The memo(s) strategy can only be used once for each input domain, because it stores everything in the same DR, regardless of the strategy s.

!1; memo(\1 -> 2); debug; // prints: 2
!1; memo(\1 -> 3); debug // prints: 2, expected: 3

It would be more reliable if it could somehow store a tuple (input term, unique id of s) as the LHS of the Memo DR.

Submitted by Tobi Vollebregt on 2 November 2011 at 14:11