Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.compilers > #486 > unrolled thread

Re: Have we reached the asymptotic plateau of innovation in programming la

Back to article view | Back to comp.compilers

Contents

  Re: Have we reached the asymptotic plateau of innovation in programming la SLK Systems <slkpg3@gmail.com> - 2012-03-09 14:36 -0500
    Re: Have we reached the asymptotic plateau of innovation in programming la "Derek M. Jones" <derek@_NOSPAM_knosof.co.uk> - 2012-03-10 00:24 +0000
      Re: Have we reached the asymptotic plateau of innovation in programming la Hans Aberg <haberg-news@telia.com> - 2012-03-10 15:06 +0100
    Re: Have we reached the asymptotic plateau of innovation in programming la glen herrmannsfeldt <gah@ugcs.caltech.edu> - 2012-03-12 05:49 +0000
      Re: Have we reached the asymptotic plateau of innovation in programming la Hans Aberg <haberg-news@telia.com> - 2012-03-13 00:10 +0100
        Re: Have we reached the asymptotic plateau of innovation in programming la Johann 'Myrkraverk' Oskarsson <johann@2ndquadrant.com> - 2012-06-07 18:15 +0000
          Re: Have we reached the asymptotic plateau of innovation in programming la Hans Aberg <haberg-news@telia.com> - 2012-06-10 01:12 +0200
    Re: Have we reached the asymptotic plateau of innovation in programming la "Jonathan Thornburg [remove -animal to reply]" <jthorn@astro.indiana-zebra.edu> - 2012-03-14 02:02 +0000

#486 — Re: Have we reached the asymptotic plateau of innovation in programming la

[Some of us who programmed in ANSI Standard Fortran 66 and PL/I 76
might take issue with the claim that C standardized procedural
programming. Standard high level procedural interfaces to operating
systems aren't new either, Burroughs had them in Algol in the 1960s.
-John]

By "significant developments" and "standardizing" I meant that for
programmers to have settled on 1 hardware/OS architecture and 1
programming language is something new, and good. The time to which you
refer was the wild west of both, with new kids on the block every
year.

Most subsequent languages have copied the C *syntax*. For example, i++
is now a fairly standard idiom. Certainly fortran is still in wide
use, but who copies its syntax? Well, I guess C did borrow from its
formatted I/O...

How many Burroughs machines are now in use? Point is that Wintel
overwhelmed all other architectures, not that it invented the system
call. I claim this is a good thing for ease of code portability and
reuse.

Why is this defacto standarization good? Because AMD can go from
nothing to a huge software base overnight. Because Apple can run
windows software. Because I can read and understand javascript without
having learned it. An if statement is an if statement, but settling on
a singe syntax for it is beneficial.

Not that I am complaining about the variety of programming languages.
Migrating from brand X to C++ or java is what keeps me in business.

http://slkpg.byethost7.com
[I'm not sure a software monoculture is an innovation, much less
an interesting one.  IBM faced antitrust suits in the 1960s and 70s
in both the US and Europe because their mainframes and OS/360 were
so dominant.  And as far as who copies Fortran syntax, every time
you write a=b+c or if(a>b)c=d, or function foo(x,y), you're
writing in Fortran. -John]

[toc] | [next] | [standalone]

#488

John,

> [I'm not sure a software monoculture is an innovation, much less
> an interesting one.  IBM faced antitrust suits in the 1960s and 70s

A language monoculture has benefits.

Greater people portability for one.

Fewer compilers needed (ok, this group's readers don't consider that a
benefit :-)

Everybody doing things the same way can also reduce faults.
The following experiment found a correlation between
percentage source code occurrences and developer knowledge
of binary operator precedence.
http://www.knosof.co.uk/dev-experiment/accu06.html

> in both the US and Europe because their mainframes and OS/360 were
> so dominant.  And as far as who copies Fortran syntax, every time
> you write a=b+c or if(a>b)c=d, or function foo(x,y), you're
> writing in Fortran. -John]

R copied Fortran syntax.  But that has been around for a while.
[R? -John]

[toc] | [prev] | [next] | [standalone]

#491

On 2012/03/10 01:24, John Levine wrote:

 > On 2012/03/10 01:24, Derek M. Jones wrote:
...
>> R copied Fortran syntax.  But that has been around for a while.
> [R? -John]

Some PLs are listed here:
   https://en.wikipedia.org/wiki/List_of_programming_languages

Most letters are already in use. :-)

Hans

[toc] | [prev] | [next] | [standalone]

#493

SLK Systems <slkpg3@gmail.com> wrote:

(snip)

> By "significant developments" and "standardizing" I meant that for
> programmers to have settled on 1 hardware/OS architecture and 1
> programming language is something new, and good. The time to which you
> refer was the wild west of both, with new kids on the block every
> year.

> Most subsequent languages have copied the C *syntax*. For example, i++
> is now a fairly standard idiom. Certainly fortran is still in wide
> use, but who copies its syntax? Well, I guess C did borrow from its
> formatted I/O...

As I understand it, Fortran introduced the multi-character variable
name, pretty much universal in programming languages, but
mathematicians haven't caught on yet.

> How many Burroughs machines are now in use? Point is that Wintel
> overwhelmed all other architectures, not that it invented the system
> call. I claim this is a good thing for ease of code portability and
> reuse.

Many systems have system calls that are similar enough to code
around in high-level languages. Things like file name convention
differences are also not so hard to work around.

