I assume that [EMAIL PROTECTED]@q is not a requirement. That is, you want [EMAIL PROTECTED]@q to compute something useful without the requirement that f be restricted to i. . The answer is, yes, there are some.
> and < are an obverse pair; >@< is the identity but <@> is not, but f&.> is the useful "each". */ and q: are an obverse pair; */@q: is the identity but q:@(*/) is not, but >:&.(q:^:_1) demonstrates that there is no largest prime. <:&.(p:^:_1) n computes the largest prime less than n and >:&.(p:^:_1) n the next larger prime. [EMAIL PROTECTED] is the identity but [EMAIL PROTECTED] is not (the imaginary part of the result of ^. has magnitude less than o.1). Find your own application of [EMAIL PROTECTED]@^ . *:@*: is the identity but *:@%: is not. ----- Original Message ----- From: Norman Thomson <[EMAIL PROTECTED]> Date: Monday, March 5, 2007 2:33 pm Subject: Re: [Jprogramming] Truncating > Programming forum <[email protected]> writes: > >The point is not the number of characters but a > >pattern that is useful in other areas. In this case > >the pattern is duality. > What makes this case rather special is that, writing p for p: and > q for > p:^:_1, we have verbs for which [EMAIL PROTECTED] is identity but [EMAIL > PROTECTED] is not. > However [EMAIL PROTECTED]@q gives a valuable outcome. Are there other > instances of the same > sort ? Roger? ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
