Re: How to Make Ada more widely used?

Re: How to Make Ada more widely used?

[Robert Dewar]

Greg Aharonian uses an interesting form of argument. His thesis is that
the mandate is responsible for the lack of libraries etc. He never gives
any hard data to support such a viewpoint (indeed it's hard to imagine how
any such data could be obtained). So apparently he feels that an acceptable]
substitute is simply to repeat the unsubstantiated claim seven times a day.

Re: How to Make Ada more widely used?

[John Bollenbacher]

Michael Feldman (mfeldman@seas.gwu.edu) wrote:

[much stuff deleted]

: I think I'm going to declare a 1-person moratorium on armchair analysis
: for a while. Anyone want to join me? 

Amen.  The water's great.  Come on in...

: 
: Mike Feldman
: ------------------------------------------------------------------------
: Michael B. Feldman -  co-chair, SIGAda Education Committee
: Professor, Dept. of Electrical Engineering and Computer Science
: The George Washington University -  Washington, DC 20052 USA
: 202-994-5253 (voice) - 202-994-5296 (fax) - mfeldman@seas.gwu.edu (Internet)
: ------------------------------------------------------------------------

--
-----------------------------------------------------------------------------
- John Bollenbacher                                        jhb@dale.cts.com -
- Titan Linkabit Corp.                                       (619) 552-9963 -
- 3033 Science Park Rd.                                                     -
- San Diego, Ca. 92121                                                      -
-----------------------------------------------------------------------------

Re: How to Make Ada more widely used?

[Bjarne Stroustrup]

>From: sampson@nosc.mil (Charles H. Sampson)
>     The impetus behind the effort that led to Ada was the DoD's conclusion
>in the 70's that the proliferation of languages being used at that time was
>a major part of the cost of DoD software.  This conclusion was based on
>studies, but I don't remember if they addressed the issue of tool costs. 
>Unfortunately, I haven't the slightest idea where you should look to find
>government documents this old or how you can request them if you know where
>to look.

For starters, look at the Ada paper for the ACM History of programming
conference (HOPL2). That paper focusses on the reasons that a single
programming language was deemed necessary and the politics needed to
design such a language and get it mandated.

Jean Isbiah was invited to write a paper on the language-technical aspects
of Ada, but unfortunately he was to busy to invest the time needed (I can
vouch that the amount of effort required to write a HOPL paper really is
very significant).

Re: How to Make Ada more widely used?

[Gregory Aharonian]

>From: sampson@nosc.mil (Charles H. Sampson)
>     The impetus behind the effort that led to Ada was the DoD's conclusion
>in the 70's that the proliferation of languages being used at that time was
>a major part of the cost of DoD software.  This conclusion was based on
>studies, but I don't remember if they addressed the issue of tool costs. 
>Unfortunately, I haven't the slightest idea where you should look to find
>government documents this old or how you can request them if you know where
>to look.

   Given the large scale use of non-Ada languages by people inside the
DoD itself presently, I doubt the value of these reports if the very same
organization's people don't pay any attention to them.  Besides, there
are a small number of DoD reports (most suppressed) showing that the 
country does not have the infrastructure to support a one language policy,
and that most of the initial promises made about Ada benefit and Ada
policies were completely wrong.

   Here's a classic example of the intellectual myopia that plagues most
Ada policy statements.  In the SEI sub-study for the Mosemann reports,
SEI states that it uses a set of criteria from IBM to assess the benefits
of using Ada and C++, and determined Ada was better, which they stated and
was touted to death.  What they didn't observe was that most of the criteria,
if not all, equally as well applied to the life-cycle programming problems
of the corporate world, a world to which IBM is saying that Smalltalk and
C++ are your future answers.

    Thus either SEI used the criteria incorrectly, IBM used the criteria
incorrectly, the criteria are incorrect (all three of which it is hard to
believe, especially in an Ada context since the DoD gives these guys tons
of money to do things correctly), or that SEI's conclusion is meaningless
for some other reason.  Given this trend in programming language analysis,
I question how relevant any of those reports you cite are, for those 
reports probably have similar problems, and even if not, too many people
inside the DoD itself right now are disobeying a federal regulation and
not using Ada.  I always thought that one rule for any soldier was to obey
all of the laws of the land.

    What has killed Ada over the years is the DoD's lack of interest in
honest assessments of its language and policies, made dishonest by basing
conclusions on only favorable data and funding people with a vested 
interest in offering conclusions that ensure that their funding will be
continued.  And for those few internal studies that are more methodical and
realistic, the DoD either suppresses them or classifies them, effectively
preventing anyone from ever finding about even the DoD's own doubts.

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

Re: How to Make Ada more widely used?

[Charles H. Sampson]

In article <1993Jun9.214716.15798@mksol.dseg.ti.com> mccall@mksol.dseg.ti.com (
fred j mccall 575-3539) writes:
>In <C83xAE.MMs@taurus.cs.nps.navy.mil> shimeall@cs.nps.navy.mil (timothy shime
all) writes:
>
>>In article <SRCTRAN.93Jun2233244@world.std.com> srctran@world.std.com (Gregor
y Aharonian) writes:
>>>   Simple.  Drop the Ada Mandate.  It is an artificial distortion of the
>>>marketplace that has sheltered Ada from competition with other languages.
>
>>Tell us Greg: If the Ada Mandate is dropped, how is DoD to retain
>>the small-language-set benefits?  
>
>Have there been any believable studies to indicate that these benefits
>have indeed been received?  In other words, is there any credible
>evidence that the likely gains and savings made are not wiped out by
>the higher cost of Ada compilers and development tools?

     The impetus behind the effort that led to Ada was the DoD's conclusion
in the 70's that the proliferation of languages being used at that time was
a major part of the cost of DoD software.  This conclusion was based on
studies, but I don't remember if they addressed the issue of tool costs. 
Unfortunately, I haven't the slightest idea where you should look to find
government documents this old or how you can request them if you know where
to look.

                              Charlie

Re: How to Make Ada more widely used?

