[Fwd: F22 completes 11% of its Flight tests]

[Fwd: F22 completes 11% of its Flight tests]

[Rocky]

[-- Attachment #1: Type: text/plain, Size: 54 bytes --]

Would someone from the Ada community care to join in?

[-- Attachment #2: Type: message/rfc822, Size: 4151 bytes --]

From: "jtarver" <jtarver@tminet.com>
Subject: Re: F22 completes 11% of its Flight tests
Date: Tue, 11 Jan 2000 18:49:05 -0800
Message-ID: <A1Se4.2088$iy5.100593@typ12.deja.bcandid.com>


Harry Andreas <andreas@baloney.org> wrote in message
news:andreas-1101001558400001@x-147-16-144-157.rsc.raytheon.com...
> In article <388299ca.2188436@news.xmission.com>, sferrin@xmission.com (D.
> Scott Ferrin) wrote:
>
> > >The people producing the parts have a contract to do so and there are
also
> > >specialty suppliers that stock obsolete parts for currently operating
> > >airplanes.  The i960 and all Intel's military foundry is gone now.
None of
> > >the specialty suppliers bought a bunch of i960 mil-spec processors
because
> > >there was no market for them.  So as with any unmaintainable airplane
system
> > >the MPP must be replaced for production; as they will be unsupportable
in
> > >the field.  This is that much more powerful than a cray central
processor we
> > >have heard so much about.
> > >
> > >John
> > >
> >
> >
> > So nobody knows this or is it something only you are privy to?  So how
> > do they plan on building these fighters with no chips to go in them?
>
> John and I have had this discussion before. He insists on maintaining
> [pun] this story even though it's not true.

What part of the story is not true Harry?  The way you get to cray like
speed with micros is to use the geometric improvement in processing power
indicated for parallel processors.

> The F-22 CIP was designed from day 1 to be able to easily upgrade the
> processor, and other chip technologies, without re-designing the entire
> system.

Writing in Ada does this for you, supposedly.

> This was a key feature of PAVE PACE architectures which are de rigeur
> today but which F-22 was the first real implementation.
> We knew during the YF and demval phases that we would be using different
> processor chips for production. It's unavoidable given today's fast moving
> market.

The processor used in the MPP was already obsolete by then.

>  The application software is written without direct calls to the hardware.
> Everything goes through the OS which acts as a buffer between the physical
> hardware and the application code. So when your processor chip goes off
> market, you replace it and part of the OS and theoretically the tactical
> software doesn't have to change [which is the biggest cost item]. The
> devil is in the details however. [BTW, I'm no software guy. There's a
> million details that makes this possible.]

It will be highly impressive to see a version of Ada that is actually
transportable.  The first thing the development of Ada compilers did was
allow extensions in direct violation of the charter for the system.  This
worked to deliver the pork to Boston, but really comprimised the intent of
the DARPA program funding Ada.

John

Re: [Fwd: F22 completes 11% of its Flight tests]

[Marin D. Condic]

> 
> It will be highly impressive to see a version of Ada that is actually
> transportable.  The first thing the development of Ada compilers did was
> allow extensions in direct violation of the charter for the system.  This
> worked to deliver the pork to Boston, but really comprimised the intent of
> the DARPA program funding Ada.
> 
> John

I worked F22 engine control software written in Ada for years - and may
yet again someday. We were having a similar problem with respect to the
M68040 becoming obsolete and needing to be replaced (probably by the
PowerPC) as the engine goes into production. I'll throw my $0.02 worth
in:

There is not now, nor has there ever been, *any* computer language that
is 100% portable. *Especially* in the field of embedded computers. So
when portability comes up, you have to ask the question as a matter of
degree.

Ada is the most portable language I have ever worked with. I have
dragged large chunks of Ada code out of an embedded environment and
compiled it with zero modifications on a Sun workstation and on a PC and
had it run just as it did in the embedded box. I have had very similar
experiences with a large number of other bodies of code (non-embedded)
so I don't think this was unique.

