Re: Ada's (in)visibility and pricing!

Re: Ada's (in)visibility and pricing!

[paul goffin]

Why hasn't Ada caught on in engineering?  Well one big reason is cost.

A little while ago, I was involved in flight simulation and we were
facing customer pressure to abandon FORTRAN and go to Ada.

Actually, the pressure wasn't very great.  The Statements of Requirements
always said "A High Level Programming Language will be Used, preferably
Ada".

Well, I believe that Ada has many advantages over Ada and, of course
'C' (which is the other language we used), so I was keen to aid the
change.  (In fact I ended up managing the first Ada project we did.)

So what did I mean by cost?  Well, apart from the training costs
and the cost of setting up development systems that actually used
Ada, (New compiler driving scripts, new way of addressing hardware,
new way of managing "global variables" - Yes, nasty as they are,
you do sometimes have to use global variables in the real world!)
there was a nasty shock in the price of the compilers.

What do I mean by nasty?  Well, we were able to use a FORTRAN
compiler on our systems for a cost of $2000 per project.
(ie per host processor.)

The Ada compiler vendors wanted $30 000 PER SEAT.  We even
had the vendors charging different prices for the same compiler
on different hosts.  ie. A compiler/toolset on a Sun SPARC 1 was
$30 000, while on a Sparc 2, the same software was $40 000.

The attitude of the vendors was "We know you've GOT to use Ada,
so just hand over the blank cheque and we'll let you."

They were, of course, wrong.  The cost (in parerwork etc.)
of getting Ada waivers was generally less than the cost of
the compilers - let alone the training costs etc.

So what do my previous employers do now?  Well they still
use FORTRAN and C, and they fight any requirement to use
Ada.  It just costs too much, and when it comes to it,
their customers - even the military ones - place
their orders on price; Ada is still on the wish list.

So, if you want the Ada market to grow, compiler
vendors, get your prices down!

Paul
-- 
+-------------+-------------------------------------------------------+
+ Paul Goffin +  Crosfield Electronics Ltd. U.K.  +44 442 230000x3357 +
+             +  Opinions expressed are mine! (Yours for a small fee) +
+-------------+-------------------------------------------------------+

Re: Ada's (in)visibility and pricing!

[Bob Kitzberger]

pdg@crosfield.co.uk (paul goffin) writes:

>                                                   We even
>had the vendors charging different prices for the same compiler
>on different hosts.  ie. A compiler/toolset on a Sun SPARC 1 was
>$30 000, while on a Sparc 2, the same software was $40 000.

Charging different prices based on the host isn't unique to Ada -- call
Microtec and ask them about their SPARCstation/680x0 C compilers,
and how much they are for SPARC 1s, 2s, etc.  Price varies by host size.

	.Bob.
----------------
Bob Kitzberger          VisiCom Laboratories, Inc.
rlk@visicom.com         10052 Mesa Ridge Court, San Diego CA 92121 USA
                        +1 619 457 2111    FAX +1 619 457 0888

Re: Ada's (in)visibility and pricing!

[agate!usenet.ins.cwru.edu!magnus.acs.ohio-state.edu!zaphod.mps.ohio-state]

In article <15390@suns5.crosfield.co.uk> pdg@crosfield.co.uk (paul goffin) writ
es:

>   Why hasn't Ada caught on in engineering?  Well one big reason is cost.
>
>   [ deleted ]
>
>   So what did I mean by cost?  Well, apart from the training costs
>   and the cost of setting up development systems that actually used
>   Ada, (New compiler driving scripts, new way of addressing hardware,
>   new way of managing "global variables" - Yes, nasty as they are,
>   you do sometimes have to use global variables in the real world!)
>   there was a nasty shock in the price of the compilers.
>
>   What do I mean by nasty?  Well, we were able to use a FORTRAN
>   compiler on our systems for a cost of $2000 per project.
>   (ie per host processor.)
>
>   The Ada compiler vendors wanted $30 000 PER SEAT.  We even
>   had the vendors charging different prices for the same compiler
>   on different hosts.  ie. A compiler/toolset on a Sun SPARC 1 was
>   $30 000, while on a Sparc 2, the same software was $40 000.


I agree, this is insane.  The Sun SPARC 1 compiler is probably
identical to the Sparc 2 ( and if it is not it should be ).  
When I asked one vendor why they had this pricing arrangement
the said that it was because of the precieved power of the
machine; the compiler was in fact identical.  I asked another
vendor why their VMS compiler was ~$50,000 when their unix
compiler was ~$10,000.  Both of these compilers were cross compilers 
to the same target.  The answer boiled down to people
expect to pay more for a VMS compiler.
 
