I chose Scryer Prolog because I think is the more reasonable Prolog out there (not that I investigated much, but swi always felt too deviant from the ISO Prolog, in the sense that in the past many times I searched for help the answer was "nah, that is just a SWI-Prolog specific thing"). Also it feels principled, simple enough, yet powerful enough. And the authors are super responsive.

I'm not an expert on Prolog, but I like how clean the Prolog solution looks.

This being said, I try to be as idiomatic as possible, using DCGs where possible and avoiding "imperative style" prolog, and trying to avoid cuts.

Any improvements or corrections you see on the code, please send them my way as only by correcting our previous knowledge we learn new stuff.

I found quite useful to have the following function around in my emacs as ediprolog does not work for me on Windows:

(defun javier/scryer ()
  "Call scryer on save"
  (when (eq major-mode 'prolog-mode)
    (message "%s" (string-trim
      (shell-command-to-string (format "scryer-prolog '%s' -g halt." buffer-file-name))))))

(add-hook 'after-save-hook #'javier/scryer)

It shows on the minibuffer whether scryer complains on the current file, which makes it somewhat easier to find syntax errors and missing punctuation symbols.

Disclaimer: I went back to read The Power of Prolog because I was feeling my answers were A) not fast enough B) not declarative. From this point on I try to minimize the use of custom "imperative" constructs I had in my src/utils.pl and instead declare relations. The results are much faster programs.