Are there any mechanism available for formal comparison of interfaces? We can derive algorithms from a set of desired properties, and compare them by the guarantees of these properties. We can prove data-structures equivalent with algebraic data-types, and properties of operations. We have guidance to only optimize as necessary, to encode invariant in the type, and to have the structure of the state-space match the problem to be solved. By contrast the means of comparing interfaces are seemingly lacking. At best we seem to be given questionable principles for design in process rebuked by practitioners and academics (e.g. Norvik) alike. Must this be a matter of philosophy, if so are there at least any which maintain the discipline of philosophy?
Just prove every possible usage of one interface has an equivalent usage in the other.
>>2
This would tell you that the other implies the one, but I'm not sure we want to hold the possible operations invariant, nor to look for strictly expanding/reducing possible operations.
What kind of interfaces? User interface?
>>4
I was thinking sets of operations, that is any interface. I would be interested in any specialization however if you have relevant information.