Can anybody give a rational explaination for these pricing
schemes.  For people who what to use Ada pricing the the 
single biggest problem.


--
pat gioannini
pat@psl.nmsu.edu

Re: Ada's (in)visibility and pricing!

[Pascal Obry]

> Can anybody give a rational explaination for these pricing
> schemes.  For people who what to use Ada pricing the the 
> single biggest problem.

They only want to make money. They have already use the reputation of Ada
(very big, very complex, for the Army, very safe ...) to sell compiler 4 time
the price of the others languages.

But it is not enought !! VMS isn't bigger that UNIX, yes !! isn't it more
complex yes !!  of course, so here they find a way to sell at a better
price (i.e. 5 time more) ... that is, a C compiler is not very hard to do, so
with this argument they make more money in any language ...

But it is the same in some other area. If you have a sport car (very fency,
you know) then try to compare the prices of some pieces almost identical ..

If you play golf (In France it is a fency sport) compare the prices of a
golf shoes with almost the same shoes (but not *special* for golf !!) ...

So, the prices, aren't they made from the reputation of the consumer ?

I'am not making an very safe application, I don't work for the army ...
But when I buy a Ada compiler I pay the price for all this.

Pascal.

--

-------------------------------------------------------------------------------
--  Pascal OBRY								     --
--  Room 2D-337				e_mail : obry@bellcore.com  	     --
--  Bellcore								     --
--  445 South Street			voice : 1 - 201 829 4039	     --
--  Post Office Box 1910						     --
--  Morristown, New Jersey 07962-1910					     --
-------------------------------------------------------------------------------

Re: Ada's (in)visibility and pricing!

[dog.ee.lbl.gov!hellgate.utah.edu!cs.utexas.edu!csc.ti.com!tilde.csc.ti.co]

In <PAT.92Sep16084002@peak.opus> pat@peak.opus (pat gioannini) writes:

>I agree, this is insane.  The Sun SPARC 1 compiler is probably
>identical to the Sparc 2 ( and if it is not it should be ).  
>When I asked one vendor why they had this pricing arrangement
>the said that it was because of the precieved power of the
>machine; the compiler was in fact identical.  I asked another
>vendor why their VMS compiler was ~$50,000 when their unix
>compiler was ~$10,000.  Both of these compilers were cross compilers 
>to the same target.  The answer boiled down to people
>expect to pay more for a VMS compiler.
> 
>Can anybody give a rational explaination for these pricing
>schemes.  For people who what to use Ada pricing the the 
>single biggest problem.

I would expect that the pricing is based on the expected number of
simultaneous users (in most cases).  Hence the more powerful the
machine, the more it costs to license software on that machine.  This,
as silly as it seems from my point of view, seems to be the rule
rather than the exception.  It's not just Ada compilers.  Practically
all software seems to license this way.

-- 
"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: Ada's (in)visibility and pricing!

[SAHARBAUGH]

pat gioannini asks:
I agree, this is insane.  The Sun SPARC 1 compiler is probably
identical to the Sparc 2 ( and if it is not it should be ).  
When I asked one vendor why they had this pricing arrangement
the said that it was because of the precieved power of the
machine; the compiler was in fact identical.  I asked another
vendor why their VMS compiler was ~$50,000 when their unix
compiler was ~$10,000.  Both of these compilers were cross compilers 
to the same target.  The answer boiled down to people
expect to pay more for a VMS compiler.
 
Can anybody give a rational explaination for these pricing
schemes.  For people who what to use Ada pricing the the 
single biggest problem.


Ok, I'll take a shot at giving an explanation:

1.  Price and cost are not necessarily related.
2.  In a competitive economy all equivalent products
are priced the same. (This is the argument suppliers
may use when accused of price fixing)
3.  It is comfortable for a vendor to price his
product based on "displaced costs".  e.g., I have
a product to sell you for $1 and you will save $2
every month.  If the buyer has the $1 he will be
very anxious to give it to the vendor.
Now as to the Ada compiler pricing:
If the compiler is for a PC the vendor knows
that it will support one software developer
at say 100 lines of code per minute.
If the same compiler is hosted on a Unix box
with 10 users then 10 software developers
can use it, each at 100 lines per minute (averaged
over say a 30 minute interval).
The user gets X10 benefit, the vendor should get
X10 price.

