[net.micro] Languages Survey

godwin@icse.uci.edu (Dave Godwin) (11/09/85)

Hi folks.
	This survey bit took a little longer than expected; replies have
continued to trickle in.  In total, I received responses from 74 locations.
These locations range from university cs departments to university research
teams to industry r&d to 'this is what we use at work every day' from large
banks and other typical sites.  The results were not ( to me ) surprising,
but I think that many of you will be, so let's talk about it.
	Keep in mind that these replies are mostly from locations working
on the cutting edge of our industry; the subject was 'what languages are
you working with now', not 'what have you been maintaining recently'.  The
survey was sent out to bboards discussing Pascal, Ada, Software engineering,
micro computers, and on site at various Rockwell, Hughes, Lockheed, Navy
and Air Force plants.  ( This may not have covered all of the population,
but it do come close. )


language	percentage use
-------------------------------
Pascal    	36.45
C	     	25.65
FORTRAN    	13.5
Modula-2   	5.40
Forth   	5.40
COBOL   	4.05
Ada   		4.05
miscillaneous	~5.00	( This includes all sorts of strange stuff )

	The miscillaneous catagory is filled with languages used at only
one location reported, and then that language is usually not highly used.
Languages mentioned include Snobol 4, Simula, PL/I, and assorted assembly
and macro stuff. 

	Thanks much to the folks at Columbia, Rutgers, UC Berkely and a certain
Hughes location for the nice, extensive replies.  These were most helpfull.

		Dave Godwin
		University of California, Irvine
		godwin@icse.uci.edu

p.s. ( There was one response I found rather amusing.  There is a large group
	at Rockwell.  Which site, and which project they are working on is not
	important.  The project contract says that the software source that
	gets turned in to the buyer must be written in either FORTRAN or the
	assembly language for the given machine.  The engineers in this group
	prefer to do things in a more comfortable fashion.  They write 
	everything in Pascal, which makes thing much easier for them.  They
	compile their Pascal programs, and test them, run them, get satisfied
	with them.  They then run the Pascal code files through the dis-
	assembler, and send the buyer a nice long assembly listing like the
	contract asked for.  Everybody winds up happy. :-) )

BEBO%SLACVM.BITNET@wiscvm.wisc.edu (11/12/85)

Date: 11 November 1985, 13:18:56 PST
From: Bebo White                (415) 854-3300 x2907 BEBO     at SLACVM
To:   INFO-ADA at MIT-MC.ARPA, INFO-MICRO at BRL-VGR.ARPA
      INFO-PASCAL at BRL-VOC.ARPA
Subject: Languages Survey

This note is a reaction to Dave Godwin's Languages Survey which recently
appeared on this net - I would respond directly to Dave, but we don't
have access to an EDUNET gateway.

I, like many others I'm sure, was pleased with the outcome of the survey
which reinforced our interest in Pascal. One can only hope that the survey
was not biased by a large number of Pascal "bigots."

I assume that this survey represents language usage over a wide range
of computer manufacturers, including IBM. If that assumption is true
and Pascal does enjoy the popularity that the survey indicates, then
why is it not viewed as a strategic language by IBM?

IBM's Pascal/VS is a Program Offering with significantly less status
than VSFORTRAN, a Program Product. The SHARE FORTRAN Project is a
large and active group with apparently considerable influence on IBM.
On the other hand, the SHARE Pascal Project barely scrapes along with
virtually all of its requirements to IBM designated as Future
Objectives (i.e., expect it when you see it). The IBM Santa Teresa
Lab has a large FORTRAN team and a skeleton Pascal support staff.

Does anyone have any ideas about what may explain this apparent
incongruity?

Thanks.

steve@BRL.ARPA (Stephen Wolff) (11/12/85)

>	Does anyone have any ideas about what may explain this apparent
>	incongruity?

You don't suppose, do you, that there is just the tiniest possibility
that the design of the survey might have introduced some bias?

kvancamp@pica-lca.arpa (LCWSL) (11/12/85)

Bebo White writes this about the languages survey conducted by Dave Godwin:

>I assume that this survey represents language usage over a wide range
>of computer manufacturers, including IBM. If that assumption is true
>and Pascal does enjoy the popularity that the survey indicates, then
>why is it not viewed as a strategic language by IBM?

I tend to question Dave's results, in terms of adequately representing
commercial users. Dave, would it be difficult to break those 74 locations
down into how many were connected to universities?  My own experience is
certainly not unbiased, but in my experience in DOD I've found a LOT more
Fortran programmers than Pascal. It's my feeling that the Pascal users seem
to be more concentrated at universities. I'd just like to mention one
observation, by the way. The copy I received of Dave's results (through
info-micro) was also forwarded to info-pascal and info-ada. Dave probably
doesn't know where his replies all originated from, but I think his request
may have reached more Pascal programmers by the nature of the way it was
distributed.

		  --Ken Van Camp <kvancamp@pica-lca.arpa>
		Army Armament Research & Development Center
		       SMCAR-FSA-E   Building 350-S
			  Dover, NJ  07801-5001
		       (201)724-3675  (AV)880-3675

	The opinions expressed herein do not necessarily reflect the 
			opinions of my employer.