Re: Fantastic Ada promotional piece from Rational (long)

[jgoodsen@trinidad.radsoft.com (John Goodsen)]

In article <34al0m$89d@felix.seas.gwu.edu> mfeldman@seas.gwu.edu (Michael Feldman) writes:

   The points that I, and other educators, have been making since Ada
   started in 1983 (actually before that) are

   (1) the vendors should seek funding based NOT on the DoD mandate, but 
       on Ada's viability for the world. That you once again cited the
       strength of the mandate as the _key_ concern tells me that, in
       fact, we were correct in our assumption that the vendors did not
       really believe in Ada, only in the potentially infinite DoD
       market they could cash in on. 

Show me one successful businessman who makes the investment to start a company
because they "believe" in something.  Hardly a wise business decision.  It
is completely legitimate for a company to enter a market with a plan to cash
in on their presence in the market.  Why do you and Greg continue to blast
companies for their decisions to follow a market rather than try to create
a market.  If someone came to you and said, "Mike, I need 2 million to start
this company who's goal is to build cool Ada things because we believe that
Ada is the best language to use, can you invest?"  If you had a couple mil in
your back pocket, can you honestly tell me that the wisest decision is to
try to push a market into existence by investing into it?  Markets develop
out of demand from consumers, not out of pushing by vendors who want to sell
products.  If I've got something to sell you but you don't want to buy it,
then am I the bad guy because you don't buy it?  I think not.

I'm getting kind of tired of hearing from people who have no
experience starting and running a business blast (directly or
indirectly) those who at least have the balls to give it a shot.

So we can hear you and Greg talk about who should get fired for
bad policy decisions within the DoD, but if any executive of a 
commercial company followed the advice you guys hand out so freely,
let me tell you, there'd be a lot of heads rolling because the
suggestions you and Greg make are not decisions that are typically
in the best interest of the shareholders in a company.  Your
suggestions of giving free handouts to universities and Greg's suggestions
that companies who make money in the Ada market are hypocrites when
they pursue business in other markets are equally ludicrous.
I for one would raise hell if I invested in a company and they decided
to ignore potentials for making money in non-Ada markets.  I'd like
to think you are not so crazy that you like your investors blowing
your money down a tube either.

It's the DoD that has killed Ada through failed policies and thereby
giving Ada a bad name to those who might think of using the language
in a non-DoD project.  Not small businesses.  Cut the liberal, "make
me feel good" business-bashing, crap and stay focused on the real
problem that Greg exposes daily.

       Turbo Pascal was just getting started at the same time; C++ started 
       even later. Somehow, Borland and its ilk (and I'm sure there were
       minicomputer software companies in the same category, but I can't
       recall now) managed to convince their backers of the viability of
       their stuff _without_ having to refer to a DoD Ada mandate. The
       playing field was, as I recall, pretty level then. 

Come on, Mike.  Pascal was taught to damn near every college student
back when Turbo Pascal came out.  No vendor intervention was required.
No one even heard what Ada was.  Not what I'd call a level playing field
by any stretch of the imagination.

       BUT the vendors - right from the start - acted as though their
       only competition was from other Ada vendors. In the DoD arena,
       that was (more-or-less) correct at the time, but any industry
       that is so fragile that its survival depends on the DoD bureaucracy
       is, in my outsider's opinion, built on a rather weak foundation.

   (2) an important part of the success of a new computing technology
       is its propagation through the universities. We tried to make
       this point over and over, starting in 1983 or earlier, but we
       were belittled as fuzzy-headed academics needing some business 
       school education. 

Well, it seems you don't always think in the pure economics of issues
which is where business is rooted.  You are right, though.  The Turbo
Pascal example has some merit, but you fail to analyze it properly.
Why did turbo pascal take off so fast?  Because college students were
learning it and using it on projects.  If you think that Borland created
the Pascal market, you are way off base.

Meanwhile, we have (or had) the same potential for Ada.  But because,
American universities have failed to promote Ada to the same level that
Pascal was promoted, it now becomes a small business failure because they
didn't give enough money and free software to people.  This argument reaks
of a liberal "can't take responsibility for my own actions" tone.  It's
not the goal of small business to promote the best technology.  It's the
goal of small business to capitalize on technology and make money on it.
Forcing Ada technology down everyone's throat isn't going to create the
market that we'd like to see.

       So instead of being brought into the picture as allies, we were
       called the "education _market_" and treated like second-class
       customers. We were offered what were called deep discounts,
       and some of us actually paid the "discounted" price, then found
       we were being gouged for alleged "support" we never used. We were
       asked to pay for the version that corrected the bugs we reported.

