Exercise 12.txt
Paul Johnson

Concerning Parameter objects, batch processing, command lines, and so forth.


Please get the newest copy of my new tutorial component,
simpleObserverBug3, out of the newest tutorial snapshot.

http://lark.cc.ku.edu/~pauljohn/ps909/Agent-Based_Models/tutorial-2.2-pre2.tar.gz

After you study that, I want you to do these things.

1. Write some code in the Model Swarm that writes some output to a
   file. I don't care what output you write, but show you can make the
   fprintf thing work as described in the handout, which is also
   posted:


http://lark.cc.ku.edu/~pauljohn/ps909/Agent-Based_Models/fprintfnote.txt

2. Look at the Parameters class and add a command line option for one
   variable/parameter.

I don't care what parameter, or even if the parameter gets used in the
model. Just show you can do the job of adding a parameter and passing
it through to the command line.

I expect any of you can do both of those steps because the path is so
incredibly clearly marked out ahead of you.

3. Here is more challenging.

Note that simpleObserverBug3 does not implement a true batch mode.

Compare against heatbugs, which is also in the new tutorial package.
Note that if you run heatbugs,

./heatbugs -b

it runs in batch mode, no pictures.  Look at main.m, note that the -b
flag causes it to use HeatbugBatchSwarm in place of
HeatbugObserverSwarm.

What i want you to do for simpleObserverBug3 is to add a Batch class.
If we can do that in a clear, nice way, I will create a new tutorial
exercise for it, simpleObserverBug4. It would be a good use of our
effort.

You can do it if you use your heads. I can give you other models that
also have BatchSwarms to compare against if you need.

Please note, if you look in Heatbugs, it is using Lisp to create the
ModelSwarm, and I don't expect you to do that for
SimpleObserverBug3. Instead, just continue with the normal way of
creating modelswarms.

pj