In general, the economists use a model called
a "demand curve".  The vertical axis is the number
of units sold and the horizontal axis is the price
per unit.  The general shape is y=1/x, i.e., as the
price per unit goes to zero, the number sold goes
to infinity; as the price per unit goes to infinity
the number sold goes to zero.
The vendor looks to maximize revenue and looks to
the demand curve for guidance.  His revenue is
price per unit times number of units sold, or
R=X*Y. He seeks to maximize R.  Geometrically
X*Y is the area of a rectangle whose upper right
corner touches the demand curve.
Now as to Ada compiler pricing:
Who knows the demand curve? Probably noone.  It
changes with the economic winds, competition etc.
The vendor must feel it out.  Have a sale, drop your
price and see what happens.  If R goes up then you
know you are going in the right direction.  If it
goes down then don't continue the sale, and next try
adding a feature or two and raising the price.
( You have heard the joke, "we lose money on each
sale but we make it up in volume" That is not a joke,
it is called "marginal pricing" where you sell below cost
but volume goes up, R goes up, you overabsorb your
fixed cost and make additional profit in excess
of the original loss.  Its a lot of fun if you're
playing with someone else's money.)

I don't vend Ada compilers but I do vend Ada related
services and products.  I have a LOT of sympathy
for Ada compiler vendors.  Their market economics
is distorted by one dominant customer (Uncle Sam, no
relation) and their users are very smart, some smarter
than they are. (What I mean is knowledgable of the
technology, not basic IQ).  They are caught where
the Capitalistic system is trying to serve the
Government.  They aren't allow but they are
They aren't allowed to make much money but they 
are allowed to go broke.  No wonder they look to
C++ where the market forces are "normal".

I hope this helps some.  I would suggest a book
on Product Management for the interested reader.
sam harbaugh saharbaugh@ROO.FIT.EDU        
-----------

Re: Ada's (in)visibility and pricing!

[Michael Feldman]

In article <9209190322.AA21359@ajpo.sei.cmu.edu> SAHARBAUGH@ROO.FIT.EDU writes:

>I don't vend Ada compilers but I do vend Ada related
>services and products.  I have a LOT of sympathy
>for Ada compiler vendors.  Their market economics
>is distorted by one dominant customer (Uncle Sam, no
>relation) and their users are very smart, some smarter
>than they are. (What I mean is knowledgable of the
>technology, not basic IQ).  They are caught where
>the Capitalistic system is trying to serve the
>Government.  They aren't allow but they are
>They aren't allowed to make much money but they 
>are allowed to go broke.  No wonder they look to
>C++ where the market forces are "normal".

By "they" I assume you mean the prime contractors. Where is it written that
a compiler vendor isn't allowed to make "much" money selling the product
commercially?

This may be an off-the-wall analogy, but the military is buying a lot of
standard 1/2-ton pickups these days. Does Chrysler sell these to the
government at an outrageous price? No, everyone knows (approximately)
the price of a pickup, so I would guess that Uncle Sam pays roughly
the same price for a Dodge Dakota that you or I would (or rather that
another fleet owner would). (Am I OK so far?)

Now if Chrysler started trying to sell those Dakotas to you and me for
10x the price of a Chevy equivalent, how many do you think they'd sell?
What the Ada compiler folks don't seem to want to focus on is that the
overall software market is a whole lot bigger than just DoD. If DoD
likes off-the-shelf pickups at regular prices, why not off-the-shelf
compilers at regular prices?

Of course there are special Ada ports for special DoD board configs,
etc., but the mainstream host=target compilers would have a big
potential market if only the vendors would treat a compiler just
like the competition treats a compiler. 

Just so we aren't hand-waving about theoretical numbers here, why not
recall Rich Pattis' list of RISC 6000 prices:

> FYI: I got IBM's RISC 6000 catalog:
>
> AIX XL C++     $ 7,000
> AIX XL FORTRAN $ 6,090
> AIX VS COBOL   $ 8,735
> AIX Ada        $30,570
> AIX XL Pascal  $ 5,235

How ON EARTH can they justify a 4-5x difference for the Ada compiler?
Obviously they do not think the market is going to be large. What they
are doing is creating a self-fulfilling prophecy here. Only those
companies forced by the mandate to do Ada will choose it. The market
will therfore be small and captive, proving their claim that the
market is small. As Charles Colson once said,

"When you've got them by the b*lls, their hearts and minds will follow."

5 or so years ago, the vendors were implying that the prices would drop
once the compilers were mature and a lot of the R&D was written off,
as well as the extraordinarily high first cost of validation. OK, it's
five years later. Why are they still not pricing stuff competitively
(against the other languages, that is)?

I've been told I need a business-school course to understand this.
Perhaps I do. Maybe only in a B-school could they rationalize this kind
of irrationality.

Or can somebody explain it in terms non-B-schools grads can understand?

Cheers -

Mike Feldman