and interacting with 
Swarm models, based on our distillation of experience and 
what we've seen happening in the user-community. Hopefully
one outcome of the meeting will be a clearer understanding of
the extent of the set of functionalities that should
be provided in a rigorous and generic fashion in the
swarm library hierarchy. We also intend to hold workshops
on how to make swarm experiments repeatable and 
comparable, so that other groups can reproduce
results in the proper scientific tradition.

  Thanks for all the feedback and contributions so far,
and you'll be hearing a lot more in the near future.

Chris Langton
for the Hive here at SFI