Why Ada has seven years to thrive or die

[world!srctran@uunet.uu.net (Gregory Aharonian)]

     A recent article on object oriented technology illustrates why the Ada
community only has a short amount of time to achieve Ada acceptance, or
seeing the language relegated to a niche role like Forth and Lisp.

     The article is titled "Users making plans for object technology
applications" and is in the August 30 issue of Network World, page 45.  In
the article, there is a table on percentage of users planning to adopt
object oriented technology, based on a survey of 186 managers at some of
the largest 1000 companies.  The Table is as follows:

		   PRECENTAGE OF USERS ADOPTING
		    OBJECT ORIENTED TECHNOLOGY

			1994		 2%
			1995		11%
			1996		34%
			1997		53%
			1998		71%
			1999		80%
			2000		88%

      Assuming that this survey reflects general industry future plans (and
no one in the Mandated world has any equivalent data), and given that 88%
pretty much closes up a market, it would appear that the Ada community has
seven years to achieve significant market share before it is impossible for
Ada to play any significant role in the software engineering and MIS worlds,
and therefore to be able to help the DoD meet its software requirements.

      If by 2000 Ada is still at its present level of about 1% market share
outside the Mandated World, then the Ada community will have failed.  Such
a market share level will not attract many programmers, many companies to
develop tools, and many third party types to provide products and services
that drive other languages like C++ and Smalltalk.

      Unfortunately, in my humble :-) opinion, I do not think Ada will be
able to have significant market share won for it for two reasons.

	1)  Current Ada tools are two to three years behind those for
	    other languages.  Other than the great Rational tools, most
	    of the Ada compiler environments and libraries are two to
	    three years behind stuff I see for C/C++ and Smalltalk, and
	    even Cobol.  Ada research tools, such as most of the STARS
	    stuff, is as far behind, if not farther.

	    In the software world, a two to three year lead is very
	    difficult to overcome, especially given the cutthroat 
	    competition now going on in the C/C++ world, and the total
	    inexperience of Ada compilers of competing where people are
	    free to choose.


	2)  Ada83 is a great language.  It failed to get much acceptance
	    not because of any deficiencies maybe corrected in Ada9X,
	    but because the Mandated world either didn't care to promote
	    the language outside the Mandated world or didn't know how.
	    Ada9X is a false promise, because the people who botched the
	    job of handling Ada83 (whoever they are since no one wants to
	    accept responsibilty) - those who didn't - still don't and
	    still won't.

    So for those of you with overhead to goto Tri-Ada, keep in mind that
Ada has a window time shorter than the age of Ada to achieve significant
market share, or be dismissed.  And think to yourself, is a community that
displays over 40 booths at its own internal trade shows, while only one
or two booths at public trade shows, is this a community that has any 
chance of actually having to compete where people are free to choose?
-- 
**************************************************************************
 Greg Aharonian                                      srctran@world.std.com
 Source Translation & Optimization                            617-489-3727
 P.O. Box 404, Belmont, MA 02178