I hope it’s not too obvious, given who wrote this and who they were talking to respectively, but formal methods and category theory are easily far more powerful than their current applications would suggest. Even mathematicians are still skeptical of these fields; many folks have a lot of catching up to do. An example in formal methods might be Metamath, a syntax-driven prover capable of handling multiple different foundations of maths. It is faster than type-driven provers, verifying thousands of proofs per second. In category theory, Chu spaces generalize matrices, but also generalize a broad range of set-like objects, and have unexplored connections to programming and digital logic.

Metamath is also very, err, quirky. The usability is pretty low, I think, compared to other proof assistants. I think metamath-zero is a promising replacement with better foundations and barely higher complexity.

I know it’s beside the point, but the title of this post seems like a grammatical illusion I haven’t seen before. The first impression upon reading it is of the intended meaning, but then you read it again and… ???

I don’t think there’s any math I know that would help computing but isn’t, but I tend to learn math to help me do computer stuff, so.

I’ve seen some of the other side, where computers would benefit from some specific math. DARPA’s Hive Challenge put together a decent amount of prize money for better graph analytics in hardware or software.

I couldn’t immediately find the results, but I think most of the prize money was not awarded. I believe only one single award was made for an improved in-memory layout for graphs that could speed up analytics.

I suspect improved graph analytics would speed up Prolog and Haskell, and improve layout algorithms for PCBs and FPGAs.

Shinichi Mochizuki claims to have proved the abc conjecture. The problem is that he spend years on this, developing some obscure theory only he and some other Japanese mathematicians really (claim to) understand, and the mathematical community just doesn’t understand it yet.

(It’s a bit more subtle in reality, as people think there is a gap in his proof, which lowers the rationale to spend years and years on this obscure theory)

TL;DR: “I don’t know”

I hope it’s not too obvious, given who wrote this and who they were talking to respectively, but formal methods and category theory are easily far more powerful than their current applications would suggest. Even mathematicians are still skeptical of these fields; many folks have a lot of catching up to do. An example in formal methods might be Metamath, a syntax-driven prover capable of handling multiple different foundations of maths. It is faster than type-driven provers, verifying thousands of proofs per second. In category theory, Chu spaces generalize matrices, but also generalize a broad range of set-like objects, and have unexplored connections to programming and digital logic.

Metamath is also very, err, quirky. The usability is pretty low, I think, compared to other proof assistants. I think metamath-zero is a promising replacement with better foundations and barely higher complexity.

Accept that you are going to be very very slow and don’t be annoyed by the fact.

The formally verified HTTPS stack that the author has heard rumors of is probably Project Everest. https://project-everest.github.io/

It is used in Firefox and uses the z3 solver, which is an SMT solver.

Imagine being the mathematician whose discovery gets sudden real world usage after decades. There must be no greater high.

Either that or incredible fist shaking at all the years people ignored you :)

I know it’s beside the point, but the title of this post seems like a grammatical illusion I haven’t seen before. The first impression upon reading it is of the intended meaning, but then you read it again and… ???

I don’t think there’s any math I know that would help computing but isn’t, but I tend to learn math to help me do computer stuff, so.

https://en.m.wikipedia.org/wiki/Garden-path_sentence

I’ve seen some of the other side, where computers would benefit from some specific math. DARPA’s Hive Challenge put together a decent amount of prize money for better graph analytics in hardware or software.

I couldn’t immediately find the results, but I think most of the prize money was not awarded. I believe only one single award was made for an improved in-memory layout for graphs that could speed up analytics.

I suspect improved graph analytics would speed up Prolog and Haskell, and improve layout algorithms for PCBs and FPGAs.

Shinichi Mochizuki claims to have proved the abc conjecture. The problem is that he spend years on this, developing some obscure theory only he and some other Japanese mathematicians really (claim to) understand, and the mathematical community just doesn’t understand it yet.

(It’s a bit more subtle in reality, as people think there is a gap in his proof, which lowers the rationale to spend years and years on this obscure theory)