From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-1.2 required=5.0 tests=BAYES_00,INVALID_MSGID,
	PLING_QUERY autolearn=no autolearn_force=no version=3.4.4
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 103376,9c6cb042c6c5955f
X-Google-Attributes: gid103376,public
From: dewar@cs.nyu.edu (Robert Dewar)
Subject: Re: Does Ada95 beat FORTRAN?!?
Date: 1996/04/21
Message-ID: <dewar.830136560@schonberg>#1/1
X-Deja-AN: 150736957
references: <00001a73+00002ce8@msn.com> <317906B6.42853EF6@cpmx.saic.com>
organization: Courant Institute of Mathematical Sciences
newsgroups: comp.lang.ada
Date: 1996-04-21T00:00:00+00:00
List-Id: <comp.lang.ada>

Linh says

"Not by the engineer or scientist's point of view."

responding to the questoin of whether Ada 95 measures up as a possible
replacement for Fortran?

Care to elaborate? I certainly know that if you consider only raw
performance on super computers, then Ada 95 may not match Fortran
performance, but it is after all conceivable that Fortran engineers
and even scientists might be interested in factors other than raw
performance (e.g. getting reliable and mainable programs, and for
that matter getting the right answers).

Here is one respect in which Ada 95 is clearly superior to Fortran
for numerical applications. In Ada 95, there are accuracy requirements
for the trig functions, there are no such requirements in Ada 95.

Someone sent me some interesting mail recently. They asked about my
statement that GNAT did not necessarily meet these accuracy requirements,
and wondered if they had to take special precautsions in writing code
using the trig functions. I explained that what I meant was that we
used the same routines that Fortran uses, and of course that menat we
had no idea if they met the accuracy requirements of Ada :-)