ATAN inconsistencies

[bowman(at)null.edu (Kenneth P. Bowman)]

Here's a platform dependency that took me quite a while to track down.

The SGI (IRIX) version of IDL returns 0.0 when you try to take the
arctan(0/0).  The Alpha version (properly) returns a NaN.

Ken Bowman


IDL Version 5.1 (IRIX mipseb). Research Systems, Inc.
IDL> print, atan(0.0, 0.0)
      0.00000


IDL Version 5.1 (OSF alpha). Research Systems, Inc.
IDL> print, atan(0.0, 0.0)
         NaNQ
% Program caused arithmetic error: Floating illegal operand

-- 
Dr. Kenneth P. Bowman, Professor                409-862-4060
Department of Meteorology                       409-862-4466 fax
Texas A&M University                            bowmanATcsrp.tamu.edu
College Station, TX   77843-3150                Replace AT with @