Re: one-liner for characater replacement

On May 28, 11:45 pm, "James Giles" <jamesgi...@xxxxxxxxxxxxxxxx>
wrote:
analys...@xxxxxxxxxxx wrote:

...

OK I misunderstood you.  Are you saying that there are millions of
lines of Fortran 77 code that is too valuable to risk  either
switching  to another language or for that matter upgrade to f90, f95
f2003  etc.?

I don't know where you got the word "risk".  It's expensive to
convert to another language.  It's less expensive to incrementally
upgrade to upward compatible versions of the same language.
Some things *have* to be upgraded (to take advantage of new
parallel hardware features, for example).


In the business world I am familar with, software is "owned" by non-
programmers. many people have to sign off before changes are made to
production programs. Risk is a primary consideration to decide
whether or not to change a production program.

You can practically never make a "business case" to merely upgrade a
program to a more modern language or dialect if it doesn't offer new
functions that are seen to benefit the business somehow.

Under the above rules of the game, unless the Fortran world announces
that as of a certain date pure f77 compilers would no longer be
available, upgradation to more modern Fortrans would be a tough sell.

So what I'm saying is that the future users of new Fortran features
will almost all be in large institutions (commercial or not) that have
too much of an investment in Fortran to afford conversion to another
language.  And they probably *will* (slowly and incrementally)
upgrade to several of the more recently added features.


Have you observed whats missing from the picture?

Yes - you seem to have studiously avoided the topic of initiation of
new Fortran projects in the commercial world.

Why should a language undergo so many changes, some wise and some
otherwise, if its only possible future use is the continuation of a
code base that was developed in the 1970s at the latest?

And yes - I see Fortran's future as a high-powered excel sans the
Graphics and interfaces to databases etc.  Fortran will live (if at
all) purely as a prototyping/throw-away code/proof-of-concept
langauge.

I don't think it's ever been that.  Nor do I think it ever will be.
If the committee decides to target that usage it really will kill
the language.  There are plenty of languages already in that
category that do that sort of thing better than Fortran (any
flavor) ever has or ever will.

By the same token, there are plenty of languages that can take and
have taken (with a vengeance) the place of Fortran for large
numerically intesive software projects.


I'll need some time to formulate my reply. But let me turn the
question around - before we decide what features we would like a
language to have and not have - shouldn't we first have a strategic
view as to who would use the language for whst purposes?

But, for Fortran that's not a very small or well-defined set. Just

Are you seeing that the fortran user base - although a shadow of what
what it was in Fortran's heyday is still too diverse to be defined
tightly?

The diversity of the user base is practically identical to what it was in
the past.  The user base has not shrunk all that much (in absolute terms).

As always, everybody handwaves on this, but there is no hard data.

The rest of the computing has grown vastly.  That might give the illusion
that Fortran's user base is shrinking.  But I think most of the shrinkage
is in the form I already mentioned - small users have moved away
never to return.  Sort of the same set of people you seem to regard as
the *only* future users are the ones that were the first to go.


I believe your data are wrong in this regard. Are you familiar with
Les Hatton's surveys of Fortran usage?


Let me ask this question: Is it conceivable that a software rewrite
project in a fairly large organization (private or public) takes
place toward Fortran and not away from it?

My only concern is that some of the truly bizarre features that seem
to be coming from folks who have never worked in the private sector
might push the remaining corporate users over the edge and decimate
Fortran uasge further.

"Decimate" is probably about right since it means "reduce by ten
percent".  In absolute terms, that's probably about where Fortran is
compared to it high point.

--
J. Giles

"I conclude that there are two ways of constructing a software
design: One way is to make it so simple that there are obviously
no deficiencies and the other way is to make it so complicated
that there are no obvious deficiencies."   --  C. A. R. Hoare

.

Relevant Pages

  • Re: one-liner for characater replacement
    ... convert to another language. ... upgrade to upward compatible versions of the same language. ... So what I'm saying is that the future users of new Fortran features ... Are you seeing that the fortran user base - although a shadow of what ...
    (comp.lang.fortran)
  • Re: public comment on fortran 2008
    ... There are Fortran programmers who find even Fortran 90 too ... doesn't serve the user base to have the language always a moving ... language, it should be addressed by the standard with some urgency. ...
    (comp.lang.fortran)
  • Re: Surprise
    ... haven't seen any Fortran libraries that do that. ... Libraries are a huge part of the reason why one language would be ...
    (comp.lang.fortran)
  • Re: Help from fellow Fortran Users
    ... and Fortran in which he introduced the subject of Pascal, ... Pascal is a useful, high-level, portable language that is easy to use ... Inasmuch as his comments considered extensions for Pascal in the '70s, ... standard differs from ...
    (comp.lang.fortran)
  • Re: A petition to J3 apropos FORTRANs future
    ... Fortran folded like an accordion and gave ... The vast majority of programmers are not doing "scientific ... In theory, programmers would just "use the best language for the job", ... But there are reasons, both real and silly, why one ...
    (comp.lang.fortran)