[Michael Feldman]

In article <1993Jun9.220019.16017@mksol.dseg.ti.com> mccall@mksol.dseg.ti.com (
fred j mccall 575-3539) writes:
>
>In other words, they acted in a way that was exactly what could be
>expected, given the total lack of economic forces on them.  Now you
>want to blame *them* for acting like businesses instead of how you
>would have preferred them to act.  Sorry, but folks who ignore history
>are doomed to repeat it and history (not to mention basic economics)
>teaches us that protected markets are generally going to behave pretty
>much like what we've seen Ada vendors do.  It was predictable.

Well, you're making some generalizations about protected markets here.
You may have a point, but I think it would be a good idea for you to show
some other examples, since you're so sure of what "history teaches us."
>
>So, minus all the "beltway bandit" emotional diatribes, you're upset
>that the companies making the compilers didn't push the language to
>get people who weren't interested to use it?  That wasn't their job
>(or in their interest).  Businesses maximize profits.  Spending a
>bunch of extra money to advertise and 'sell' a language that they've
>been gifted with a captive market for would merely be stupid on their
>part. 

Clearly there's room in the world for several opinions on this. Mine
happens to be that is _was_ in their interest to take a technology that
was _manifestly_ "dual purpose", that they had already invested in
developing, and spend money to make money, selling this product to
a much wider clientele. You sau "businesses maximize profits." I
certainly agree; so they should. But you are not distinguishing between
the profits in the next quarter and the profits in the next decade.

You opine that it was predictable - and in their interest - to go for the 
short-term bucks and neglect building up the long-term demand for their
products. I opine that this next-quarter obsession is a relatively
recent thing in US business and is a mentality that is preventing US
business from making the investments that will keep the economy going in 
the long haul. Much of American business has, for the last, say, 15 years,
focused more on pushing paper, buying each other, building empty buildings,
etc., than on making long-term investment. Much of American business got
so caught up in the Defense buildup that they never focused on what they
would do in the inevitable build-down. 

I knew something was wrong when Westinghouse stopped selling lightbulbs
and Singer stopped selling sewing machines. We will all be stuck, one way
or another, paying for these companies to get back to their roots in the
broader-than-defense sector.

There needs to be a balance between short and long term.

Whether, in a protected market, _any_ business would've gone the way of
the Ada business, is - I opine - just your opinion, to which you are
certainly entitled.

I think I'm going to declare a 1-person moratorium on armchair analysis
for a while. Anyone want to join me? Let's see if we can get the group back 
on technical stuff. We can certainly take the flamewars to e-mail.

Mike Feldman
------------------------------------------------------------------------
Michael B. Feldman -  co-chair, SIGAda Education Committee
Professor, Dept. of Electrical Engineering and Computer Science
The George Washington University -  Washington, DC 20052 USA
202-994-5253 (voice) - 202-994-5296 (fax) - mfeldman@seas.gwu.edu (Internet)
------------------------------------------------------------------------

Re: How to Make Ada more widely used?

[cis.ohio-state.edu!magnus.acs.ohio-state.edu!usenet.ins.cwru.edu!howland.]

In <1993Jun8.040313.1135@seas.gwu.edu> mfeldman@seas.gwu.edu (Michael Feldman) 
writes:

>In article <SRCTRAN.93Jun7230852@world.std.com> srctran@world.std.com (Gregory
 Aharonian) writes:
>>
>>Of course, thanks to the Mandate, there are virtually no Ada support
>>libraries for most applications domains, while despite the problems
>>of C, there are hundreds of commercial and thousands of public domain
>>support libraries in C.  Thus for a few things, Ada programmers do
>>quite well, while for most other applications, Ada programmers can't
>>do anything at all.  The Mandate has allow the destruction of a
>>commercial Ada business, which despite DoD denials, hurts DoD
>>software development and drives up costs.
>>
>Hogwash. DoD did not force the Ada vendors to take, for ten years,
>their characteristic myopic approach to business. For all that I
>think the (congressional) mandate was a mistake, I nevertheless
>pin the sorry state of the Ada industry squarely where IMHO it
>belongs: on a bunch of companies that had the chance to act like
>Silicon Valley entrepreneurs, developing a government and a commercial
>market in parallel, as so many companies have done. 

>Alas, they didn't act like Silicon Valley entrepreneurs. They acted 
>like Beltway Bandits, jumping on to what they thought was an infinitely 
>long DoD gravy train, blowing off other market sectors and the universities,
>until the declining post-Cold War defense situation shook them out of their
>awful complacency. I do NOT believe that DoD asked them to be, or
>intended that they be, captive DoD contractors. They did it themselves.
>Don't blame their shortsightednes and lousy management on the mandate.
>It surely _allowed_ these companies to become as they are, but nobody
>_forced_ them to. They blew it, Greg, all by their lonesomes.

In other words, they acted in a way that was exactly what could be
expected, given the total lack of economic forces on them.  Now you
want to blame *them* for acting like businesses instead of how you
would have preferred them to act.  Sorry, but folks who ignore history
are doomed to repeat it and history (not to mention basic economics)
teaches us that protected markets are generally going to behave pretty
much like what we've seen Ada vendors do.  It was predictable.

So, minus all the "beltway bandit" emotional diatribes, you're upset
that the companies making the compilers didn't push the language to
get people who weren't interested to use it?  That wasn't their job
(or in their interest).  Businesses maximize profits.  Spending a
bunch of extra money to advertise and 'sell' a language that they've
been gifted with a captive market for would merely be stupid on their
part. 

-- 
"Insisting on perfect safety is for people who don't have the balls to live
 in the real world."   -- Mary Shafer, NASA Ames Dryden
------------------------------------------------------------------------------
Fred.McCall@dseg.ti.com - I don't speak for others and they don't speak for me.

Re: How to Make Ada more widely used?

[cis.ohio-state.edu!magnus.acs.ohio-state.edu!usenet.ins.cwru.edu!howland.]

