Re: Datapro announces survey of ObjectOrient languages

Re: Datapro announces survey of ObjectOrient languages

[cis.ohio-state.edu!magnus.acs.ohio-state.edu!csn!hellgate.utah.edu!dog.ee]

In article <SRCTRAN.93Apr1131550@world.std.com> srctran@world.std.com (Gregory 
Aharonian) writes:
>    From an article in April 1993 SunObserver:
>
>    The US and European object oriented development market will grow from
>$865 million to $4 billion by 1997, but the number of OO programming
>languages will shrink in the process, according to a new Datapro survey.
>About 80 languages now exist, but an industry shakeout is on the way, and
>at present, C++ and Smalltalk are ahead of the pack, said Datapro's Frank
>Teti.

The rest of this article I chose not to include since I think that this 
initial paragraph is sufficient to make my point.

Greg, often times I find your postings to be insightful and valuable.  But
quite often I find them to be much less insightful, this posting is one of
those instances.

The comparative language statistics that you reference are all statistics
on "present" usage of "OO programming languages".  Considering that much of
the OO community (actually, I would be willing to argue that most of the
OO community) does not consider Ada an OO programming language, you have
to admit that the absence of Ada-83 in this arena is not surprising.  
When Ada-9X becomes available then perhaps the absence of Ada in the OO
community would carry a bit more credibility.

Just my two cents worth.

Regards,
Steve Case

Re: Datapro announces survey of ObjectOrient languages

[Gregory Aharonian]

>The comparative language statistics that you reference are all statistics
>on "present" usage of "OO programming languages".  Considering that much of
>the OO community (actually, I would be willing to argue that most of the
>OO community) does not consider Ada an OO programming language, you have
>to admit that the absence of Ada-83 in this arena is not surprising.  
>When Ada-9X becomes available then perhaps the absence of Ada in the OO
>community would carry a bit more credibility.

Steve,
	Thank you for the interesting comment, which like my initial comment
has been and is, and probably is true for some time.  However, consider the
following points.   Ada83 does have some object oriented capabilities (if
you refer to the taxpayer funded literature on Ada83, you will see a fair
amount of claims for this), with generics maybe not great capabilities but
at the time of its introduction enough to be involved with the OO community.
With some of the Ada systems like DRAGOON which bring more OO stuff to Ada83,
there is certainly enough grounds for the Ada community now to participate
in the OO world.
	But they don't.  Almost every OO conference and trade show has
absolutely no Ada presence, either Ada83 or Ada9X.  There are few Ada
articles in the OO journals, and from talking with contacts at many major
computer publications, little activity on the part of the Ada community in
contacting the trade press to make the case for Ada.
	So we both agree that Ada83 has been poorly evangelized by the
Defense community to date and that's why it does not appear in such reviews
as the Datapro survey.

	You then make the very wishful hope that "When Ada-9X becomes
available, then perhaps the absence of Ada in the OO community would carry
a bit more credibility".  Let me explain why this unfortunate circumstance
will come to pass, mainly for considerations that you as a student have yet
to encounter in the real world.
	I'll concede that Ada9X wil probably shape up tobe a great OO
language.  When it comes to the technical competence of the Ada9X committee
headed by Christine Anderson, I have only praise.  Thus it will be easy,
if anyone cares to, to make the case for Ada9X in the non-Mandated world.
However, under current DoD leadership, this will never happen, for the
following reasons.
	Anyone with any business sense, and therefore concluding they should
be responsible for spending Ada tax dollars, knows that many companies around
the country and making their committments to objected oriented programming
over the next few years, and as now as we speak (or email).  It will probably
take another few years for most companies to select on their OO technology,
and begin making investments at their companies.
	However, once companies make their investments, they will be very
reluctant to switch to a new language or methodology.  Unlike the people in
the Pentagon, who have ample tax dollars to start and stop software initiatives
and change them in mind stream on a whim (like their software reuse efforts),
companies do not have this luxury.  You get fired or demoted for doing things
like that at companies in the non-Mandated world.  Further executives are
going to be reluctant for their company to choose a language that they have
never heard of (from their cursory skimming of Computerworld - "ADA? What
has the disability act have to do with programming?").
	Thus the window of opportunity for becoming a major OO language only 
will remain open for a few more years.  Given that it will take that long for
Ada9X to be finalized, fully approved, standardized, and have compilers
debugged and tested, by the time the Ada9X community is ready to make the
case for Ada9X (if they even care at all), most windows will be shut.  And
given the lousy track record of the DoD (especially AJPO), the compiler
vendors and the Ada software initiatives contractors (STARS) in promoting
Ada outside the mandated world, where they have to spend their own money,
I doubt highly the case for Ada9X will be made any better than any case for
Ada83.  Sometime ask Tucker Taft what percent of the money Intermetrics
receives for its Ada consulting to the DoD that it spends on evangelizing
Ada in the non-Mandated world.   Can you spell Z E R O.
	Whichever socialists insisted on, and pushed for, the Ada Mandate
should have realized that it would only work if someone was out there making
the case for Ada with a great capitalistic marketing job for Ada.  As my
many posts on the complete absence of Ada in the non-Mandated world (i.e.
by measuring ads for compilers, help wanted ads, booth space at trade shows,
articles in magazines, source code in university theses, commercial seminars
for programming languages), it is obvious that no one is making the case
for Ada, no one wants to make the case for Ada, leading to its virtual
absence in the non-Mandated world, which undermines implicit assumptions
used to justify the imposition of the Mandate, which leads to the thought
that under present conditions of the apathy of Ada's health outside the
Mandated world, that the Mandate is a potential threat to national security
by making it increasingly difficult for the DoD, with shrinking budgets,
to meet all of its programming needs.

