18 February 2014

Post date: Feb 18, 2014 4:15:23 PM

Brainstorming for topics to cover next week when Max Albert comes to visit to talk about programming issues:

• What is IPython and what can it do for me?
• When is it worth using parallel programming techniques for simulation?
• There are tools out there like KNITR, an automatic report generation package that helps to integrate the output of simulations with R analyses. This seems similar to the use of Python notebooks to generate interactive documentation of simulation code. Should we be using these kinds of tools more?
• Question of how to save program state, output data and parameter lists when simulations get complicated.
• Why should I use a version control system?
• What is refactoring and why should I learn about it?

Additional discussion:

Some comments from Chris Hughes on his model of innovation in a network of firms.

Discussion of a model of Roman trading networks by Tom Brughmans, question about how to represent the basics of trade (supply, demand, stockpiling decisions, etc.) Suggestion to look at the Sugarscape model with its two-good trading system, Tesfatsion's intro to agent-based computational economics, intro-to-econ textbook on supply and demand, etc.

Brendan Neville on his model of homophilic social networks for trading. Importance of getting the mechanisms for making and breaking connections right. Difficulties of having agents infer their similarity to other agents based on their response to common situations (products / artefacts) and then making or breaking connections accordingly.