So basically this interpreter checks if a loop contains only linear transformations of the variables in the lexical scope, and if so it builds the corresponding

So basically this interpreter checks if a loop contains only linear transformations of the variables in the lexical scope, and if so it builds the corresponding matrix and computes the appropriate power of it. If so that's fine but it's probably a rare use case, since loops usually have conditional branches, don't they? If they don't the algorithm is poorly written in the first place as it should have been noticed that there was a linear transformation involved.