The trick with programming long lived embedded computers is knowing how
to isolate the things which are going to be processor/system specific.
That way, when you have to upgrade to new hardware, the problem isn't as
horrific. The biggest problem is that changing anything in the box means
you have to go through huge amounts of revalidation of the system, and
this is an issue independent of the processor or language you choose to
use. Ada has always been a big help in enhancing portability, but it can
never solve all of the problems. No language can.

MDC
-- 
=============================================================
Marin David Condic   - Quadrus Corporation -   1.800.555.3393
1015-116 Atlantic Boulevard, Atlantic Beach, FL 32233
http://www.quadruscorp.com/
m c o n d i c @ q u a d r u s c o r p . c o m

Visit my web site at:  http://www.mcondic.com/

"Capitalism without failure is like religion without sin." 
        --  Allan Meltzer, Economist 
=============================================================

Re: [Fwd: F22 completes 11% of its Flight tests]

[jtarver]

Marin D. Condic <mcondic-nospam@quadruscorp.com> wrote in message
news:387CC1C0.4C57E34C@quadruscorp.com...
> >
> > It will be highly impressive to see a version of Ada that is actually
> > transportable.  The first thing the development of Ada compilers did was
> > allow extensions in direct violation of the charter for the system.
This
> > worked to deliver the pork to Boston, but really comprimised the intent
of
> > the DARPA program funding Ada.
> >
> > John
>
> I worked F22 engine control software written in Ada for years - and may
> yet again someday. We were having a similar problem with respect to the
> M68040 becoming obsolete and needing to be replaced (probably by the
> PowerPC) as the engine goes into production. I'll throw my $0.02 worth
> in:
>
> There is not now, nor has there ever been, *any* computer language that
> is 100% portable. *Especially* in the field of embedded computers. So
> when portability comes up, you have to ask the question as a matter of
> degree.

"It all depends", as they say.

> Ada is the most portable language I have ever worked with. I have
> dragged large chunks of Ada code out of an embedded environment and
> compiled it with zero modifications on a Sun workstation and on a PC and
> had it run just as it did in the embedded box. I have had very similar
> experiences with a large number of other bodies of code (non-embedded)
> so I don't think this was unique.

Ada was originally envisioned to be platform transparent.  When Data General
produced the only 100% compliant compliler and the Ada development program
went for the noncompliant VAX as their standard platform that idea was out
the window.

> The trick with programming long lived embedded computers is knowing how
> to isolate the things which are going to be processor/system specific.
> That way, when you have to upgrade to new hardware, the problem isn't as
> horrific. The biggest problem is that changing anything in the box means
> you have to go through huge amounts of revalidation of the system, and
> this is an issue independent of the processor or language you choose to
> use. Ada has always been a big help in enhancing portability, but it can
> never solve all of the problems. No language can.

Well you know I'd like to argue with you, but there is little in what you
wrote to dispute.

John

Re: [Fwd: F22 completes 11% of its Flight tests]

[Bill Greene]

jtarver wrote:

> Ada was originally envisioned to be platform transparent.  When Data General
> produced the only 100% compliant compliler and the Ada development program
> went for the noncompliant VAX as their standard platform that idea was out
> the window.

I don't understand your statement.  IIRC, there was a validated Ada 83
compiler for the Data General (written by Rational?) and there were
several validated Ada 83 compilers for the DEC VAX (written by SofTech,
DEC, and others).  What do you mean by "noncompliant VAX"?

FWIW, my experiences with porting Ada (83 or 95) code have been quite
satisfactory.

-- 
William R. Greene                              1100 Perimeter Park Drive
Ganymede Software, Inc.                                        Suite 104
http://www.ganymede.com                       Morrisville, NC  27560 USA
Phone: (919) 469-0997, ext. 280                      Fax: (919) 469-5553

Re: [Fwd: F22 completes 11% of its Flight tests]

[David Tannen]

Hi,

>I mean the first rule of the compliance was that there be no
extensions and
>VAX immediately violated that rule.


Could you specify exactly what you mean by "no extensions and
VAX immediately violated that rule"?  I have worked extensively
with the DEC Ada compiler and I did not see any "extensions" outside
of the LRM.

