Is C/C++ the future?

[srctran@world.std.com (Gregory Aharonian)]

   The October 1994 issue of UPSIDE (a yuppy kind of entrepreneurial magazine
popular in Silicon Valley) has an article on one of the roundtable discussions
of industry leaders, in this case predicting what technology will be like in
the year 2000.

   On the panel were Gordon Bell (father of the VAX), Robert Lucky (VP of
applied research at Bellcore), Nathan Myhrvold (VP of advanced technology at
Microsoft), Jef Raskin (one father of the Macintosh GUI), and John Warnock
(CEO of Adobe).

   One of the questions was: "What will be the dominant programming language?"
with the following responses:

BELL:  Visual Basic, Mosaic markup language, C++, Cobol, Fortran, Telescript
LUCKY: C++. There will be too much investment in code to change this.
MYHRVOLD: C and C++
RASKIN: BASIC
WARNOCK: C

   Admittedly a very small sample, tho from representatives of companies with
a much bigger influence in determining the future of programming than anyone
in the Ada Mandated world, especially in light of industry trends.

  As an example of what Lucky is referring to, Taligent (an IBM/Apple/HP joint
effort) is releasing this summer a developer release of the TAE (Taligent
Application Environment) - a collection of 100 frameworks, comprising over
2000 C++ classes and over 30,000 member functions, and who knows how many
million lines of C++ code.  This dwarfs anything in the non-Mandated part of
the Ada world.  Along with Taligent, Microsoft and Sun (whose OpenStep has
already been shipped to 100,000 users - larger than the installed Ada base)
are also coming out similarly large and complex C++ systems that will be 
adopted by large sectors of the corporate software world. Who will want to
adopt other languages once companies start investing in these systems?  Why
switch away from these industry standards?  Just to get a compiler that stops
when it encounters an error?

(And guess who funded tons of the academic R&D that is being used on these
large C++ environments.  ARPA, and it still is so funding, apparently in
cahoots with the Air Force [KBSA] and the SEI. So much for military loyalty.
The Ada9X academic campaign is a complete waste of time and money because
ARPA already has cornered DoD influence of the academic world and ARPA has no
intention of allowing any other branch of the DoD to seriously encroach on 
their turf with Ada).

   Convert? Certainly not any customer of TeamAda member IBM.  IBM's future 
OO plans will be based on its' VisualAge and VisualGen.  To supplement these
tools, IBM intends to get back into the compiler business in a big way with
C++, object oriented Cobol and perhaps even object oriented PL/1.  Also
coming are Smalltalk, object oriented REXX, Visual RPG and Microsoft's Visual
Basic.  IBM intends to deliver fully compatible versions of most compilers
across all its strategic systems, which now includes OS/2, AIX, OS/400 and
MVS.   BUT NOT ADA!!!!!!!  Imagine IBM prefering an object oriented REXX over
Ada.   Having milked all of the Ada pork it can out of the DoD, why should IBM
invest in a dead-end language?  Why should anyone, if as the DualUse plan
shows, even the DoD is unwilling to invest in commercializing Ada?

   Don't believe me?  Well, someone is giving a very rational lecture at the
upcoming weeklong C++ WORLD conference (Austin, TX, 11/14-11/18) on rules of
thumb for managing industrial-strength object-oriented C++ projects.  It
will probably be full of rational tips for using some company's products
as a rationale for using C++ on these large OO projects that are dominating
industry.  Obviously this rational lecture reflects a rational trend by
rational corporate software developers, many of whom will be using either
Taligent's, Microsoft's, or Sun's environments and need strong C++ tools,
rationally.

   Nothing DISA and the ASA is doing with its DualUse plans will have any
effect (assuming they care to measure) whatsoever on industry use of Ada.
All their plans will do is to further entrench Ada as a niche language for
those very large, critical systems that are too rare to be a basis for a
thriving industry.  Other than for that need, both outside and INSIDE the
DoD, Ada won't be used, no matter how many meaningless and conflicting
mandates the DoD issues (like Mosemann's AI memo that strategically ommitted
mentioning Ada, probably the inspiration for the Defense Science Board not
to cover Ada in its study).


Greg Aharonian