silber@voodoo.ucsb.edu (10/13/89)
-Message-Text-Follows- During the 'workstation wars' in our dept. , prospective buyers of stations have been doing lots of benchmarking. We have received various complaints about the Decstation libraries AND some complaints cross ref'd to the standard VMS libraries. Some of the math routines return results which have been criticized (e.g. returning 0 instead of an error when evaluating sqrt(-1) or similarly when evaluating pow(-1,<fractional exponent>) Now I know that DEC systems, VMS in particular, have been used in scientific computing for a long time. Are these behaviours 'unacceptable oversights' or just 'features'???