In <C83xAE.MMs@taurus.cs.nps.navy.mil> shimeall@cs.nps.navy.mil (timothy shimea
ll) writes:

>In article <SRCTRAN.93Jun2233244@world.std.com> srctran@world.std.com (Gregory
 Aharonian) writes:
>>   Simple.  Drop the Ada Mandate.  It is an artificial distortion of the
>>marketplace that has sheltered Ada from competition with other languages.

>Tell us Greg: If the Ada Mandate is dropped, how is DoD to retain
>the small-language-set benefits?  

Have there been any believable studies to indicate that these benefits
have indeed been received?  In other words, is there any credible
evidence that the likely gains and savings made are not wiped out by
the higher cost of Ada compilers and development tools?

>Factor into your calculations that
>there IS NO STANDARD for many commonly-used languages (incl. C and C++)

Not true.  There has been an ANSI and ISO standard for C for some time
now.  There is pretty much a de facto standard for C++, minus a few
bells and whistles.  

>, and, in fact, there are large portability problems for non-trivial 
>non-Ada-based applications across even rather similar environments.  

This also happens with Ada-based applications, if you are moving from
one operating system to another, does it not?  Oh, it may be more (or
less) difficult in any specific case, but anyone who thinks they are
going to take a major program that pushes the hardware and OS and
simply move the source to a different OS and recompile is kidding
themself (because if they're pushing limits, they are going to be
using some machine-specific things -- in which case the ease or
difficulty of porting has more to do with what those are and how well
they were isolated (which Ada helps in the case of developers who
aren't used to keeping that sort of thing foremost in their minds)). 

>The fact of the matter is that Ada is MEANT to be a niche language.
>The language sponsors, DoD, don't much care (except where it 
>impacts their costs) if Ada EVER is a huge commercial success.
>Even as it is today, it is VERY useful in the DoD context.

As are lots of other languages.  The point behind the complaints about
a niche language with a significant learning curve is that it makes
development significantly more costly than it has to be to get the
same kinds of benefits.

>				Tim
>Disclaimer: The preceding is NOT an official statement for any
>governmental or nongovernmental organization.  It is a personal
>opinion of the author.

This obviously is the case for me, as well.

-- 
"Insisting on perfect safety is for people who don't have the balls to live
 in the real world."   -- Mary Shafer, NASA Ames Dryden
------------------------------------------------------------------------------
Fred.McCall@dseg.ti.com - I don't speak for others and they don't speak for me.

Re: How to Make Ada more widely used?

[Doug Smith]

In article <dewar.739543517@schonberg> dewar@schonberg.NYU.EDU (Robert Dewar) w
rites:
>Greg Aharonian uses an interesting form of argument. His thesis is that
>the mandate is responsible for the lack of libraries etc. He never gives
>any hard data to support such a viewpoint (indeed it's hard to imagine how
>any such data could be obtained). So apparently he feels that an acceptable]
>substitute is simply to repeat the unsubstantiated claim seven times a day.


"...proof by repeated assertion."

       quoted from Peopleware.
       
Doug
smithd@software.org

Re: How to Make Ada more widely used?

[Gregory Aharonian]

>From: dewar@schonberg.NYU.EDU (Robert Dewar)
>Greg Aharonian uses an interesting form of argument. His thesis is that
>the mandate is responsible for the lack of libraries etc. He never gives
>any hard data to support such a viewpoint (indeed it's hard to imagine how
>any such data could be obtained). So apparently he feels that an acceptable]
>substitute is simply to repeat the unsubstantiated claim seven times a day.

Rob,
	My claims are as well substantiated as the claims in the original
Mosemann studies comparing Ada and C++.  Since you didn't complain about
their unsubstantiated claims, I don't why you are complaining about mine.
In fact, I try to at least be self-consistent in what I post, unlike the
Mosemann studies, which contradicted each other and consciously excluded
much programming language demographic data.

	So fine, flame me, as long as you flame the others to the same
degree.  And keep in mind that I do this Ada stuff with my own money,
while the others are spending tax dollars, and should be held to higher
standards as public servants.  Be fair or be silent.

Greg

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

Re: How to Make Ada more widely used?

[Kenneth Anderson]

In comp.lang.ada Michael Feldman writes:

>[stuff deleted]

>With some investment in their Mac system, and some nicely targeted
>advertising of it, I'll bet they could make it more popular than it is.
>They have the market all to themselves; I think they are not really
>exploiting it. Mac fans: can we make them hear us?

I would kill for an affordable (for students) Ada 9X compiler that allows
full access to the Mac toolbox. I add to your question above:
What is the best way of making our desires known?



>Cheers -

>Mike Feldman

Ken Anderson

P.S.  I used Ada 9X in the sentence above because I think it is better
suited to Mac programming than Ada83.  I just finished reading the
Intro to Ada 9X document by Barnes, and boy, am I excited.  I can't wait
to get my hands on ANY Ada 9X compiler so I can start using that
language!

Re: How to Make Ada more widely used?

[Beth Walker]

Mike Feldman writes:

>With some investment in their Mac system, and some nicely targeted
>advertising of it, I'll bet they could make it more popular than it is.
>They have the market all to themselves; I think they are not really
>exploiting it. Mac fans: can we make them hear us?

I hope so.  One of the things I really would like to see is a more
"Mac-like" interface along the lines of "Think-Ada".  The other
biggie on my wish list is for the Ada code to be compatible with
the GUI generator applications.

My reasons for this is that, when I write Ada code on my Mac, it is
for my own use, rather than for work.  That being the case, easier
and quicker (coding wise, not compiler speed wise) are two of my
main drivers.  If I can use C and an interface builder and take one
week as opposed to using Ada and hand-coding the interface (or 
having to use an un-Mac-like Ascii interface), I will use C.  If 
the time & effort are about equal, and the result is the same, I
would much rather use Ada.  Currently, I am being forced into the
C camp if I want a Mac-like interface, simply because of a dearth
of tools.

Re: How to Make Ada more widely used?

