Sociology of Programming Language Use

Sociology of Programming Language Use

[Richard Pattis]

I thought that the following post (From: eric@tfs.com (Eric Smith)) might
be of interest to those discussing why Ada hasn't "caught on" as much as
it "should".

---------------------------------------------------------------------

C++ seems to be gaining "critical mass" for the same reason Pascal did
in the late 1970's.  Pascal became widely available at low cost when
UCSD published their version at a very low price.  It rapidly gained
popularity during the rest of the 1970's and early 1980's.  Borland's
Turbo Pascal gave it an additional kick.

The same thing happened to C a few years after Pascal.  There were lots
of companies publishing low cost high quality compilers for it so it
gained critical mass.

The same thing has been happening with C++ recently, and that is IMHO
the real reason why C++ is more popular than Eiffel.  When people
choose a new programming language, it has nothing to do with what
language they used before.

Programmers satisfied with their present programming language are less
likely to try to find better alternatives than those who are
dissatisfied.  So a C programmer who wants to find a better language
would probably have a bias against C++ in favor of getting further away
from C.

However, all such biases probably have much less effect than they get
credit for.  The real motivating forces might be almost purely
economic.  Programmers want to use whatever language lets them get the
most work done with the least effort, provided they don't have to pay a
lot of money to get started with it.

Suppose I could use Eiffel at work but just C++ at home.  If I spend
20% of my time programming at home on personal projects, doesn't it
make sense that I would want to use the same language at work, just to
avoid daily switching between them?  Especially if that is the only way
I could use the same class libraries etc.?

A 486/33 with 64 megabytes of ram costs around $5000.  A good C++
compiler for it usually adds less than a few hundred dollars to that.
I have been programming for more than 20 years, but have never had
access to any computer more powerful than that.  Not even the
mainframes I used to program were anywhere near that powerful.

Compare that to what is available with Eiffel.  I don't know of any
system of such low cost that will even run Eiffel.  And I suspect the
Eiffel compiler alone would probably cost a large fraction of that.
It would be wonderful to have Eiffel as readily available as C++, but
until then, it's more like a fairy tale than a real product.

Meanwhile C++, even though less elegant than Eiffel, is genuine magic.
In a race between genuine magic and a fairy tale, the genuine magic
usually wins.




-- 
------------------------------------------------------------------------------
  Richard E. Pattis			"Programming languages are like
  Department of Computer Science	 pizzas - they come in only "too"
    and Engineering			 sizes: too big and too small."

Re: Sociology of Programming Language Use

[mcsun!uknet!ukc!yorkohm!minster!mjl-b]

In article <1991Oct21.175825.10745@beaver.cs.washington.edu> pattis@cs.washingt
on.edu (Richard Pattis) writes:
>I thought that the following post (From: eric@tfs.com (Eric Smith)) might
>be of interest to those discussing why Ada hasn't "caught on" as much as
>it "should".
>
>---------------------------------------------------------------------
>
>C++ seems to be gaining "critical mass" for the same reason Pascal did
>in the late 1970's.  Pascal became widely available at low cost when
>UCSD published their version at a very low price.  It rapidly gained
>popularity during the rest of the 1970's and early 1980's.  Borland's
>Turbo Pascal gave it an additional kick.
>
>The same thing happened to C a few years after Pascal.  There were lots
>of companies publishing low cost high quality compilers for it so it
>gained critical mass.
>
>The same thing has been happening with C++ recently, and that is IMHO
>the real reason why C++ is more popular than Eiffel.  When people
>choose a new programming language, it has nothing to do with what
>language they used before.
>
>Programmers satisfied with their present programming language are less
>likely to try to find better alternatives than those who are
>dissatisfied.  So a C programmer who wants to find a better language
>would probably have a bias against C++ in favor of getting further away
>from C.

Indeed, if you read the US Air-Force report on the comparison of Ada and
C++, Ada scores over C++ in almost every catagory. Two of the catagories
that Ada doesnt't win are "popularity" and "acceptance".

It seems that Ada is up against "brand loyalty" here. The very people who
admitted that Ada was better than C++ in each of the catagories mentioned in
the report also said that they were happier with C++!

>  Richard E. Pattis			"Programming languages are like
>  Department of Computer Science	 pizzas - they come in only "too"
>    and Engineering			 sizes: too big and too small."

| Mathew Lodge              | "Baldrick, are the words 'I have a cunning plan' 
|
| mjl-b@minster.york.ac.uk  |  heading towards this conversation with ill      
|
| Summer: lodge%alsys@uknet |  deserved confidence?" -- Blackadder III         
|