Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!linus!philabs!cmcl2!seismo!hao!noao!terak!doug
From: doug@terak.UUCP (Doug Pardee)
Newsgroups: net.lang
Subject: Re: Language use
Message-ID: <1048@terak.UUCP>
Date: Thu, 20-Feb-86 18:39:48 EST
Article-I.D.: terak.1048
Posted: Thu Feb 20 18:39:48 1986
Date-Received: Mon, 24-Feb-86 04:57:44 EST
References: <2337@burdvax.UUCP>
Distribution: net
Organization: Calcomp Display Products Division, Scottsdale, AZ, USA
Lines: 51

> I am looking for a good reference that will tell me what percent of
> the programs written are in each programming language.
> 
> I have heard that 75% of all programs are written in COBOL, 24% are
> written in FORTRAN, and 1% are written in other languages, but
> I would like a reference.

There are too many variables to give anything that even approximates
an answer.  For example,

Do you want to know
  a)  The percentage of programs ever written, or
  b)  The percentage of programs written for computers still in use, or
  c)  The percentage of programs still being used themselves, or
  d)  The percentage of programs still being maintained, or
  e)  The percentage of programs currently being written?
  
Do you count copies on different systems individually?  Or is the
VAX/Unix 4.2 C compiler on this system not count if you've already
counted it on another VAX?

Do you count one, two, or three different programs for "lint" running on
VAX, PDP/11, and Motorola 68000?

Do you count different versions as the same program?  Is Unix 4.2
and System V the same program?  What about the two versions of PCC
we use here:  one generates VAX code, one generates NS32000 code?

Do you count a program that is run once and then discarded to be the
same as a program like "ls" that is run many times a day?

Do you count a shell script as a program?

Getting even more esoteric:  do you count anything programmable, or
is there some specific definition of a "program" that you're thinking
of?  Does the microcode inside a 68000 count as a program?  Or maybe
it counts as a hundred programs, one for each instruction?  What about
the cooking instructions on my microwave oven?

Getting serious again...  the figures that were suggested above are
probably based on "customer-developed mainframe applications" -- that
is, the programs that run on big IBM/Honeywell/CDC/etc computers;
programs that were written by programmers at the computer site; programs
whose results are supposed to benefit non-computer people.

That estimate was probably made 15 years ago or more; there's no way
that anyone could make any sort of rational estimate now.  By the time
you specify enough parameters to make it measurable, the result would
be meaningless.
-- 
Doug Pardee -- CalComp -- {hardy,savax,seismo,decvax,ihnp4}!terak!doug