Algorithm = logic + control

[dcnorris]

In this famous paper of Kowalski's, I've encountered this lovely idea:

... with well-chosen names for the interfacing procedures, data-structure-independent programs are virtually self-documenting. For the conventional program which mixes procedures and data structures, the programmer has to provide documentation, external to the program, which explains the data structures. For the well-structured, data-independent program, such documentation is provided by the interfacing procedures and is part of the program. [p.430]

Could anyone point to some exemplary code in Scryer's libraries that illustrates this principle?

[bakaq]

I'm not sure the Scryer core libraries are the best place for finding this. Most of it is low level code designed to make writing this type of thing easier, but I can think of one example that gets close to this: library(pairs).

[dcnorris]

This suggestion surprises me a bit! To me, library(pairs) feels very 'close-to-the-metal', being wedded to the representation of lists of terms with primary functor (-)/2. The way I'm interpreting Kowalski here, his ideal might be something more like a Haskell typeclass that provides an interface to (say) a mapping from some finite set of terms to another set of terms, and allows us freely to swap out a pairlist for an assoc.

BTW, I've found it necessary to add related functionality in https://codeberg.org/dcnorris/DEDUCTION/src/branch/main/pairutils.pl

[bakaq]

To me, library(pairs) feels very 'close-to-the-metal' [...]

Well, I did say that the core Scryer libs aren't a good place to find that.

[dcnorris]

Maybe some of the contributed libraries operating at a higher level, then?

[bakaq]

Actually, I've found a perfect example! library(queues)! Really elegant.

[dcnorris]

This does bear close study, thanks!

[jjtolton]

queues is mind boggling. first described AFAICT in Craft of Prolog which O'Keefe describes as "your 3rd or 4th book on Prolog", and this version here I understand is actually the perfected version of what is presented by O'Keefe.

[UWN]

queues is mind boggling. first described AFAICT in Craft of Prolog

1983.

[dcnorris]

Tracing queues's evolution since 1983 should prove especially instructive! The 1983 code has an imperative flavor which has since been abstracted away, it seems.

[jjtolton]

I would also remark that @triska has a pretty nice commentary on that subject (A=L+C) in Sparrows on Eagles (one of my all times faves). Skip to around 40:30 for the punchline.