Module Names

[thorisson]

One of the issues in building a set of modules is the naming. If a module is an encapsulation of a particular function in a system, then the name of that module should reflect the name of the function. The trick is to pick the name of a module to describe the function that the module serves, implying no more and no less in its name. Otherwise what will happen is that the name of the module will, at best, confuse the system builders over the course of building the system. At worst it will mess up message type naming, and thus the naming of other modules, and negatively affect the design of the architecture, severely degrading the clarity of the design.

[vignir]

I think it can also help to use suffixes in module naming which express which category the module belongs to. If a module is a perception module that perceives when the agent is out of breath it's name could be OutOfBreathPerceptor. A decider module that handles the breath of the agent would then be called BreathDecider

[thorisson]

This is a very good idea; we have started working on a message type ontology for cognitive systems -- this would call for the equivalent for processes, i.e. process type ontology for cognitive systems. This might in fact be something that the MINDMAKERS community could collaborate on. (Of course there will not be a single ontology for any of this, but at least where people agree there is no need to use different naming conventions to confuse matters more than necessary.)