Oh by the way, VAX doesn't produce the manual or spec that is V&V
(ie ACVC tested).  Ada is defined by the LRM.  That is what every
vendor has to test against.  I know I use to work for an Ada vendor
and testing using the ACVCs was part of my responsibility.

>> FWIW, my experiences with porting Ada (83 or 95) code have been
quite
>> satisfactory.
>
>What did you port as the new platform?


I have been involved in maintaining code (about 1msloc) that
targeted three different processors and ran on three different
hosts (9 combinations).  With all of that code we had about
90% of the code in common and 10% varied because of the
platform or the target h/w.

Ada is a very portable language and the features it provides
make it extremely portable.  If an application isn't portable
that is fault of the developers and most importantly the
managers and PMs for that project.

An ambassador for Christ
David Tannen (tannen@jcdisciples.org)
John 14:21 "Whoever has my commands and obeys them, he is the one who
            loves me.  He who loves me will be loved by my Father,
            and I too will love him and show myself to him."
Check Out:   http://www.jcdisciples.org/davidtannen/index.html

Re: [Fwd: F22 completes 11% of its Flight tests]

[Harry Andreas]

In article <85j53h$ic0@chronicle.concentric.net>, "David Tannen"
<tannen@jcdisciples.org> wrote:

> I have been involved in maintaining code (about 1msloc) that
> targeted three different processors and ran on three different
> hosts (9 combinations).  With all of that code we had about
> 90% of the code in common and 10% varied because of the
> platform or the target h/w.
> 
> Ada is a very portable language and the features it provides
> make it extremely portable.  If an application isn't portable
> that is fault of the developers and most importantly the
> managers and PMs for that project.

This has pretty much been our experience, too, and is what I was alluding
to in a prior message.
Ada can be written to be portable if care is taken to make it happen.
And on this program it was.

-- 
Harry Andreas
the engineering raconteur

replace baloney with computer to reply

Re: [Fwd: F22 completes 11% of its Flight tests]

[Bill Greene]

jtarver wrote:
> 
> Bill Greene <wrg@Ganymede.com> wrote in message
> news:387CEE4A.3965@Ganymede.com...
> > jtarver wrote:
> >
> > > Ada was originally envisioned to be platform transparent.  When Data
> General
> > > produced the only 100% compliant compliler and the Ada development
> program
> > > went for the noncompliant VAX as their standard platform that idea was
> out
> > > the window.
> >
> > I don't understand your statement.  IIRC, there was a validated Ada 83
> > compiler for the Data General (written by Rational?) and there were
> > several validated Ada 83 compilers for the DEC VAX (written by SofTech,
> > DEC, and others).  What do you mean by "noncompliant VAX"?
> 
> I mean the first rule of the compliance was that there be no extensions and
> VAX immediately violated that rule.

The only mechanism I know of for gauging compliance is formal
validation, which included passing the thousands of tests in the ACVC
(now ACATS) test suite.  There were and are a number of validated Ada
compilers for the VAX.

But what does it mean to say that "VAX" violated a rule?  Are you
talking about a particular Ada compiler for the VAX?  If so, which
compiler is it and what sort of extensions are you referring to? 
Implementation-defined attributes and pragmas?

Also, it's not entirely accurate to say that Ada was intended to be
platform transparent.
One of the design goals for the language was to allow low-level access
to the underlying architecture, e.g., in code inserts, which are
certainly not portable.  But then, no one writing a code insert for a
VAX would ever expect that code to be portable to a non-VAX platform.

> 
> > FWIW, my experiences with porting Ada (83 or 95) code have been quite
> > satisfactory.
> 
> What did you port as the new platform?

I've ported among VAX, Sun, Apollo, Windows, OS/2, IBM mainframes, and
Macintosh, to name a few off the top of my head.

