swarm-modeling
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Swarm-Modelling] (no subject)


From: glen e. p. ropella
Subject: Re: [Swarm-Modelling] (no subject)
Date: Thu, 23 Nov 2006 09:20:36 -0800
User-agent: Thunderbird 1.5.0.7 (X11/20060927)

hilit wrote:
> Although my dad, brother and husband are all very good with programming,
> I haven't the slightest idea!!

Gary's advice is the eminent advice, then.  Sit down with one or more of
these people and ask them questions.  You should have plenty of
questions after having read all the opinionated hooha on this list.
Ultimately, if you have local programming expertise, then ask and follow
their advice, because such local resources can make or break your
project.  But, don't _allow_ them to do any of the work for you. [grin]
 That will just make your life harder later on.

> Could you write for me the steps in need to go through in building a model?
> 
> For example: 1. learn C++
> 
> 2. Download repast
> 
> 3. Write down a flow chart, etc….

1) Download Repast and Mason and follow their instructions until you can
run the examples.
2) Run the examples, tweaking parameters, run them again, etc. until
ideas start to spark in your head.
3) Write a hello-world Java program to get a bare working knowledge of it.
4) Modify one of the examples you played with above, recompile it, rerun
it to see your changes.
5) Repeat step (4) until you're almost comfortable enough to try your
own simple simulation.  Then try to write a trivial simulation from scratch.
6) Stop.  Back off.  Wait a week or so without doing _any_ work on this
in order to let you head settle.
7) Design _your_ cat model.  All this means is to try to think about and
write down every relevant aspect you _must_ implement to get what you
want.  It doesn't matter whether it's flow charts, box and arrow
diagrams, UML, or chicken scratches on a napkin.
8) Try to implement it.
9) Then when you realize it's not what you want, throw it away and try
again. [grin]

> Do I need to learn a programming language for the model or are these
> toolkits, like netlogo, gonna do it for me?

No tool will do anything for you, despite what the tool's proponents
say.  You _will_ have to become technical enough to iterate between
building your model, executing it, and debugging it.  And this process
is a slippery slope to developing programming competence.  So, you may
as well bite the bullet and learn to program.  BUT!!! you don't have to
become an expert.  You don't need to take a programming course or write
a bunch of programs.  Just get familiar with the underlying language of
the tool you choose.  I recommend Repast for what I've heard from you
and that means learn some Java.

> Is repast like netlogo like swarm? Or each is different?

They're all different.  Netlogo is higher level and more limited in what
it can do.  Gary's had good success with it and that means you might,
too.  Avoid Swarm, it is low level and allows you to make some pretty
arcane mistakes.  Java is a "safer" language than C (or Objective-C); so
it's easier to write programs that "just work".

> Sorry if it all sounds stupid, but I come from pure zoology so bare with
> me…

It's not stupid at all!  This type of conversation is _exactly_ why we
formed the SDG.

-- 
glen e. p. ropella, 971-219-3846, http://tempusdictum.com
Arms are the only true badge of liberty. The possession of arms is the
distinction of a free man from a slave. -- Andrew Fletcher


reply via email to

[Prev in Thread] Current Thread [Next in Thread]