[comp.lang.fortran] Fortran 90 Definition

cowie-james@cs.yale.edu (James Cowie) (06/28/91)

Alright, with all this discussion of the upcoming NAG release, 
does anyone have a pointer to a (published or) online version of the
F90 _definition_, official or unofficial?

--jim (cowie@cs.yale.edu)

francis@hanauma.stanford.edu (Francis Muir) (06/28/91)

James Cowie writes:

  Does anyone have a pointer to a (published or) online version of the
  F90 _definition_, official or unofficial?

I have open in front of me:

	PROGRAMMER'S GUIDE TO FORTRAN 90

	Walter S. Brainerd
	Charles H. Goldberg
	Jeanne C. Adams

	InterText Publications
	McGraw-Hill Book Compny

The authors are members of the relevent ANSI and ISO committees, either
or both. At $39.95 a pop it is not cheap (its a paperback), but it does
seem to be boringly complete. A supersubset of the authors also has a
book:

	THE FORTRAN 90 HANDBOOK

which I do not know except as a reference.

							Fido

wsb@boise.Eng.Sun.COM (Walt Brainerd) (06/28/91)

In article <1991Jun27.184733.24437@leland.Stanford.EDU>, francis@hanauma.stanford.edu (Francis Muir) writes:
> James Cowie writes:
> 
>   Does anyone have a pointer to a (published or) online version of the
>   F90 _definition_, official or unofficial?
> 
> I have open in front of me:
> 
> 	PROGRAMMER'S GUIDE TO FORTRAN 90
> 
> 	Walter S. Brainerd
> 	Charles H. Goldberg
> 	Jeanne C. Adams
> 
> 	InterText Publications
> 	McGraw-Hill Book Compny
> 
> The authors are members of the relevent ANSI and ISO committees, either
> or both. At $39.95 a pop it is not cheap (its a paperback), but it does
> seem to be boringly complete. A supersubset of the authors also has a
> book:
> 
> 	THE FORTRAN 90 HANDBOOK
> 
> which I do not know except as a reference.
> 
> 							Fido

I hate to suggest that anyone NOT buy our book, but it really is not
a complete description of Fortran 90.  It usually discusses only one
"best" way to do most things (e.g., of the millions of ways to declare
an array of character strings of kind KANJI, one is discussed and used
consistently).

The spec for the NAG implementation is the Fortran 90 standard.  They
say it is no more and no less.

Unfortunately, the HANDBOOK probably won't be out until near the end
of the year.  It is an attempt to describe the complete language, but
done in a more understandable way; this takes a lot of careful hard work!

While I'm on, I have been a little surprised by all the discussion
of the execution efficiency of the NAG implementation.  In their
own announcement, they say the objective is not to provide a
highly optimized environment, but one in which code development
and education can take place; Mike Metcalf's preliminary experiences
reported here seem to indicate it will be pretty good at that.  And
as someone did say, perhaps it can be transformed into something that
produces very efficient code.
--
Walt Brainerd               Sun Microsystems, Inc.
walt.brainerd@eng.sun.com   MS MTV 5-40
                            Mountain View, CA 94043
                            415/336-5991