Henry, I've been hoping someone else would ask about or clarify your kr2 answer, but since no one has I have to ask, how the right tine of kr2 works when almost no part of it works, but the whole does work? Below are my experiments that produce length errors as well as the whole which you used.
Btw, I tried looking at it with dissect but that did not help. When I finally ran "trace'a (*"0 1"1"1 2) b'" I may have gotten a hint at something different when the last step was parsed by "5 trident" and not "2 dyad". But that still has me lost. $a *"1"1 2 b |length error | $a *"1"1 2 b $a *"1 2 b |length error | $a *"1 2 b $a *"0 1"1"1 2 b 2 3 2 3 $a *"0 1"1 b |length error | $a *"0 1"1 b $a *"0 1 b |length error | $a *"0 1 b -- (B=) ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