> Why is this defacto standarization good? Because AMD can go from
> nothing to a huge software base overnight. Because Apple can run
> windows software. Because I can read and understand javascript without
> having learned it. An if statement is an if statement, but settling on
> a singe syntax for it is beneficial.

The logical IF seems to work about the same in many languages.
The exact syntax varies, such as the need for parentheses and
the use of THEN and ELSE keywords.

> Not that I am complaining about the variety of programming languages.
> Migrating from brand X to C++ or java is what keeps me in business.

(snip)
> [I'm not sure a software monoculture is an innovation, much less
> an interesting one.  IBM faced antitrust suits in the 1960s and 70s
> in both the US and Europe because their mainframes and OS/360 were
> so dominant.  And as far as who copies Fortran syntax, every time
> you write a=b+c or if(a>b)c=d, or function foo(x,y), you're
> writing in Fortran. -John]

Well, the arithmetic IF was Fortran original. Logical IF didn't come
until later. I believe not until Fortran IV, though I am not so sure
what did and didn't come in the different versions of Fortran II.

I would have thought that the logical IF came from ALGOL before
Fortran.

I have "History of Programming Languages" edited by Richard Wexelblat,
which might explain some of this. A good reference for the origins of
many programming language features.

-- glen
[You're right about the Algol stuff.  Dunno whether Hopper's commercial
translator predated Fortran wrt multi-character variables. -John]

[toc] | [prev] | [next] | [standalone]

#496

On 2012/03/12 06:49, glen herrmannsfeldt wrote:

> As I understand it, Fortran introduced the multi-character variable
> name, pretty much universal in programming languages, but
> mathematicians haven't caught on yet.

Math started out with sentences like "add the first unknown quantity to
the second unknown quantity", but over the centuries, it was eventually
shortened to expressions like "x + y".

So computing takes a step back in evolution, in part due to a limited
character set. But that is slowly changing in view of Unicode and STIX
fonts, which are already in use in proof assistants, for example, Isabelle.

Otherwise, there are a lot of multi-character symbols in use in math,
for example, standard functions. Users of TeX know that these are
typeset tighter than the corresponding variables. So $sin$ will be
typeset as three variables with extra space between them indicating
implicit multiplication, whereas to get the function name one would have
to use $\sin$.

Hans

[toc] | [prev] | [next] | [standalone]

#677

Hans Aberg <haberg-news@telia.com> writes:

> On 2012/03/12 06:49, glen herrmannsfeldt wrote:
>
>> As I understand it, Fortran introduced the multi-character variable
>> name, pretty much universal in programming languages, but
>> mathematicians haven't caught on yet.

Mathematicians have an implied operator between symbols.  You can't do
that with multi character names.  Not in a way that can make sense, to
me.

Operator overloading, if done incorrectly, is a maintenance nightmare so
I don't think there will ever be a language with user defined implied
operators.

  %let * be the implied operator

Then one could write

  K = a b + c^2 a k;

which would mean

  K = a * b + c^2 * a * k;

where ^ is "to the power of" and +, * the usual mathematical operators.

And while I find it easier to read the former, I doubt it'll be
universally recieved since slightly later in the program one might
encounter

  %let ^ be the implied operator

  L = c 2;

and L is suddenly c squared.


Maybe I'll try a toy language with implied operators one day.

--
   Johann Oskarsson                http://www.2ndquadrant.com/    |[]
   PostgreSQL Development, 24x7 Support, Training and Services  --+--
                                                                  |
   Blog: http://my.opera.com/myrkraverk/blog/
[It's been done. Probably before you were born. -John]

[toc] | [prev] | [next] | [standalone]

#687

On 2012/06/07 20:15, Johann 'Myrkraverk' Oskarsson wrote:
> Mathematicians have an implied operator between symbols.  You can't do
> that with multi character names.  Not in a way that can make sense, to
> me.

In mathematics, the spacing is different - check out TeX. For example,
the name "sin" is typeset somewhat tighter than implicit multiplication
of the three variables "s", "i" and "n". Unicode has various spaces that
might be used.

Hans

[toc] | [prev] | [next] | [standalone]

#501

SLK Systems <slkpg3@gmail.com> wrote:
> By "significant developments" and "standardizing" I meant that for
> programmers to have settled on 1 hardware/OS architecture and 1
> programming language is something new, and good. [[...]]

I see no sign that programmers have settled on 1 programming language.
Fortran, C, and C++ are certainly in widespread use, but then again,
so are SQL, Lisp, Perl, Mathematica, Matlab, and TCL (to name only a
few others that *don't* have syntax-and-semantics-similar-to-C).


> Point is that Wintel
> overwhelmed all other architectures, [[...]]

I see no sign that Wintel has overwhelemed all other architectures.
It's widely used for desktops, but outside of that space it's rare.
E.g., what fraction of cellphones use it?  What fraction of the TOP500
list of supercomputers?  What fraction of the code in a modern
fly-by-wire airliner is running on a Wintel architecture?  As a final
example, a few years ago I read that the world's #1 most common
computer peripheral device was a fuel-injected carburetor; I rather
doubt it runs on either the software or the hardware part of "Wintel".

--
-- "Jonathan Thornburg [remove -animal to reply]" <jthorn@astro.indiana-zebra.edu>
   Dept of Astronomy, Indiana University, Bloomington, Indiana, USA
   "C++ is to programming as sex is to reproduction. Better ways might
    technically exist but they're not nearly as much fun." -- Nikolai Irgens

[toc] | [prev] | [standalone]

Back to top | Article view | comp.compilers

csiph-web