-- 
William R. Greene                              1100 Perimeter Park Drive
Ganymede Software, Inc.                                        Suite 104
http://www.ganymede.com                       Morrisville, NC  27560 USA
Phone: (919) 469-0997, ext. 280                      Fax: (919) 469-5553

Re: [Fwd: F22 completes 11% of its Flight tests]

[jtarver]

Bill Greene <wrg@Ganymede.com> wrote in message
news:387D011B.2DC@Ganymede.com...
> jtarver wrote:
> >
> > Bill Greene <wrg@Ganymede.com> wrote in message
> > news:387CEE4A.3965@Ganymede.com...
> > > jtarver wrote:
> > >
> > > > Ada was originally envisioned to be platform transparent.  When Data
> > General
> > > > produced the only 100% compliant compliler and the Ada development
> > program
> > > > went for the noncompliant VAX as their standard platform that idea
was
> > out
> > > > the window.
> > >
> > > I don't understand your statement.  IIRC, there was a validated Ada 83
> > > compiler for the Data General (written by Rational?) and there were
> > > several validated Ada 83 compilers for the DEC VAX (written by
SofTech,
> > > DEC, and others).  What do you mean by "noncompliant VAX"?
> >
> > I mean the first rule of the compliance was that there be no extensions
and
> > VAX immediately violated that rule.
>
> The only mechanism I know of for gauging compliance is formal
> validation, which included passing the thousands of tests in the ACVC
> (now ACATS) test suite.  There were and are a number of validated Ada
> compilers for the VAX.

LOL.  Validating that the VAX does what it says it does is not the same as
the compiler meeting the Ada specification.

> But what does it mean to say that "VAX" violated a rule?  Are you
> talking about a particular Ada compiler for the VAX?  If so, which
> compiler is it and what sort of extensions are you referring to?
> Implementation-defined attributes and pragmas?

"extended instruction set" is what I am refering to.

> Also, it's not entirely accurate to say that Ada was intended to be
> platform transparent.

That is exactly the original selling point for the funding of Ada.

> One of the design goals for the language was to allow low-level access
> to the underlying architecture, e.g., in code inserts, which are
> certainly not portable.  But then, no one writing a code insert for a
> VAX would ever expect that code to be portable to a non-VAX platform.

The compiler would necessarily be in the immage of the hardware, but the
code itself was to be independent.

> > > FWIW, my experiences with porting Ada (83 or 95) code have been quite
> > > satisfactory.
> >
> > What did you port as the new platform?
>
> I've ported among VAX, Sun, Apollo, Windows, OS/2, IBM mainframes, and
> Macintosh, to name a few off the top of my head.

VAX to Apollo in Ada sucessfully?  How much effort was the transfer?

John

Re: [Fwd: F22 completes 11% of its Flight tests]

[Bill Greene]

jtarver wrote:
> > The only mechanism I know of for gauging compliance is formal
> > validation, which included passing the thousands of tests in the ACVC
> > (now ACATS) test suite.  There were and are a number of validated Ada
> > compilers for the VAX.
> 
> LOL.  Validating that the VAX does what it says it does is not the same as
> the compiler meeting the Ada specification.

The ACVC suite does not "validate that the VAX does what it says it
does."
Individual ACVC tests are designed (and named) to establish a compiler's
compliance to specific chapters and paragraphs of the Ada Reference
Manual, which *is* the Ada specification.

> 
> > But what does it mean to say that "VAX" violated a rule?  Are you
> > talking about a particular Ada compiler for the VAX?  If so, which
> > compiler is it and what sort of extensions are you referring to?
> > Implementation-defined attributes and pragmas?
> 
> "extended instruction set" is what I am refering to.

???

> VAX to Apollo in Ada sucessfully?  How much effort was the transfer?

I had to run the compiler and linker.

-- 
William R. Greene                              1100 Perimeter Park Drive
Ganymede Software, Inc.                                        Suite 104
http://www.ganymede.com                       Morrisville, NC  27560 USA
Phone: (919) 469-0997, ext. 280                      Fax: (919) 469-5553

Re: [Fwd: F22 completes 11% of its Flight tests]

[Bill Greene]

jtarver wrote:
> 
> Bill Greene <wrg@Ganymede.com> wrote in message
> news:387D0C1B.9D2@Ganymede.com...
> > The ACVC suite does not "validate that the VAX does what it says it
> > does."
> > Individual ACVC tests are designed (and named) to establish a compiler's
> > compliance to specific chapters and paragraphs of the Ada Reference
> > Manual, which *is* the Ada specification.
> 
> Who wrote the Ada reference manual you V/V to?  Was it Digital Equipment?

