But now that you mention it, I don’t know that I need that flexibility and therefore my code is perhaps more complex than it needs to. I should perhaps have waited until I needed that flexibility and then refactored the code.

The dynamic nature of Ruby may have been the way to go. I’ll play around with both approaches and see what feels most natural to me.

Thanks for the input.

Of course there are other reasons for creating factory classes, but often they are created just in order to be able to inject mocks. While this is necessary in languages like Java (at least without delving into bytecode), in Ruby its possible to use meta-programming to achieve the same effect without cluttering code with factory classes.

I actually like not having calls directly to the constructor from my production code. I like the way the class gets decoupled from an actual implementation of the interface it needs.