Open Project to Investigate Money and Economic Systems

Posts tagged python

A New Direction

:: code, ipython-notebook, python, optimaes

OPTIMAES is, once again, stirring slightly in its endless sleep.

We have, once again, a new blog. This time a flat site made with Frog. (I decided that the infrequency of posting here made the overhead of Wordpress unnecessary.)

So I wrote a quick importer for the old posts.

All of which would be pointless if it weren’t for a more interesting announcement. I’m moving OPTIMAES to IPython Notebook.

OPTIMAES was always meant to be a project that was about public play and exploration. While Python allows that; and sharing code via GitHub is pretty public; OPTIMAES was never very visible. IPython Notebook solves that problem, allowing us to turn programs into interactive documents with dynamic graphs.

The new code will be up on GitHub shortly. And there’ll be some instructions.

Thoughts on the codebase

:: code, python, repast

As usual, I’m in two minds about the future of the OPTIMAES code-base.

The original code was written in Python. Recent experiments have been done with Repast.Py.

There’s no doubt that Repast.py lets new users get going very quickly. It has a GUI, which lets you set up a simulation in a few minutes. And you get graphing and neat grid and network diagrams free. As one of the objectives of OPTIMAES is educational and particularly learning-by-exploring toy economic worlds, Repast.py fits well. It’s a good educational tool.

On the other hand, it’s not really Python, so you can’t use the full expressivity of the language to define complex models, nor can you integrate with other Python libraries. There are more powerful simulation and graphing libraries available which would be nice to use. But, these aren’t currently integrated into a package comparable with RepastPy.

At present I’m thinking of following the following strategy : the next few experiments will be done in Repast. However the original code-base will still be available to those who are interested. The Repast pseudo-python will be easy enough to adapt for use by anyone who’d like to work with this code-base.