Markus Triska

And further, the issue may be confined to **lists**. For comparison, the following works exactly as expected: ?- T = f(A,f(A,[])), A = a, assertz(l(T)). T = f(a,f(a,[])), A =...

For comparison, just using `.` instead of `f`: ?- T = f(A,f(A,[])), A = a, assertz(l(T)). T = f(a,f(a,[])), A = a. ?- T = .(A,.(A,[])), A = a, assertz(l(T))....

And also, as already noted in https://github.com/mthom/scryer-prolog/issues/2668#issuecomment-2848598803, everything works exactly as expected when the exact same goals appear as part of a *program* instead of appearing only as part of...

Thank you a lot for working on this, I greatly appreciate your interest in improving `library(reif)`, and I think many Prolog programmers will eventually benefit from this! There may well...

> And now the most important point, after all of this I didn't observe any noticeable performance improvement! Am I doing something completely wrong? One important test would to check...

> In any case, the text of Scryer's library(reif) was apparently manually copied and rewritten. I think it would be good to start with restoring the **canonical** `library(reif)` straight from...

`library(reif)` was first added to Scryer more than 5 years ago in https://github.com/mthom/scryer-prolog/commit/b7dd76bf37cfbf8341a18d52f4ec61df4efacf93, long before Scryer had fully conforming Prolog syntax, a `meta_predicate/1` declaration, goal expansion and many other features that are...

Thank you so much for working on this, and congratulations on these awesome results! Regarding the remaining questions, my personal opinion: 1. It is a very worthwhile goal to **keep...

I have added a preliminary version of `statistics/2` in #2440 so that the benchmarks work as expected and with as few modifications as possible!

This is starting to look very good, very closely following the master copy of `reif.pl`! Well done! I still note the following differences that I think can safely be eliminated...