That’s a very cool construction of regular expressions, just using matrices and semirings. I loved the explanation of why F and G were as they were as well. I h

That’s a very cool construction of regular expressions, just using matrices and semirings. I loved the explanation of why F and G were as they were as well. I had two comments: > Assume we have a regular language (R, 1, ., 0, +, * ) Shouldn’t this data (a semiring with * -operation or “closure”) not be one regular language, but _all_ regular languages? A single regular language should be an element of R. I also think that the power series A* = 1 + A + A^2 + ... looks more like 1/(1-A) rather than e^A, which makes sense given the defining equation A* = 1 + A(A*), but that’s not much of an issue since neither of these properly make sense in this semiring.