[cis.ohio-state.edu!math.ohio-state.edu!darwin.sura.net!seas.gwu.edu!mfeld]

In article <1993Jun8.152222.14698@sei.cmu.edu> gartm@ajpo.sei.cmu.edu (Mitch Ga
rt) writes:
>
[stuff deleted]

>   As another example, Alsys has put major efforts into making Ada
>products for environments that are thought of as non-DoD platforms.
>Years ago the conventional wisdom was that most Ada users would
>work on Vaxes or Unix workstations.  We went ahead and pioneered Ada 
>for the IBM PC.  This was a big success.  We also put major, 
>expensive, development efforts into producing Ada compilers for IBM 
>mainframe environments such as MVS and VM/CMS with the idea of 
>getting MIS applications to switch to Ada.  This was not such a success.  
>But you can't say we didn't try.

I'm delighted to see Alsys responding to this goading :-)

For years, Alsys essentially refused to talk to the universities. This
seems to be changing, but was certainly true until very recently. To
respond about the VM products, I always wondered why, if indeed Alsys was
trying to target the business world, they didn't try to target the
business _schools_. We have a nice VM machine sitting around at GW.
The CS and engineering students have moved to Unix, but a lot of business
students still write Cobol on the mainframe. Look around that environment;
you'll see that GW's not alone.

If that compiler is not "moving" in business, how 'bout donating it to
schools with VM machines? AdaWorld is a decent interface. Set up a
_small_ budget for supporting it; schools using compilers for _teaching_
require precious little support.
>
>   Right now Alsys (and Meridian) are working on Ada for Windows NT.  
>I can't speak for Meridian, but inside Alsys the hope is again that
>this product will help Ada to expand into the general programming 
>market.

Well, according to my limited reading (InfoWorld), Microsoft has placed
_thousands_ of beta copies of NT with developers (I think 50,000 was the
number I saw). Mebbe Alsys oughta get on the stick, get its NT compiler
into beta form, and drop it for a song into some of those NT beta sites,
before everyone gets completely locked into C++ there. How's that for
a constructive idea? If the NT compilers are as late as the Windows
compilers were, they won't make a dent because everyone will have
locked into Microsoft or Borland. Will the Alsys NT compiler hit the
market as soon as NT does? Will it be priced comparably to the C++
high-end products for developers? If not, forget it.
>
>   Yet another example is Ada 9X itself.  The Ada vendors will be
>making major investments in implementing 9X, and the object-oriented
>and information systems features that will be added are more for the 
>general market, to compete with C++ and COBOL, than for the DoD market.  
>Upgrading Ada compilers to 9X will be expensive and will be done 
>largely to try to compete in the general programming market.

Many of us in the education world would be DELIGHTED to act as beta
testers of your 9X compilers as they come together. Given the amount
of hand-wringing the Ada world has done about Ada in the universities,
I'd have expected by now that the vendors would be ringing my phone -
and my colleagues' phones - off the wall, trying to get us involved
- under nondisclosure if necessary - in the 9x work. Guess what?
They're not. We ain't gonna beg this time.
>
>   I share the frustration of Mike, Greg, and others that Ada hasn't
>been as big a success in the non-Mandated world as we would like.  
>But I think their statements that lay all the blame on the Ada vendors 
>are untrue and unfair.  We *are* trying to expand the use of Ada.

Well, I'm not exactly _blaming_ the vendors, merely continuing to remark
on the myopia that most of them have - either publicly or privately -
admitted. I'm not making this up; surely you know that.

I don't want to be unfair to Alsys. Alsys' attitude shows definite
signs of paradigm shift, and I applaud it. Keep it up, guys!
Keep the faith - maybe it's not too late!

Mike Feldman
------------------------------------------------------------------------
Michael B. Feldman -  co-chair, SIGAda Education Committee
Professor, Dept. of Electrical Engineering and Computer Science
The George Washington University -  Washington, DC 20052 USA
202-994-5253 (voice) - 202-994-5296 (fax) - mfeldman@seas.gwu.edu (Internet)
------------------------------------------------------------------------

Re: How to Make Ada more widely used?

[cis.ohio-state.edu!math.ohio-state.edu!darwin.sura.net!seas.gwu.edu!mfeld]

In article <1993Jun8.132425.4553@sei.cmu.edu> ae@sei.cmu.edu (Arthur Evans) wri
tes:
>
[stuff deleted]

>Probably some did; surely not all.  One company I'm familiar with came
>into the business intending to specialize in the highest possible
>quality of compiled code.  They quickly realized that this criterion was
>not terribly important in many markets and so concentrated on embedded
>applications.  Although many of their customers are in the so-called
>"mandated" world, others are using Ada because they perceive that it
>best meets their needs, and are using this company because they really
>care about code quality.

This sounds like Tartan or DDC-I. I think it would be really neat if 
someone who knows someone there could ask if they have any non-mandated
customers who'd be willing to see their names turn up on a list like mine.

>
>This company has no products for platforms such as PCs or Macs because
>such users are rarely willing to pay a lot extra to save the last few
>microseconds at run time.  I certainly wouldn't pay on my Mac -- I'm
>much more concerned with ease of use.
>
Absolutely. For many desktop PC's with human users (not embedded ones),
nearly any application that's not heavy into number-crunching will run
faster than its human user. Fast enough is fast enough. I've written
a _whole lot_ of programs with Ada/Ed DOS, which, on my 386-33, run
fast enough so that their execution performance is indistinguishable
to the _human_ user from a genuine executable program. And this is
with an interpretive back end running around a P-code file. 

I'm not suggesting that we all go out and write everything with Ada/Ed; 
rather, I took it as an extreme example of how "fast enough is fast 
enough."

>Otherwise, I agree with the point Mike has made.  Why is there only one
>Ada vendor for the Mac?  It would be instructive if some of the vendors
>would reply about how they see the market place.  I expect they would
>raise some points we haven't seen here up to now.
>
There have been 3 vendors who developed compilers for the Mac. Meridian is
still in the game; Alsys and TeleSoft (yes, I know they're the same now)
dropped out. TeleSoft never released a MacOS compiler, only an A/UX one.
Alsys had a MacOS system but dropped it _just_ as the typical Macs were
getting big and fast enough to run a good Ada system well. (Meridian on
my IIci is quite respectable; I haven't tried Alsys.) I thought it quite
foolish of Alsys to drop it prematurely. Apparently they aren't Mac-savvy
enough to realize the degree to which the hardware has caught up with Ada.
(The typical new Mac is a 68030 with 4-8 mb RAM, running fast enough
to make even a slow Ada compiler tolerable.)