Borland has the same policy.  Sounds like good business to me, so maybe
it's not to far off to classify you and your peers as the "education market",
because I don't hear a lot of business sense coming out of your keyboard yet.

       Over and over I was told that vendors were just not interested
       in hooking the freshmen, because they would only be influential
       after 5 years. 

If Ada is so great, why to people have to be "hooked" through commercial business
support.  Again, I mention the history of Pascal.  That language
became a defacto standard teaching language because it was identified
and accepted as the norm.  Commercial businesses had little to do
with the incorporation of Pascal into so many CS programs.  How come
Ada can't experience the same defacto standardization unless businesses
fork out free software, support and pump money into promotional campaigns?
You have yet to make this connection between the 2 languages.

And don't get me wrong, of course it makes business sense for a vendor
to promote a market, but to put the blame on small businesses for the
failed Ada market is a much more myopic view of capitalism from a
college professor than I would expect to hear these days.

   We are now ten years into Ada's life, and (I guess) the vendors
   have finally awakened to their myopia of the past. I have heard 
   vendor folks on TA panels beating their breasts about their lack of
   a "market orientation" in the early days. Well, it's not like
   some folks weren't trying to turn their heads; I think they
   were simply too stubborn to listen, or too paralyzed by the
   ups and downs of the mandate to act on it.

   That C and C++ propagated like a virus through the universities while
   the Ada vendors'  backs were turned (or their heads were stuck in the
   sand, whatever), is just plain historical fact. If it is "vendor
   bashing" to set the record straight, so be it.

I suppose that it was the responsibility of Ada compiler vendors
to sniff out every student and make sure that they weren't bitten
by the C/C++ virus?  Meanwhile, instructors and students alike
remain ignorant of new languages like Ada.  Are you saying that the
last 10 years of failure to get Ada in as a defacto standard programming
language in CS curriculums is the fault of Ada vendors?  come on.


   Instead, the vendors essentially walked away from Ada, preferring, 
   apparently, to go for the larger C++ market, and advertising
   accordingly. So they did not really work at expanding the Ada
   market when C++ was not a strong competitor, and, having blown it
   the first time around, prefer to diversify into C++. This does NOT
   show the world a strong confidence in Ada; instead the message is
   "see, even its own suppliers jump off the train just because DoD
   blows a different-toned whistle."

Allow me to paraphrase this:

     "Ada vendors have failed expand the Ada market faster than the C++
      market has expanded, and now that they have "blown it", they
      are making the wise business decision to move into the C++ market
      as well."

2 response to this:

      A) It is a invalid premise that the reason for the a failed Ada market
         is due to vendors who are business savvy enough to know how to
         grow a market.  Remember, markets grow much faster and bigger when
         user's demand a product.  Where is the demand for Ada in this supply-side
         economy that we live in?  How come people are still graduating from
         college and haven't heard of the language?  I suppose it's the
         fault of Ada vendors for not giving out free software
         and support  (btw, I'm still waiting for that Borland free giveaway
         offer to students - ain't gonna happen)


       B) Hurray for the vendors that have enough business sense to carve out
          their place in the proven C++ market!  If they had my investment bucks
          and DIDN'T do this, *THEN* you might have something interesting to
          complain about.  Business is about making money, not creating markets
          based upon technological beliefs.

          And a followup on this note:  IMHO, the Ada9X project has failed
          miserably in pitting Ada against new OO languages, including C++.
          In particular, the introduction of yet another set of terms
          for a concept that is storming the industry (I'm talking about
          the decision to use "tagged types" in place of "class" and the
          lack of multiple inheritance).  I've reviewed the email/news/and lsn's
          and it still seems to me that Tucker et. al. on the Ada 9X team
          are sitting too high on their hobby horses to understand the most
          basic concepts about how to market Ada 9X into the OO community.
          Forcing people to discover that a tagged type is similar to a class,
          with a few noticable syntactic and structural caveats was a piss
          poor decision that smacks of the Ada arrogance that you and Greg so
          often dispell.  Add to this the decision to not provide direct
          support for multiple inheritance in the language and you have 2 BIG 
          reasons why Ada 9X isn't going to go far against C++.  Yeah, I've
          heard Tucker et. al. argue from a technical standpoint and it's
          an argument that either side can win, but I contend that the bottom
          line is that the Ada community (language designers, users, ...) at
          large is responsible for lack of any fast growing Ada markets.
          I've talked to 2 ex-Ada programmer's this week.  Both were aware of
          Ada 9X but neither was aware that it supported object oriented programming.
          I did a little experiment.  The first programmer, I told about tagged types.
          The terminology along (tagged types) caused him to go "huh?"  I had to
          draw on the board before he understood.  After the work required to 
          teach the first guy what Ada 9X had in store, I introduced tagged types
          to the second engineer under the guise of a "class type".  Guess what?
          I only had to explain that a "class type" in Ada was similar to a "class"
          in C++ except that the member functions are not stored as part of the object.
          Once he understood, I told him, "BTW, it's not called "class type", but rather
          "tagged typed" in Ada.  Just an example of why you don't present new
          terminology for old concepts to potential customers.  People don't want
          to have to learn what a "tagged type" is before they decide that Ada has
          what C++ has to offer...  

