Roger Hui wrote:
> See also the essay in the J Wiki
> http://www.jsoftware.com/jwiki/Essays/Logarithm_and_Exponential_(Extended_Precision)

Perhaps it is worth pointing out that exp and ln are sensitive to the
representation of the argument, as in the discussion of 1p1. Naively
equivalent arguments can have significantly different values, and the
differences are magnified by the precision being used.


   0j40": exp 0.002408597
1.0024115000000053605232609400393432850079
   0j40": exp 2408597r1000000000
1.0024115000000053504529356718870476062262
   0j40": exp x: 0.002408597
1.0024115000000029884975365923722529060429


John


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

Reply via email to