Just my one dollars worth, which is what I have to pay with my own money
to get Internet access.

Anyways, thanks for the comment on the post.


Greg Aharonian
Source Translation & Optimization

-- 
**************************************************************************
Greg Aharonian
Source Translation & Optimiztion
P.O. Box 404, Belmont, MA 02178

Re: Datapro announces survey of ObjectOrient languages

[Christopher J. Henrich]

In article <SRCTRAN.93Apr2100042@world.std.com> srctran@world.std.com (Gregory 
Aharonian) writes:
>	Thus the window of opportunity for becoming a major OO language only 
>will remain open for a few more years.  Given that it will take that long for
>Ada9X to be finalized, fully approved, standardized, and have compilers
>debugged and tested, by the time the Ada9X community is ready to make the
>case for Ada9X (if they even care at all), most windows will be shut.  And
Here we seem to be looking at a big difference between two cultures.
If the fans of C++ had waited until that language was finalized, they
would not have been heard from yet.  Because it isn't finalized yet.
(I may be wrong but I think the ANSI committee is still considering
changes to C++.)  Instead, they have been going ahead with
preliminary tools, either accepting or positively enjoying the fact
that the language was in a state of flux.  Clearly this has been
appropriate for that part of the world of softwre development which
builds applications for PCs.

Is the much more sedate pace of Ada-9X evangelization (yes, Virginia,
there is such a thing) equally appropriate for a different part of
the marketplace?  I am referring to systems where the computer is
part of a large, expensive piece of equipment, development projects
involve many people, and large bodies of code must be maintained
over times measured in decades.  In this sector, Ada-83 seems to have
been successful, regardless of the Mandate.  I think Ada-9X will
do very well, but maybe Mr. Aharonian is right in wanting us to
sing its praises more loudly and clearly.

There is one important feature of Ada (83 or 9X) that C++ utterly
lacks: the program library.  In Ada-land, we take it for granted that
the compiler knows all about the entire set of modules that have been
compiled into one "program library."  This is especially important in
long-term multi-person projects.  I understand that there are
significant enhancements in library management in Ada9X.  These
should be an important talking and selling point.

There you have it.  It would be nice if we could spread light,
inexpensive, preliminary Ada9X systems around the PC world, and give
lots of programmer a chance to try it out, write some neat packages
of object-oriented code, and so on.  (I want one for my Mac, dammit!)
But I don't see how Ada-9X can get there from here.  Likewise, it
would be nice to make program library management integral to C++, but
I do not see how C++ can get there from where it is now.

Regards,
Chris Henrich

Re: Datapro announces survey of ObjectOrient languages

[Gregory Aharonian]

>Is the much more sedate pace of Ada-9X evangelization (yes, Virginia,
>there is such a thing) equally appropriate for a different part of
>the marketplace?  I am referring to systems where the computer is
>part of a large, expensive piece of equipment, development projects
>involve many people, and large bodies of code must be maintained
>over times measured in decades.  In this sector, Ada-83 seems to have
>been successful, regardless of the Mandate.  I think Ada-9X will
>do very well, but maybe Mr. Aharonian is right in wanting us to
>sing its praises more loudly and clearly.

    There are many large, systems engineering in the non-Mandated world
being done in a variety of languages, most not in Ada.  In all of these
cases, vendors offer products and services very aggressively, with
active sales campaigns and discounts and deals to push their products.
So Ada has choice but to be marketed as agressively as possible now.
The sedate approach is for losers, or people who really don't believe 
in their product.
    Also, you are wrong about Ada's success, at least in the non-Mandated
world.  Dollar wise and percentage wise, Ada is not doing very well being
choosen, which is different from not being chosen at all.

>There is one important feature of Ada (83 or 9X) that C++ utterly
>lacks: the program library.  In Ada-land, we take it for granted that
>the compiler knows all about the entire set of modules that have been
>compiled into one "program library."  This is especially important in
>long-term multi-person projects.  I understand that there are
>significant enhancements in library management in Ada9X.  These
>should be an important talking and selling point.

   I will argue non-technically that this is a lousy feature of Ada.
The key is non-technically.  As many have pointed out, one reason C/C++
does so well is that the languages are deficient in some aspects, like
program libraries.  This attracts companies to offer products to meet
these needs, which generates more business activity and advertising,
which gives more recognition to the languages, which makes them easy to
be accepted by management.  Industries are driven by this third party
activity (look at PCs).   Since Ada is so great, it offers very such
opportunities, resulting in little activity to attract the corporate eye.

    In the case of program library, there are too many such needs to be
met by the one solution Ada offers, which discourages people from using
the language in the non-Mandated world.  For C/C++ I am able to choose
from many commercial products with program library capability that best
fits my needs.  Freedom of choice - it's what is supposed to be defending.

    In any event, Ada's success and role in national security relies on
great mastery of the US software industry market forces.  Those pushing for
the Mandate were irresponsible in not planning sufficiently for the
marketing and fostering of Ada outside the Mandated world, if they ever
understood the problem of marketing, which is normally not a soldierly
thing to do.  No one expects the DoD to be a marketing organization -
that's what the free markets are for.  But in pushing for the Mandate, the
DoD implicitly became a marketing organzation.  Until it explicitly
recognizes it, Ada will languish and be niched in the Mandated world,
which in a era of decreasing defense budgets, is a potential threat to
national security.

Greg Aharonian
Source Translation & Optimization
-- 
**************************************************************************
Greg Aharonian
Source Translation & Optimiztion
P.O. Box 404, Belmont, MA 02178