For Ada 83, "The Reference Manual for the Ada Programming Language" is
ANSI/MIL-STD-1815A-1983.  It's also FIPS-PUB-119/NAC.  It's also an ISO
standard, but I don't have the document number with me.  You can find a
link to the individual chapters at
http://www.adaic.com/standards/ada83.html.  No single author is
identified.

Nowadays, Ada means Ada 95, and links to its reference manual and
rationale are at http://www.adaic.com/standards/ada95.html.

HTH.
-- 
William R. Greene                              1100 Perimeter Park Drive
Ganymede Software, Inc.                                        Suite 104
http://www.ganymede.com                       Morrisville, NC  27560 USA
Phone: (919) 469-0997, ext. 280                      Fax: (919) 469-5553

Re: [Fwd: F22 completes 11% of its Flight tests]

[Keith Thompson]

"jtarver" <jtarver@tminet.com> writes:
> Bill Greene <wrg@Ganymede.com> wrote in message
> news:387D0C1B.9D2@Ganymede.com...
> Who wrote the Ada reference manual you V/V to?  Was it Digital Equipment?

No.  The Ada reference manual and the validation suite are the same
for all compilers and systems.

> > > > But what does it mean to say that "VAX" violated a rule?  Are you
> > > > talking about a particular Ada compiler for the VAX?  If so, which
> > > > compiler is it and what sort of extensions are you referring to?
> > > > Implementation-defined attributes and pragmas?
> > >
> > > "extended instruction set" is what I am refering to.
> >
> > ???
> 
> That is what I thought.

What is what you thought?

I'm afraid I have no idea what you mean by "extended instruction set".
Certainly Ada compilers for the VAX can provide
implementation-specific pragmas and attributes, this is allowed by the
standard.  They also provide machine-code insertions (which aren't
expected to be portable) and additional library packages (which are
implemented *in* Ada; they aren't extensions to the language itself).
None of this makes it particularly difficult to write portable code
that will work properly on a VAX or anywhere else.

Can you provide an example of what you're talking about?

-- 
Keith Thompson (The_Other_Keith) kst@cts.com  <http://www.ghoti.net/~kst>
San Diego Supercomputer Center           <*>  <http://www.sdsc.edu/~kst>
Welcome to the last year of the 20th century.

Re: [Fwd: F22 completes 11% of its Flight tests]

[Richard D Riehle]

In article <Dt7f4.2222$iy5.92651@typ12.deja.bcandid.com>,
	"jtarver" <jtarver@tminet.com> wrote:

>
>Bill Greene <wrg@Ganymede.com> wrote in message
>news:387D011B.2DC@Ganymede.com...

>> Also, it's not entirely accurate to say that Ada was intended to be
>> platform transparent.
>
>That is exactly the original selling point for the funding of Ada.

Well, not exactly.  One of the selling points for a language such
as Ada was portability, not transparency.  There is a difference.
In practice, most of Ada software has been portable.  The exceptions
have been low-level routines that incorporate platform-specific
information in their implementation.  For example, the interrupt
scheme for a VAX will be quite different from that on an PowerPC.

I think the extensions you cite are packages provided by the compiler
publisher that enable the programmer to take advantage of capabilities
of the targeted platform.  Such packages are quite common.  They do
not extend the language.  Rather, they use the abstraction model of
Ada exactly as originally intended.  

When measured against other languages for portability, Ada turns out
to be one of the most portable languages available -- especially the
current Ada 95 standard.  It is even more portable than Java since it
can be used in place of Java to create applets that run on the JVM. 

As to "selling points," one needs to look at the many other selling
points of Ada, as well.  These include one very big point:  the ability
of the compiler to catch the maximum number of errors as early as possible
in the development process.  From the perspective of software economics,
this turns out to be one of the most significant of Ada's "selling points."

Richard Riehle

 

Re: [Fwd: F22 completes 11% of its Flight tests]

[Ted Dennison]

In article <Dt7f4.2222$iy5.92651@typ12.deja.bcandid.com>,
  "jtarver" <jtarver@tminet.com> wrote:

> VAX to Apollo in Ada sucessfully?  How much effort was the transfer?