   What is it about this group of companies that puts them so often in
   a "we should have done X" mode? Can they get it right? Maybe they 
   should open their ears to a few more fuzzy-headed academics, even if
   we haven't got business degrees.

Sorry Mike, but again, if you ask many Ada vendors what they would have
done differently, they will tell you that the biggest mistake they made
was believing the government when they said that Ada would be mandated
across DoD projects and then never followed through, after millions of
dollars in private funds were spent by vendors in preparation to make
this market roll...  Very few Ada vendors that I know will say that 
the biggest mistake they made was *NOT* investing enough in the Ada 
market.  They invested.  A lot of them heavily.  Uncle Sam did the screwing.

   >The main problem faced by software developers today is no longer the language 
   >they use but the availability of tools, interfaces, and components.  In a way, 
   >the availability of quality interfaces and components is what makes 
   >development in C++  efficient in spite of the lower quality of the language. 

   Yes, of course. But these tools did not come out of the blue sky.
   A lot of them were built or at least prototyped in the universities,
   where -- guess what -- C++ was handed to them on a platter. I don't
   know where the funding for most of the GNU stuff came from, but
   I'll guess it was not mostly from Uncle Sam. And the GNU stuff -
   with source code available - is what made, and is making, so much
   of this development possible in the universities.

So what's your point on this?  I think you make mine nicely :-)
Don't tell me that C++ was handed on a platter to anyone from a
vendor.  Nearly every piece of netware I pull down was developed
with the gcc/g++ compilers.  Free compilers.  No vendor intervention
was required there.  So why is it so necessary for vendors to give
away their wares and tools to universities in order for Ada to succeed.
The answer is that it is not necessary.  It's just a nice sounding
panacea and if the vendors start giving aware more compilers and mentioning
Ada in their ads more, then they get to feel good about working on
the problem - regardless if it will amount to any level of success in
addressing the problem.

This is the same argument that people use to justify social welfare
programs.  "Those with money need to fund programs that the rest of
us believe in".  Government beauracracies might work this way.  Businesses
die this way.

   Yes, I know, GNAT will do the same for Ada 9X. But of course the
   vendors screamed bloody murder at the idea that they would have
   free-software competition. I think this group is still unable to
   see the big picture.

   >For Ada to be used, Ada vendors must find ways to supply these components and 
   >tools.  This means financing them and deriving enough revenue to be able to  
   >continue.  So if a vendor chooses to make such tools language-independent to 
   >amortize the costs on a larger  population  of customers, why not?

   I am not in the least opposed to this. I _am_ opposed to vendors
   not advertising that their stuff is multi-language. If indeed you
   see this as a strength, why not tell the world? It is very rare to
   see any of these things advertised as compastible with Ada _or_
   C++. Mostly it is only the C++ connection one sees. I find this
   exceedingly hard to explain to myself; the only answer I can come
   up with is that the vendors feel their Ada connection taints them,
   and so they want to be perceived as making a clean break.

I think you're right on with this one.  But, is it the vendor's fault
that Ada leaves a sour taste in peoples mouths?  Is it the vendor's
responsibility to bastardize an advertisement with sometime that will
leave that sour taste.  Vendors are in business to make money.  Period.
Injecting a sour taste into the reader of an advertisement is hardly
a wise use of the advertising dollar.  The deeper problem is "Why does Ada
leave the impression it does with people?"   I only point back to 

   If I am way off base on this, and there is a defensible reason why
   vendors from IDE to Intermetrics to DDC-I to Rational cannot use
   the A word in their ads along with C++, I'll be glad to be corrected,
   either publicly or privately.


I think you hit the nail on the head.  Why muff an ad that is focused
towards C++ developers by mentioning Ada and leave a bad taste in
their mouth?  If I were placing C++ product ads right now, I'd be
taking the same approach.
-- 
--
John Goodsen                         Currently on-site at:
The Dalmatian Group                       JP Morgan 
User Interface Specialists                60 Wall St., New York City
jgoodsen@radsoft.com                      jgoodsen@jpmorgan.com