Roger Hui wrote:
> I know that SHARP APL (and SAX) behaves the same as J
> in this regard.  See: 

I don't have SHARP APL (and SAX) installed. From the old Iverson (Sharp) APL
reference manual,it said
"The trend (signum) is not affected by []ct, comparison tolerance"
(the original text printed "not" in italic)

I guess this meant comparison with zero is exact. Did later SAX change it to
current behavior of J's signum?

IMO tolerant <.0 and >.0 is more useful and intuitive, just like the case *0
being tolerant.  Any chance that it will be incorporated?

-- 
regards,
bill
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to