FWIW: We wrote a real-time scheduler in Ada that compiles and runs
withouth modification on vxWorks using GreenHills AdaMulti and on NT
using Gnat. Obviously it isn't exactly real time when running on NT, but
we can hardly blame Ada for that. :-)

Our entire flight sim would probably be runnable non-realtime on NT if
you could configure an NT PC with the right hardware.

--
T.E.D.

http://www.telepath.com/~dennison/Ted/TED.html


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: [Fwd: F22 completes 11% of its Flight tests]

[Jeff Carter]

In article <85ko9c$7e$1@nnrp1.deja.com>,
  Ted Dennison <dennison@telepath.com> wrote:
> In article <Dt7f4.2222$iy5.92651@typ12.deja.bcandid.com>,
>   "jtarver" <jtarver@tminet.com> wrote:
>
> > VAX to Apollo in Ada sucessfully?  How much effort was the transfer?
>
> FWIW: We wrote a real-time scheduler in Ada that compiles and runs
> withouth modification on vxWorks using GreenHills AdaMulti and on NT
> using Gnat. Obviously it isn't exactly real time when running on NT,
but
> we can hardly blame Ada for that. :-)
>
> Our entire flight sim would probably be runnable non-realtime on NT if
> you could configure an NT PC with the right hardware.

Those of us with actual experience with Ada do not consider this
unusual. I was involved in real-time communications protocol stacks for
a big-endian PowerPC under VxWorks that were used without change on a
little-endian PC under Win9X. I have developed numerous applications
that ran without change on VAXs, UNIX boxes, and PCs. One has to
conciously intend the code to be portable, but other than that
there's "nothing to it".
--
Jeff Carter
"Now go away or I shall taunt you a second time."
-- Monty Python and the Holy Grail


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: [Fwd: F22 completes 11% of its Flight tests]

[Tarjei T. Jensen]

jtarver wrote :
>
>So we have your personal guarantee the hardware change will be completely
>trouble free?


Nobody can guarantee that you won't make stupid choices.

Ada is a tool that enables you to create portable code, but it will not force
you to do so.


Greetings,

Re: [Fwd: F22 completes 11% of its Flight tests]

[jtarver]

Ted Dennison <dennison@telepath.com> wrote in message
news:85ko9c$7e$1@nnrp1.deja.com...
> In article <Dt7f4.2222$iy5.92651@typ12.deja.bcandid.com>,
>   "jtarver" <jtarver@tminet.com> wrote:
>
> > VAX to Apollo in Ada sucessfully?  How much effort was the transfer?
>
> FWIW: We wrote a real-time scheduler in Ada that compiles and runs
> withouth modification on vxWorks using GreenHills AdaMulti and on NT
> using Gnat. Obviously it isn't exactly real time when running on NT, but
> we can hardly blame Ada for that. :-)

Sure I could.

> Our entire flight sim would probably be runnable non-realtime on NT if
> you could configure an NT PC with the right hardware.

Cool.

John

> http://www.telepath.com/~dennison/Ted/TED.html

Re: [Fwd: F22 completes 11% of its Flight tests]

[Marin D. Condic]

jtarver wrote:
> I mean the first rule of the compliance was that there be no extensions and
> VAX immediately violated that rule.
> 
Not exactly true. Extensions have always been allowed if, by
"extensions" you mean packages which are provided by the compiler for
features not provided in the standard. (Math libraries, etc.) If by
"extensions" you mean language constructs not provided for by the syntax
of the language, then this is also not exactly accurate. The requirement
is that the compiler provide a mode in which it will parse the standard
Ada syntax and reject anything that is not standard. However, if the
compiler has an "extended" mode in which it will interpret non-standard
syntax, that is quite all right too. The goal here was that a project
should have a means of restricting itself to the standard language to
enhance portability, but if that were not a concern and a compiler had
additional capabilities, it should be quite all right to use those
facilities.

