Re: Robustness Check and "A growing body of ad-hoc analysis solutions"

[Marcus G. Daniels]

>>>>> "B" == Benedikt Stefansson <address@hidden> writes:

B> For example, is the '@address@hidden' approach dead? 

It's deprecated, but not dead.

B> Which simple
B> ASCII format could become the chosen serialization file format and
B> an eventual parameter sweep file format .

We have two serialization backends: Lisp and HDF5. 
Swarm can read one format and write the other.

XML has been embraced by Microsoft, so that would be another useful
backend for interoperability with their apps.

Another useful backend would be SQL, so that folks could describe
agents, landscapes, etc. using standard database programs.  
           
      Popular Programmable Extensible Random-Access Efficient Text Prereq.
Lisp  -       +            +          -             -         Y    Lisp
HDF5  -       -            -          +             +         N    R or HDF5
XML   +       -            +          -             -         Y    MS Excel
SQL   +       -            +          +             -         N    MS Access

In 2.0, mousetrap's batch configuration file will use the Lisp backend.
Heatbugs now uses the Lisp backend.  Swarm Sugarscape uses the HDF5 backend.

Mousetrap's batch configuration file looks like this:

(list
 (cons 'batchSwarm
       (make-instance 'MousetrapBatchSwarm
                      #:loggingFrequency 1))
 (cons 'modelSwarm
       (make-instance 'MousetrapModelSwarm
                      #:gridSize 40
                      #:triggerLikelihood 1.0
                      #:numberOutputTriggers 4
                      #:maxTriggerDistance 4
                      #:maxTriggerTime 16
                      #:trapDensity 1.0)))

B> Keep in mind the lowest common denominator set of knowledge of
B> other tools/languages. 

Which, sadly, so often means "punt and invent something ordinary
and gratuitously different", which increases insulation.  Maybe
that is not relevant from the pedagogical perspective, but from
the practical research and modelling perspective it is a bad thing since
you can't talk to other modern tools.

B> Great to leverage off Scheme/HDF5/R but
B> pedagogically speaking they make beginners want to run for the
B> hills.

It would help if you explain why you think this is so (besides the
usual observation that the documentation for Swarm is poor).  I can see how
learning what a Scheme symbol, keyword, and a cons is would take a little
time.  Further, I know people that modify the .swarmArchiver file without
really knowing Scheme, I'm inclined to think there isn't a problem.

I can also see how it would take some time to learn how to make rudimentary
use of R.  But there is better documentation for this stuff than there
is for Swarm or Objective C.

Object archiving was added as a usability feature, but not necessarily
a `pedagogical' usability feature.  Maybe XML could help with that.  Or some
other format. 

Part of our mission is to support research, and to some
extent free tools like Swarm proliferate themselves in research
circles without a traditional teaching effort.


                  ==================================
   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.