>>>>> "NC" == Nicholas Clark <[EMAIL PROTECTED]> writes:
NC> If junctions are sets, and so a|b is identical to b|a, then isn't NC> it wrong for any implementation of junctions to use any NC> short-circuiting logic in its implementation, because if it did, NC> then any active data (such as tied things will side effects) may NC> or may not get called depending on whether a junction happened to NC> be stored internally with a first, or with b first? NC> (unless the implementation can prove to itself that nothing it's NC> dealing with has side effects, so short circuiting will have no NC> effect. Of course, this is an implementation detail, and aside NC> from speed, must not be determinable from outside) it also depends on the junction function :). i would think one() would need to check all the elements. but i agree with you about a junction being an unordered set. you have to treat it that way to make sense since you can add elements to a junction (can someone clue us in with a code example?). on the other hand we expect 'and' and friends to short circuit and execute from left to right and that is taken advantage of in many ways. so i say 'a or b' is not semantically the same as 'a | b' in that there is no guarantee of order in junctions. but there is no guarantee of evaluating all of the elements in a junction, it can short curcuit as soon as it can determine a correct boolean result (assuming a boolean result is wanted). uri -- Uri Guttman ------ [EMAIL PROTECTED] -------- http://www.stemsystems.com --Perl Consulting, Stem Development, Systems Architecture, Design and Coding- Search or Offer Perl Jobs ---------------------------- http://jobs.perl.org
