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

[Miles Parker]

>> Marcus G. Daniels <address@hidden> 02/02/00 01:38PM >>>

>>From one perspective, it's good if the various ABM toolkits focus on 
performance in a practical way, i.e. if anyone should be fast at
SugarScape, it ought to be AScape, eh?

No doubt. :-)

>>However, it would be a pity if we start bending designs around naïve
performance questions and comparisions.  The only good I can see
coming from that kind of effort is that makes users more aware of
common bottlenecks, and their associated infrastructure: "Boy, if you
want fast permuted Java callouts over homogenous sets, use Swarm's
FActionForEachHomogenous interface".  

I absolutly agree with these sentiments; a framework _user_ should not have to 
worry about how these optimizations work..it should be the framework 
_designers_ job to worry about these things, and they should be transparent and 
typically not exposed. At the same time, we might hide details so much that the 
users aren't aware of the tradeoff decisions they are inherently making. 
Definitly a rich design tension.

>>It is probably possible to use
heatbugs or SugarScape or whatever as a benchmark, but people should
realize that it says as much about the coding skill that goes into
making the model efficient as it does the underlying infrastructure.

Generally agree, esp. in terms of overall infrastructure, but in some cases the 
framework infrastucture can provide the speedup seemlessly. In the case of 
sugarscape, in one case the code takes advantage of one inherent and 
transparent opportunity for optimization, and in another case the framework 
user must provide a little extra help, essentially by having cells let the 
framework know when they have been updated. IMO, this kind of thing is OK as 
long as it is conceptually striaghtforward and the default situation is not to 
have to worry about it.

-Miles




Miles T. Parker
Software Engineer
The Brookings Institution  1775 Mass. Ave. NW  Washington, DC  20036
http://www.brook.edu/es/dynamics/models/ascape
mailto:address@hidden  voice 202.797.6136  fax 202.797.6319



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