_RunVariations_ should run the model and actual function itself, so it can compare results and error messages.

[pmikkelsen]

Right now, the _RunVariations_ operator takes two functions as operands: the primitive being tested, and a user-defined function which is supposed to be an exact "model" of the primitive.

I want to be able to use _RunVariations_ such that when I give it some input data, it runs both the operand functions on the data (or permutations/modifications of the data), and compares the result values, or the error messages produced.

Currently, the operator expects as part of its argument, the data which has already been run through the model function. This prevents me from using it in any way, when I expect the model function to produce an error. I see no need for me to run the model function before running _RunVariations_, as it is already given as an operand.