|
From: | Marcus G. Daniels |
Subject: | Re: [Swarm-Modelling] Something Glen said |
Date: | Wed, 22 Nov 2006 20:36:09 -0700 |
User-agent: | Thunderbird 1.5.0.8 (Windows/20061025) |
Rob Bowers wrote:
So is it the opinion of the Swarm community that a new user would be better off investing in Repast?There are a lot interesting ideas out there and toolkits. NetLogo, IMT, Brahms,. EcoLab, MASON just being a few examples. I'd say try some different packages out. If you just want to write a program that has graphical widgets and probes, just take a look at what Visual Studio or XCode can do. If you just want to get started your priorities may be different than someone that wants rigorous modeling abstractions. Or maybe you want total control or maybe you want maximum scalability.
I think if there was a self-evaluating, distributed, high performance programming language and runtime platform that would be 90% of the battle. Like OpenMP with distributed objects, minimally, or perhaps a multithreaded R with native code on-the-fly native code generation. If something like that existed and really worked (and scaled) then it might be interesting to think about some domain-specific abstractions. That's where the mechanical complexity is.Has a merger of energies been considered?
[Prev in Thread] | Current Thread | [Next in Thread] |