User talk:AlexKeane01

Feedback on Ita
Given your writing style, I presume you are probably the Itamaraca author, or a friend of his. I've done a bit of analysis on Ita and here's some feedback I have on it. Additionally, the site that you cite as providing evidence that Itamaraca is notable appears to just be copying the then-current list of random number generators off of Wikipedia, and the only reason Itamaraca was there was because you put it there. Irene Twoth (talk) 18:36, 9 May 2023 (UTC)
 * The claim that Itamaraca is "aperiodic" is hard to support. As actually implemented, it appears to not be, and the Itamaraca paper does not explain how to implement an aperiodic version of it. The WELLDOC paper, on the other hand, does explain how it creates aperiodic generators as far as I can tell.
 * I believe the title of the English version of the slides claim it might be a CSPRNG. Empirical analysis suggests that it is not. The slides themselves do not claim that, so that's good.
 * I do not believe it is accurate to state that Itamaraca is "based on the absolute value function". The actual mixing seems to be done by the PN*XRN calculation, which suggests it would be more accurate to say that it, like many generators, is based on multiplication.
 * The objective performance of Itamaraca on rigorous statistical tests is quite poor. It fails PractRand sooner than a 32-bit LCG outputting its lower 16 bits does, despite having a much larger state. I can't figure out how to determine how large the state is, but I would consider a double-precision based implementation to have a state of 257-321 bits; in any case, such a poor performance given a state of that size suggests Ita may have no advantages over existing, more common generators.