Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!yale!cmcl2!lanl!jlg
From: jlg@lanl.gov (Jim Giles)
Newsgroups: comp.society.futures
Subject: Re: (pssst...fortran?)
Message-ID: <62905@lanl.gov>
Date: 13 Sep 90 21:54:45 GMT
References: <1990Sep10.193130.20463@rodan.acs.syr.edu> <62688@lanl.gov> <1990Sep12.231240.2922@rodan.acs.syr.edu>
Organization: Los Alamos Natl Lab, Los Alamos, N.M.
Lines: 85

In article <1990Sep12.231240.2922@rodan.acs.syr.edu>, isr@rodan.acs.syr.edu (Michael S. Schechter - ISR group account) writes:
> In article <62688@lanl.gov> jlg@lanl.gov (Jim Giles) writes:
> >From article <1990Sep10.193130.20463@rodan.acs.syr.edu>, by isr@rodan.acs.syr.edu (Michael S. Schechter - ISR group account):
> >To be sure, ther are _superficial_ similarities between Pascal and
> >C.  But, there are superficial similarities between _any_two_ procedural
> >languages you care to mention.
> The programming that will be done will be _superficial_. No lists, [...]

Lists are quite common in Fortran.  As they are in _all_ other popular
procedural languages.  Each language presents different problems when
implementing lists.  C is not obviously superior in this respect (or
perhaps only to novices).

> [...]  No
> pointers,

C pointers are semantically _equivalent_ to one dimensional Fortran
array indexing with bounds checking turned off.  Aside from C allowing
a slightly more compact notation, there is no difference.  Since most
Fortrans allow bounds checking to be turned off, Fortran is not inferior
to C in this respect.

Having said that, I should point out that C pointers are _MUCH_ worse
than Pascal pointers.  The only operations that pointers should allow
are assignment and dereferencing.  If you want array indexing, use arrays.
Casts of pointer types should be completely impossible.

> [...] no searching, [...]

Huh?  Most searching algorithms I ever heard of were originally written
for Algol/Fortran applications (or predated programming altogether).  All
such algorithms work in all popular procedural languages.

> [...] no nothing except real-time control. [...]

Most real-time control is still done in assembly.  Fortran is mostly
for numerical/scientific work - for which it is _still_ probably the
best available language.

> [...]
> >Why do you want him to switch from a useful language to a lower level
> >and harder to use one?
> 
> Because it will break my heart to see displays like
>       M A I N   M E N U
>       -----------------
>                1)   Blah
>                2)   Blah
>                3)   Blah
>         Enter your choice:
> on the Macintosh. And while I haven't yet seen what the 
> FORTRANS for it are like, with the way that everything is done
> with handles and structures, I can see programming the user interface
> being a nightmare if FORTRAN is used. 

Many of the best user interfaces I know of were in Fortran, assembly,
PL/1, or even Cobol.  Most of the really _bad_ user interfaces I know
of are written in C for UNIX.  User interfaces can be written (or mis-
written) in just about any language.  Their quality depends on the
design of the interface, not the language used to implement it.  Of
course, the language chosen _could_ make a difference in the efficiency,
maintainability, and cost of developement for the software - all of which
C is relatively dismal at.  Stick to Modula 2 or Pascal for this kind
of stuff.

(By the way, I don't understand your example.  The Macintosh _does_ work
mostly with menus - even for things which would be better as commands
types from the keyboard.  The Macintosh tends to be novice friendly and
inhibits experts.  I personally don't like it much.)

> [...]
> >of the problems with Fortran are errors of omission - that is, Fortran
> >lacks features which sooner or later (one hopes) will be added.  On the
> >other hand, C suffers mostly from errors of commission - that is, C has
> The features aren't there NOW. That's why I want him to use something
> else, whether it's C or Pascal, isn't THAT important

Yes, if you need features that Fortran doesn't have, by all means switch
to Pascal (or Modula, or Ada, or Lisp, or ...).  If you don't need such
features and Fortran is familiar to you, stick with it.  Either way, I see
no (non-political) reason to even consider C.  The fact is (as I said
about errors of omission vs. errors of commission), Fortran has a future,
C only has a past.  And this _IS_ comp.society.futures after all.

J. Giles