The following is based in small part on contact I've had, in large part
on educated guessing. All three companies built their Ada compilers because 
they thought they saw some big DoD or NASA contracts coming down for the Mac 
platform. When these did not materialize, the vendors saw no particular 
reason to stay with the Mac, which they perceived as a "small market."

This was a good example, Art, of (what I perceive to be) the narrow
Uncle-Sam-oriented-ness of the Ada companies. When the government
contracts did not materialize for the Mac, bye-bye Mac compilers.

Meridian has yielded to pressure from "the rest of us" to stay in the
Mac game, but they don't invest a heck of a lot in their compiler. They
do support a very usable, if "thin", System 7 toolbox binding. Meridian
has told me that they make a profit selling their DOS and Mac systems at 
retail, and even their student-priced systems bring positive cash flow.
(Meridian has authorized me to say stuff like this.)

With some investment in their Mac system, and some nicely targeted
advertising of it, I'll bet they could make it more popular than it is.
They have the market all to themselves; I think they are not really
exploiting it. Mac fans: can we make them hear us?

Cheers -

Mike Feldman
------------------------------------------------------------------------
Michael B. Feldman -  co-chair, SIGAda Education Committee
Professor, Dept. of Electrical Engineering and Computer Science
The George Washington University -  Washington, DC 20052 USA
202-994-5253 (voice) - 202-994-5296 (fax) - mfeldman@seas.gwu.edu (Internet)
------------------------------------------------------------------------

Re: How to Make Ada more widely used?

[cis.ohio-state.edu!news.sei.cmu.edu!ajpo.sei.cmu.edu!gartm]

Michael Feldman writes of the Ada vendors:

> (some stuff)
>
> Alas, they didn't act like Silicon Valley entrepreneurs. They acted 
> like Beltway Bandits, jumping on to what they thought was an infinitely 
> long DoD gravy train, blowing off other market sectors and the universities,
> until the declining post-Cold War defense situation shook them out of their
> awful complacency. 
>
> (some more stuff)

   Totally untrue.  Believe me, Alsys has really tried to promote Ada
outside the Mandated world.  We have done ad campaigns, trade shows, 
articles, press tours, every idea we can think of to sell Ada.  We 
have spent lots of time, money, and energy.  In past years we have done 
lots of trade shows that are not specifically for Ada or DoD, and this 
year we will do at least a few.  In another recent message on 
comp.lang.ada, Greg Aharonian complained that the Ada vendors will not be 
selling Ada at ObjectWorld.  True, but a counterexample is that we
(Alsys) will be selling Ada at the Embedded Systems Conference, which 
is mostly a non-Ada crowd.

   As another example, Alsys has put major efforts into making Ada
products for environments that are thought of as non-DoD platforms.
Years ago the conventional wisdom was that most Ada users would
work on Vaxes or Unix workstations.  We went ahead and pioneered Ada 
for the IBM PC.  This was a big success.  We also put major, 
expensive, development efforts into producing Ada compilers for IBM 
mainframe environments such as MVS and VM/CMS with the idea of 
getting MIS applications to switch to Ada.  This was not such a success.  
But you can't say we didn't try.

   Right now Alsys (and Meridian) are working on Ada for Windows NT.  
I can't speak for Meridian, but inside Alsys the hope is again that
this product will help Ada to expand into the general programming 
market.

   Yet another example is Ada 9X itself.  The Ada vendors will be
making major investments in implementing 9X, and the object-oriented
and information systems features that will be added are more for the 
general market, to compete with C++ and COBOL, than for the DoD market.  
Upgrading Ada compilers to 9X will be expensive and will be done 
largely to try to compete in the general programming market.

   I share the frustration of Mike, Greg, and others that Ada hasn't
been as big a success in the non-Mandated world as we would like.  
But I think their statements that lay all the blame on the Ada vendors 
are untrue and unfair.  We *are* trying to expand the use of Ada.

	Mitch Gart

Re: How to Make Ada more widely used?

[dog.ee.lbl.gov!overload.lbl.gov!agate!usenet.ins.cwru.edu!magnus.acs.ohio]

mfeldman@seas.gwu.edu (Michael Feldman) says (in part):
> ... the sorry state of the Ada industry [is due to] a bunch of
> companies that had the chance to act like Silicon Valley
> entrepreneurs, developing a government and a commercial market in
> parallel, as so many companies have done.

> Alas, they didn't act like Silicon Valley entrepreneurs. They acted
> like Beltway Bandits, jumping on to what they thought was an
> infinitely long DoD gravy train, blowing off other market sectors and
> the universities ...

Probably some did; surely not all.  One company I'm familiar with came
into the business intending to specialize in the highest possible
quality of compiled code.  They quickly realized that this criterion was
not terribly important in many markets and so concentrated on embedded
applications.  Although many of their customers are in the so-called
"mandated" world, others are using Ada because they perceive that it
best meets their needs, and are using this company because they really
care about code quality.

This company has no products for platforms such as PCs or Macs because
such users are rarely willing to pay a lot extra to save the last few
microseconds at run time.  I certainly wouldn't pay on my Mac -- I'm
much more concerned with ease of use.

Otherwise, I agree with the point Mike has made.  Why is there only one
Ada vendor for the Mac?  It would be instructive if some of the vendors
would reply about how they see the market place.  I expect they would
raise some points we haven't seen here up to now.

Art
----------------------------------------------
Arthur Evans, Jr, PhD           Ada Consultant
461 Fairview Road
Pittsburgh PA  15238-1933
412-963-0839
ae@sei.cmu.edu

Re: How to Make Ada more widely used?

[Gregory Aharonian]

>I don't intend to slam C, but there ARE portability problems with
>that language that are much reduced in Ada, due to the compiler
>verification process and the emphasis on standard development.
>Much of that process and emphasis derives from the Mandate.

