cog-bind-first-n is archaic

[linas]

... and should probably be removed. That is, unless I am having a brain-fart ... but if I'm not wrong, these functions don't do anything that cog-execute! does not already do. It would simplify code and simplify examples, if we reduce the number of functions.

I don't think there's any performance impact from this, either ...

[linas]

Note that bug #646 does seem to show a difference between cog-execute! and cog-satisfying-set but I'm not sure why there would be one.

[linas]

This is low priority, though, and is perhaps a matter of taste. We still need the "get-first-n" versions, since there is no way to have a generic "get-first-n" function in in cog-execute ...

Maybe even such low priority that this should be closed, without changing anything. Hmmm...

[linas]

See also #822 for related issues

[linas]

Partly done in #1457

[linas]

Partly done in opencog/opencog#2957

[linas]

Also handled in opencog/opencog#2958

[linas]

More removals in #1501

[linas]

To remove the cog-bind-first-n, we should invent a CutLink that would abort the pattern search, when some condition is met. For example, when the desired number of solutions has been found.

Similarly, cog-bind-af should be made deprecated, and any queries that need this should just use a predicate (GroundedPredicateNode) to test for the desired values of attention-focus.

[linas]

Suggestion #1502 seems to be a better solution for fixing cog-bind-first-n and cog-bind-af

[linas]

Closing. This can be done in pure Atomese, although it is currently verbose:

(Bind
    (And
         ...
      (SequentialAnd
         (True
            (SetTV
               (Concept "cnt")
               (PredicateFormula
                  (Number 1)
                  (Number 1)
                  (Plus (Number 1) (CountOf (Concept "cnt"))))))
         (GreaterThan (Number 2.5) (CountOf (Concept "cnt")))))
))