I originally tried asking this on StackOverflow, but it was too subjective :-(. I am interested in methods of defining the power of programming languages. Turing completeness is one, but it is almost universally satisfied. What would be nice is to define a measure of power that discriminates among programming languages that are actually in used. For example, can anyone propose a non-subjective method that would discriminate between assembly and Java?
Turing completeness means that a language is maximally powerful in what it can output (which pretty much means it can do anything non-time based in the real world). So if we want to define a stronger measure of power, we need to take another approach. Shortness was suggested in the original question, but this is not easy to define at all. Does anyone have any other suggestions?