Of course, thanks to the Mandate, there are virtually no Ada support
libraries for most applications domains, while despite the problems
of C, there are hundreds of commercial and thousands of public domain
support libraries in C.  Thus for a few things, Ada programmers do
quite well, while for most other applications, Ada programmers can't
do anything at all.  The Mandate has allow the destruction of a
commercial Ada business, which despite DoD denials, hurts DoD
software development and drives up costs.

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

Re: How to Make Ada more widely used?

[Michael Feldman]

In article <SRCTRAN.93Jun7230852@world.std.com> srctran@world.std.com (Gregory 
Aharonian) writes:
>
>Of course, thanks to the Mandate, there are virtually no Ada support
>libraries for most applications domains, while despite the problems
>of C, there are hundreds of commercial and thousands of public domain
>support libraries in C.  Thus for a few things, Ada programmers do
>quite well, while for most other applications, Ada programmers can't
>do anything at all.  The Mandate has allow the destruction of a
>commercial Ada business, which despite DoD denials, hurts DoD
>software development and drives up costs.
>
Hogwash. DoD did not force the Ada vendors to take, for ten years,
their characteristic myopic approach to business. For all that I
think the (congressional) mandate was a mistake, I nevertheless
pin the sorry state of the Ada industry squarely where IMHO it
belongs: on a bunch of companies that had the chance to act like
Silicon Valley entrepreneurs, developing a government and a commercial
market in parallel, as so many companies have done. 

Alas, they didn't act like Silicon Valley entrepreneurs. They acted 
like Beltway Bandits, jumping on to what they thought was an infinitely 
long DoD gravy train, blowing off other market sectors and the universities,
until the declining post-Cold War defense situation shook them out of their
awful complacency. I do NOT believe that DoD asked them to be, or
intended that they be, captive DoD contractors. They did it themselves.
Don't blame their shortsightednes and lousy management on the mandate.
It surely _allowed_ these companies to become as they are, but nobody
_forced_ them to. They blew it, Greg, all by their lonesomes.

Mike
------------------------------------------------------------------------
Michael B. Feldman -  co-chair, SIGAda Education Committee
Professor, Dept. of Electrical Engineering and Computer Science
The George Washington University -  Washington, DC 20052 USA
202-994-5253 (voice) - 202-994-5296 (fax) - mfeldman@seas.gwu.edu (Internet)
------------------------------------------------------------------------

Re: How to Make Ada more widely used?

[agate!overload.lbl.gov!lll-winken.llnl.gov!taurus.cs.nps.navy.mil!shimeal]

In article <CBARBER.93Jun4174524@apricot-fddi.bbn.com> cbarber@apricot-fddi.bbn
.com (Chris Barber) writes:
>In article <C83xAE.MMs@taurus.cs.nps.navy.mil> shimeall@cs.nps.navy.mil 
>(timothy shimeall) writes:
>
>   , and, in fact, there are large portability problems for non-trivial 
>   non-Ada-based applications across even rather similar environments.  
>   (Right now, I'm trying to port the Aegis project-control system from 
>   Sun to Iris, and the bug-chasing is annoyingly complex...)
>
>Is this the fault of the languages used or of differences between 
>operating systems?

Actually, it's the fault of the support libraries.  There are subtle
and gross differences between the libraries.  Even when the syntax
is identical, the semantics are apparently a bit different.

I don't intend to slam C, but there ARE portability problems with
that language that are much reduced in Ada, due to the compiler
verification process and the emphasis on standard development.
Much of that process and emphasis derives from the Mandate.
					Tim

Re: How to Make Ada more widely used?

[cis.ohio-state.edu!pacific.mps.ohio-state.edu!math.ohio-state.edu!howland]

In article <C83xAE.MMs@taurus.cs.nps.navy.mil> shimeall@cs.nps.navy.mil 
(timothy shimeall) writes:

   Factor into your calculations that there IS NO STANDARD for many
   commonly-used languages (incl. C and C++)...

Minor nitpick: actually the ANSI standard for C was passed years ago,
although ANSI compliant compilers have only recently become available
for many systems.  The C++ ANSI standard is still several years away....

   , and, in fact, there are large portability problems for non-trivial 
   non-Ada-based applications across even rather similar environments.  
   (Right now, I'm trying to port the Aegis project-control system from 
   Sun to Iris, and the bug-chasing is annoyingly complex...)

Is this the fault of the languages used or of differences between 
operating systems?


--
Christopher Barber
(cbarber@bbn.com)

Re: How to Make Ada more widely used?

[timothy shimeall]

In article <SRCTRAN.93Jun2233244@world.std.com> srctran@world.std.com (Gregory 
Aharonian) writes:
>   Simple.  Drop the Ada Mandate.  It is an artificial distortion of the
>marketplace that has sheltered Ada from competition with other languages.
>Compiler prices are uncompetitive with those of other languages, no one
>is producing the Ada add-on libraries so useful for development in other
>worlds because there is no market.  Not only is the Ada Mandate bad
>economics, but it is also bad policy, as evidence by the ongoing defections
>from following by many inside the armed forces.  In short, because of Ada
>policies and not because of the language itself, Ada is uncompetitive and
>not attractive enough to be more widely used.

However, the Ada Mandate has also allowed the armed forces to drop
maintenance of litterally thousands of special-purpose compilers,
created during the 60s and 70s to support development of single
projects.  It has also allowed a large decrease in
maintenance-programmer-training costs...  In short, the ORIGINAL
purpose of the Ada Mandate, to allow DoD to move to a small subset
of languages (Ada, and selected others where economically necessary)
from the pre-Ada hugely multilingual days.

Tell us Greg: If the Ada Mandate is dropped, how is DoD to retain
the small-language-set benefits?  Factor into your calculations that
there IS NO STANDARD for many commonly-used languages (incl. C and C++)
, and, in fact, there are large portability problems for non-trivial 
non-Ada-based applications across even rather similar environments.  
(Right now, I'm trying to port the Aegis project-control system from 
Sun to Iris, and the bug-chasing is annoyingly complex...)
Given the DoD radical maintenance requirements, this may be a 
serious concern.

