Yep, All the pieces fit, don't they. :-)
I guess that if you did want a fuzzy fit with a zero you could always establish the bounds through calculation and comparison. There really isn't a primitive that will do that. Cheers, bob On Nov 29, 2014, at 10:57 PM, Roger Hui <[email protected]> wrote: > See > http://www.jsoftware.com/jwiki/Essays/Tolerant_Comparison#Comparisons_with_0 > > On Sat, Nov 29, 2014 at 10:22 PM, robert therriault <[email protected]> > wrote: > >> It looks to me as if the fact that 0 is being compared is significant. >> >> 0 (~:!.0) 1e_45 >> 1 >> 0 (~:) 1e_45 >> 1 >> >> I would expect the intolerant ~:!.0 to return 1, but the tolerant ~: to >> return 0, except that I think that 0 may be multiplied by the default >> tolerance to produce the intolerant fit of 0 >> >> You could also do a definition using match (-:) instead of not equal (~:) >> >> matchnotreal =. -.@-: + >> #I.0=(iscomplex=matchnotreal)1 j. -:^:a:1 >> 1 >> >> J can be so surprising. >> >> Cheers, bob >> >> On Nov 29, 2014, at 10:09 PM, Henry Rich <[email protected]> wrote: >> >>> Yes, I believe it would. >>> >>> Whether that's desirable is a different question. >>> >>> Henry Rich >> >> ---------------------------------------------------------------------- >> For information about J forums see http://www.jsoftware.com/forums.htm >> > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
