New arXiv math.CT preprint featuring Scryer

[dcnorris]

The effort I presented at an inchoate stage back in Nov at the 2nd Scryer Meetup has now been through the wringer of ACT2025 and finally has appeared in a mature form on the arXiv!

Being able to carry out reliable computations using the declarative integer arithmetic of @triska 's library(clpz) (in its monotonic execution mode, of course!) was crucial for the success of this work; I used the fair enumeration principles and comprehensive search mentality I learned from Markus to discover the several counterexamples elaborated in the diagrams on p.10, saving me the labor of proving false theorems! The ease of extending Scryer to include probability distributions for my simulations was also important (I'm hoping to contribute these via PR this summer). Perhaps the most indispensable contribution Prolog made here was through its DCG facility, which enabled the declarative specification of a simulation (see Appendix A) demonstrating the feasibility of a trial protocol built on the categorical machinery developed in the paper. I can hardly imagine a more parsimonious or 'face-valid' way of doing this. I am also delighted to have been able to cite the now-published DCG Standard in this connection! For clinical trial applications, the existence of such standards is enormously important.

[triska]

Wow, congratulations David! Amazing work and news!

[hurufu]

Code style in the project looks pleasant to look at – tests just after predicate definitions also help, because shape of the data can be easily understood and remembered. Interesting approach. I will also utilize quads in my next Prolog programs.