Path: utzoo!attcan!uunet!lll-winken!ames!haven!umbc3!mbph!hybl
From: hybl@mbph.UUCP (Albert Hybl  Dept of Biophysics  SM)
Newsgroups: comp.lang.fortran
Subject: An exercise in futility
Keywords: Fortran_standards resemble_a Demolition_derby
Message-ID: <587@mbph.UUCP>
Date: 16 Jan 89 21:54:30 GMT
Organization: University of Maryland, School of Medicine, Baltimore, MD 21201
Lines: 95


The United States is about to spend some 4 gigabuck mapping the human
genome.  This is one of the best uses for which our tax dollars
can be used.  Already the knowledge gained by genetic mapping is
contributing a better understanding of emphysema, hypertension, and
many genetic diseases.  Progress toward treatment and cures are
just around the corner or even in clinical trial.

The mapping project will generate long strings of 'A's, 'C's,
'G's and 'T's.  Sequences can be 1 to 1 million or more nucleic acids
in length.  When searching for a promoter region on a gene
or asking which restriction enzymes will cut a DNA sequence,
a researcher needs to locate the position of a short recognition
sequence on the long DNA sequence.  He could, for example, use
the intrinsic INDEX function in an expression like:
     J = INDEX(DNA,'CGGCCG')

At first glance FORTRAN appears to be a useful language to use for
such projects.  However, because it is so full of quirkish ambiguities
and because it is no longer a portable language, using it is
engaging in an exercise in futility.

In message <3704@s.cc.purdue.edu> ags@s.cc.purdue.edu (Dave Seaman,
Purdue University) claims:
> ...THEY ARE ENFORCING THE STANDARD  ...

Consider the following program:
      PROGRAM FUTILE
*  Repeat this program after increasing the value of K until the program fails.
      PARAMETER (K=24)
      CHARACTER AND(K)*1
      CHARACTER DNA*(K)
      EQUIVALENCE (DNA,AND)
*
      DNA(K-23:K) = 'An exercise in futility!'
      PRINT *,K
      PRINT *,(AND(J+K-24),J=1,24)
      STOP
      END

What does the ANSI X3.9-1978 standard say are the maximum allowable
sizes for the character array AND(K)*1 and the character variable
DNA*(K)?  Are the maximum allowable sizes the same?  Does it allow
the equivalence statement?  I believe that most of the standards
committee intended that values should be the same and the limit
determined by the amount of available core in the processor executing
the job.  The equivalence is legal so long as the variables are legal.

Rush to you vendor's FORTRAN Language Reference Coloring Book,
what does it say and in what color does it say it?  Why should
DNA*256 be LEGAL on most systems but ILLEGAL when using RM fortran on
an IBM PC/AT or the SVS fortran on an AT&T 3B1?  Why is DNA*501
illegal when using CMS fortran on an IBM4341?  Why is DNA*513
illegal when using FTN fortran on an UNISYS 1100?  Why is DNA*1001
illegal when using **X Fortran on a VAX VMS machine?  Of the compilers
that I have used, only the XLA+ fortran on an AT&T 3B2 machine
does what I think the standard specifies.  DNA*750000 works!
When I increased the value of K to 1000000 and above I obtained
to following diagnostics: 
(k=1000000) Ld fatal: Cannot complete output file a.out. Write error
(K=1097152) Assembly: Trouble Writing; probably out of temp-file space

Surely a FORTRAN language validation suite would have prevented such
abominable implementations of the standard.  Not even the most
convoluted interpretation of sections 1.3.2.(4) and 1.4 can justify such
capriciousness.  The purpose of the Fortran ANSI X3.9 1978 standard
"is to promote portability of FORTRAN programs for use on a variety
of data processing systems."  Pray tell, based on the above examples how
can anyone say that the maximum length for a string variable is
predicable or portable?

Steve Weintraub has shown that the USENET can be used to rapidly
discover ambiguities in a language standard.  Discussions in this
newsgroup can propose corrective language.  The real problem is how
to incorporate the new language into the standard.

Clearly sections 1.3.2.(4) and part of section 1.4 MUST
be deleted!  The  standards making body should produce a
"Cumulative Annual Pocket Part" that is published in FORTRAN forum
and other appropriate media.  There must be an adjudication 
mechanism.  The standards organization must have a registered
trademark for the FORTRAN Language.  A Fortran Validation Suite must
be created, updated and revised annually.  Before a compiler
is allowed to use the trademark, it must pass the validation tests.
Unless a Fortran Standardizing Organization is willing and able to
legally enforce the Standard, this discussion will have been nothing
but an exercise in futility.

----------------------------------------------------------------------
Albert Hybl, PhD.              Office UUCP: uunet!mimsy!mbph!hybl
Department of Biophysics       Home   UUCP: uunet!mimsy!mbph!hybl!ah
University of Maryland                CoSy: ahybl
School of Medicine
Baltimore, MD  21201                 Phone: (301) 328-7940 (Office)
----------------------------------------------------------------------