The fact of the matter is that Ada is MEANT to be a niche language.
The language sponsors, DoD, don't much care (except where it 
impacts their costs) if Ada EVER is a huge commercial success.
Even as it is today, it is VERY useful in the DoD context.
				Tim
Disclaimer: The preceding is NOT an official statement for any
governmental or nongovernmental organization.  It is a personal
opinion of the author.

Re: How to Make Ada more widely used?

[Holmes S. Liao]

In article <SRCTRAN.93Jun2233244@world.std.com> srctran@world.std.com (Gregory 
Aharonian) writes:
>
>    And to beat a dead horse, Ada will remain a dead, niche language as
>long as IBM refuses to give it as much attention as it does C++ and Smalltalk.

Well, I agree with what you said about free market and Ada mandate. But
I fail to see why IBM's endorsement to Ada has anything to do with this. 

For one thing: IBM didn't endorse Unix nor Windows. Any yet look at them on 
the market today. (I'm not saying Unix and Windows are on the same level. :-) 

holmes liao 

Re: How to Make Ada more widely used?

[David Helken]

I agree that a subset of Ada is no go. The cultural attitude is shaped in large
measure by access to an idea, thing or activity that provides feedback to the
psyche. Ada compilers are not accessable to Joe/Jane Programmer unless they can
afford such luxury. In fact, Ada has certain snob appeal if one cares to hear
it. GNAT may very well open the Ada accessability path for many who would like
to use it, but cannot afford it. 
  Also, the run time system has to become more affordable (in every way) so
  that high quality Ada support can be used at home, in the schools, and yes,
  even at work! Make Ada more accessable and it will become more widely used.
  Not so much as I would like because the problem domain for which Ada is well
  suited, though large, is just too much behind programs and products written
  in other languages e.g. accounting, games, editors). One could write EMACS in
  Ada, but why bother?

                   Dave

e-mail: helkenn@dale.cts.com

Re: How to Make Ada more widely used?

[Gregory Aharonian]

>I would like to start a discussion on what it would take to make
>Ada a more "widely used" language than it is today.
>QUESTION:
>What would it take to make Ada a more "widely used" language than
>it is now?

   Simple.  Drop the Ada Mandate.  It is an artificial distortion of the
marketplace that has sheltered Ada from competition with other languages.
Compiler prices are uncompetitive with those of other languages, no one
is producing the Ada add-on libraries so useful for development in other
worlds because there is no market.  Not only is the Ada Mandate bad
economics, but it is also bad policy, as evidence by the ongoing defections
from following by many inside the armed forces.  In short, because of Ada
policies and not because of the language itself, Ada is uncompetitive and
not attractive enough to be more widely used.

    Second, market and evangelize the language.  Outside the Mandated
world, few know much about Ada.  Why?  Because they rarely see Ada vendors
and Ada contractors (like the entire STARS program) ever marketing Ada
at trade shows and conferences.  How can people be expected to make any
decisions about a language they never receive information about?  But
the DoD doesn't care, the Ada contractors don't care and the Ada compiler
vendors barely care, and what does appear is made confusing by an equal
amount of C++ marketing from these same people.

    As a business process, the fostering of Ada has been a disaster from
day one.  And the same people are still in charge of stuff.  Until these
people are replaced by people who understand free markets, and these 
policies are changed, Ada use will remain stagnant.

    And to beat a dead horse, Ada will remain a dead, niche language as
long as IBM refuses to give it as much attention as it does C++ and Smalltalk.

Greg Aharonian

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

Re: How to Make Ada more widely used?

[David Emery]

Much of the stuff you're asking for, particularly dealing with
directories, are not language issues, but operating system issues.
That's why there are C and Ada (and FORTRAN) bindings to POSIX.  

The POSIX/Ada binding covers both command line arguments and access to
directories, as well as a lot of other "unix-like" stuff.  

Look for some string handling packages from the Ada Technology
Insertion Program later this year (coming attractions...).

				dave

Re: How to Make Ada more widely used?

[Michael Feldman]

In article <EMERY.93Jun1155219@goldfinger.mitre.org> emery@goldfinger.mitre.org
 (David Emery) writes:
>
>"Sequential" Ada (Ada without tasks) is a feasable subset, but there's
>not that much cost in compiling tasks.  There is a fair amount of cost
>in the runtime system.   Removing tasking from the language would
>seriously cripple the language for use in embedded applications...
>
Even Ada/Ed, now available FREE from NYU and other sources (WITH sources!)
handles tasking and generics quite competently. The code it compiles
is for a virtual machine. A compiler-savvy person putting in the effort
could produce a decent back end that would translate the virtual code
to executable. The runtime system is there, with sources and all.
If you wished, you could do this for a DOS machine. You'd have to
distribute the result under GNU rules, which is to say you'd have
to supply your source code. But think of the service you'd do.

In my opinion, a subset Ada83 compiler is a dead issue, a solution in
search of a problem. There's nothing wrong with the Meridian compilers
that a small investment couldn't fix; their basic technology is, IMHO,
quite sound, and they do the full language. Same with Janus and IntegrAda.

Meridian makes a small profit selling their compilers to students, at onesies
retail prices of $99-199, depending on features. I don't think Meridian
has tested the elasticity in dropping their commercial prices to, or 
slightly above, student levels, which would put them in direct competition
with Borland products, and certainly with the high-end "professional"
C-family compilers.

I maintain that (to paraphrase Bill Clinton and co.) it's the attitude, stupid.
It's an attitude change that would make Ada more widely used. It is STILL
the case that the non-Ada world is MUCH more ignorant of Ada than hostile
to it. And Ada's friends do little to rectify the situation, after all
these years. With friends like this, who needs enemies?

