Comment by tromp
1 day ago
> In the Callas Normal Form, the factors are integers p = 2^{n-1} and q = 2^{m+1}, where n ≤ m, and p and q are ideally prime, but don’t have to be.
The paper's formatting clearly went wrong here, as it should have read p = 2^n - 1 and q = 2^m + 1.
The "Proposed Quantum Factorisation Evaluation Criteria" are excellent, but for measuring progress, the required minimum factor size of 64 bits is too large. A good milestone would be a quantum circuit that can factor the product of any pair of 5-bit primes {17,19,23,29,31}.
I checked in with Scribble as he did the typesetting. He apologizes for the error but says working without opposable thumbs makes the work more challenging.
I have some ethical concerns here. footnote 6 clearly states that Scribble did not do enough work to merit coauthor credit, but if he was one of the primary researches for section 5 of the paper and was responsible for typesetting the entire paper, denying such a good boy sufficient credit for his work is a serious breach of scientific standards.
I think 8 bit primes is probably a better minimum. 5 bits is still small enough that randomly choosing a 5 bit factor will succeed 40% of the time. This is especially problematic since Shor's algorithm only has a 50% success probability per round, so you need some extra bits to be able to distinguish a correctly working quantum computer from a random number generator.
Thanks. The flawed superscripting was bugging me too. Easily detectable but a reviewer should have caught it before publication.