Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!uflorida!novavax!hcx1!bill From: bill@ssd.harris.com (Bill Leonard) Newsgroups: comp.lang.fortran Subject: Re: Two Fortran Standards Message-ID: Date: 7 Sep 89 17:28:51 GMT References: <282@unmvax.unm.edu> <303@unmvax.unm.edu> <1598@convex.UUCP> <1989Aug28.221032.21797@agate.uucp> Sender: news@hcx1.UUCP Organization: Harris Computer Systems Division Lines: 28 In-reply-to: jerry@violet.berkeley.edu's message of 28 Aug 89 22:10:32 GMT > I can find no justification for extra precision in assignment statements, > although I may have overlooked it. I think Jerry has hit on the real problem here. The question that the average programmer wants answered is: If I initialize a variable in a DATA statement, will I get the same results that I would get if I use an assignment statement? Rather surprisingly, neither F77 nor F8x address this question. I personally think that a compiler that tried its best to get the most exact answer possible from a given code would be the most desirable. > It's ridiculous that some standard conforming programs running on a VAX > will only run with DECs compiler and others only with the BSD compiler. I think this is a bit strong. Jerry's example tests for equality of floating-point numbers: a definite no-no. Anyone who does that deserves what they get. Nevertheless, I think Jerry has a point, but good luck writing a standard to enforce that. The standard cannot enforce the exact value of answers to floating-point computations: there's too much room for variance (e.g., architecture, optimizations, etc.). This is the type of thing that can only be enforced by the marketplace. -- Bill Leonard Harris Computer Systems Division 2101 W. Cypress Creek Road Fort Lauderdale, FL 33309 bill@ssd.harris.com or hcx1!bill@uunet.uu.net