Mike Feldman
------------------------------------------------------------------------
Michael B. Feldman
co-chair, SIGAda Education Committee

Professor, Dept. of Electrical Engineering and Computer Science
School of Engineering and Applied Science
The George Washington University
Washington, DC 20052 USA
(202) 994-5253 (voice)
(202) 994-5296 (fax)
mfeldman@seas.gwu.edu (Internet)

"The most important thing is to be sincere, 
and once you've learned how to fake that, you've got it made." 
-- old show-business adage
------------------------------------------------------------------------

Re: How to Make Ada more widely used?

[cis.ohio-state.edu!magnus.acs.ohio-state.edu!usenet.ins.cwru.edu!howland.]

In article <1993Jun2.004023.10578@leeweyr.sccsi.com> bill@leeweyr.sccsi.com (Bi
ll Lee) writes:
>Some background:
>
>
>The answer as I see it is the POSIX Ada bindings. I am using them
>at work, but the "designers" who read through the POSIX specs and
>decided what was "required" unfortunately knew (know?) nothing
>about Ada. And the bindings we have been given have left out the
>most rudimentary elements: e.g., the generic read/write procedures
>in the I/O spec!
>

Just so nobody misunderstands: the "designers" were those at work,
not the POSIX bindings people. 1003.5 is as good a piece of work
as you could ask for. Too bad that the IEEE is so dead-set against Ada
that they are trying to force thin bindings on everything else.


Bill

Re: How to Make Ada more widely used?

[cis.ohio-state.edu!magnus.acs.ohio-state.edu!usenet.ins.cwru.edu!howland.]

Some background:

I've been "into" Ada technology since 1985. I have managed large
Ada projects. I have developed a large amount of Ada software. I
believe in the "rightness" of Ada.

IMHO:

One of the biggest problems encountered while trying to use Ada
for something other than the "large" developments is the lack of
readily available interfaces to the kind of things you need for
writing programs: portable, standard interfaces to the underlying
operating system.

As an example, how do you deal with directories in Ada? Easy to do
in C. Good "standard" libraries.

How do you deal with string parsing in Ada? Easy to do in C. Good
"standard" libraries.

How do you get command line arguments in Ada?  Easy to do in C.
It's part of the language.

And on and on and on..........

I develop Ada software for a living. I like the language. I would
prefer to use it all the time. But when I need a quick-and-dirty
that I can't do in a shell script or sed or something, I will turn
to C. (Yech! There, I said it. Pooey! 8-{)

The answer as I see it is the POSIX Ada bindings. I am using them
at work, but the "designers" who read through the POSIX specs and
decided what was "required" unfortunately knew (know?) nothing
about Ada. And the bindings we have been given have left out the
most rudimentary elements: e.g., the generic read/write procedures
in the I/O spec!

What I want is for the Ada vendors to bundle a FULL POSIX 1003.5
library with the compiler. And when it is available, a .20 library,
too! THAT will make Ada easier to use and will remove many of the
arguments that allow the old guard to stay in their C ways.

Regards,

Bill Lee

Re: How to Make Ada more widely used?

[David Emery]

A subset of Ada like you describe is very close to Modula-2.  (so why
not just use Mod-2?)  And, it makes the resulting subsetted language
very hard to use.  In particular, the behavior of Ada programs without
exceptions is almost impossible to describe.  Removing generics
severely weakens the type model, and makes it almost impossible to do
I/O on user-defined types. 

"Sequential" Ada (Ada without tasks) is a feasable subset, but there's
not that much cost in compiling tasks.  There is a fair amount of cost
in the runtime system.   Removing tasking from the language would
seriously cripple the language for use in embedded applications...

The SIGAda ARTEWG people have been discussing the notion of a standard
interface between a compiler and a RTS for a long time.  Part of the
problem is that this interface is very compiler-specific, and has a
serious impact on the code generation, etc.  And, this interface is
often also very dependent on the target architecture, O.S., etc.  It's
not clear what applications programmers would gain from this
interface, as they should not be calling these services directly.
Instead, they write code that the compiler then translates into calls
to the RTS.   

I am unconvinced that there is a market out there for Ada RTS
developments.  Some people have discussed companies to do this, but
there is not a large set of compiler-independent Ada RTS developers
out there.  In part this is due to the very compiler-dependent aspects
of a good, efficient RTS (for *any* language, not just Ada.)

				dave

How to Make Ada more widely used?

[Mark Johnson]

I would like to start a discussion on what it would take to make
Ada a more "widely used" language than it is today.

QUESTION:
What would it take to make Ada a more "widely used" language than
it is now?
(and please, no side issues on how widely used it is now, etc.)


MY ANSWER: (2 Parts)

1) Make a subset language of Ada 83, called say 'A+', which does
not include tasking, generic's, and exceptions.  This language
would not require a special runtime, would be much cheaper to
build a compiler for, and would still provide many of the good
qualities that Ada has over some other languages, such as,
readability, strong type checking, encapsulation, etc..

OK, so maybe you wouldn't be able to use this subset on defense
work.  No big deal, so you use Ada instead of 'A+'.  But, since
'A+' would be a subset of Ada, all of the A+ code could be reused
during Ada development.  And some of the Ada development could be
done on cheaper 'A+' compilers.

I believe that this subset would be more widely use by commercial
companies and educational facilities (than Ada presently is). 
For one, the compiler would be cheaper.  Two, you would not have
to pay a licensing fee for every application you sold because you
wouldn't have the specialized runtime.

Plus, It would be great to have something like this for embedded
development.  (I can add my own exception checking if needed.)

2) Create and "strongly support" a standard interface between the
Ada (NOT A+) compiler and the Ada runtime.  First of all, I say
"strongly support" because the compiler vendors could not be
forced to change their runtime designs overnight.  May be force
the standard interface by 1998 or 2000?  A standard interface
would benefit everyone.  Lets face it, a good compiler and a good
runtime do no always go together.  It would create a new market
for third-party Ada runtime's which would give developers a
chance to choose a runtime which would better fit their needs. 
(no more tweaking the one you are forced to use).

Are these bad ideas?

Mark Johnson
Harris Corp (GASD)
mjohnson@su19bb.ess.harris.com

My opinions are my own, but may be shared by others.

"Imperfections are the differences between perfection and
perceived perfection"