MDC
-- 
=============================================================
Marin David Condic   - Quadrus Corporation -   1.800.555.3393
1015-116 Atlantic Boulevard, Atlantic Beach, FL 32233
http://www.quadruscorp.com/
m c o n d i c @ q u a d r u s c o r p . c o m

Visit my web site at:  http://www.mcondic.com/

"Capitalism without failure is like religion without sin." 
        --  Allan Meltzer, Economist 
=============================================================

Re: [Fwd: F22 completes 11% of its Flight tests]

[Ed Falis]

In article <387F8E50.11D27E14@quadruscorp.com>,
  "Marin D. Condic" <mcondic-nospam@quadruscorp.com> wrote:
 The requirement
> is that the compiler provide a mode in which it will parse the standard
> Ada syntax and reject anything that is not standard. However, if the
> compiler has an "extended" mode in which it will interpret non-standard
> syntax, that is quite all right too. The goal here was that a project
> should have a means of restricting itself to the standard language to
> enhance portability, but if that were not a concern and a compiler had
> additional capabilities, it should be quite all right to use those
> facilities.

I can't think of any Ada compiler that went that far, though.  All
"extensions" I'm aware of have either been domain-oriented packages
(encouraged by the standard), or implementation-defined pragmas
(allowed by the language).  DEC Ada featured mostly the former in its
Starlet package that enabled access to OS facilities, much as the
POSIX/Ada binding enabled access to POSIX compliant OS's.

- Ed


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: [Fwd: F22 completes 11% of its Flight tests]

[Bill Greene]

Ed Falis wrote:
> 
> In article <387F8E50.11D27E14@quadruscorp.com>,
>   "Marin D. Condic" <mcondic-nospam@quadruscorp.com> wrote:
>  The requirement
> > is that the compiler provide a mode in which it will parse the standard
> > Ada syntax and reject anything that is not standard. However, if the
> > compiler has an "extended" mode in which it will interpret non-standard
> > syntax, that is quite all right too. The goal here was that a project
> > should have a means of restricting itself to the standard language to
> > enhance portability, but if that were not a concern and a compiler had
> > additional capabilities, it should be quite all right to use those
> > facilities.
> 
> I can't think of any Ada compiler that went that far, though.  All
> "extensions" I'm aware of have either been domain-oriented packages
> (encouraged by the standard), or implementation-defined pragmas
> (allowed by the language).  DEC Ada featured mostly the former in its
> Starlet package that enabled access to OS facilities, much as the
> POSIX/Ada binding enabled access to POSIX compliant OS's.

Earlier in this somewhat-long thread I asked:
> But what does it mean to say that "VAX" violated a rule?  Are you
> talking about a particular Ada compiler for the VAX?  If so, which
> compiler is it and what sort of extensions are you referring to?
> Implementation-defined attributes and pragmas?
and jtarver replied
> "extended instruction set" is what I am refering to.

Perhaps he is really complaining about the instruction set of a CPU or
processor architecture (the VAX), and not about a programming language
(Ada) at all.  Note his assertion that "VAX" violated some alleged
design goal of Ada.  Having tried unsuccessfully to discover his meaning
by direct questions, all I can conclude is that he is blaming the Ada
programming language (which he apparently thought was designed by DEC)
for the fact that the VAX is non-compliant with the Data General
instruction set!

-- 
William R. Greene                              1100 Perimeter Park Drive
Ganymede Software, Inc.                                        Suite 104
http://www.ganymede.com                       Morrisville, NC  27560 USA
Phone: (919) 469-0997, ext. 280                      Fax: (919) 469-5553

Re: [Fwd: F22 completes 11% of its Flight tests]

[Keith Willshaw]

Bill Greene wrote in message <387FCA73.3A61@Ganymede.com>...
>Ed Falis wrote:
>
>Perhaps he is really complaining about the instruction set of a CPU or
>processor architecture (the VAX), and not about a programming language
>(Ada) at all.  Note his assertion that "VAX" violated some alleged
>design goal of Ada.  Having tried unsuccessfully to discover his meaning
>by direct questions, all I can conclude is that he is blaming the Ada
>programming language (which he apparently thought was designed by DEC)
>for the fact that the VAX is non-compliant with the Data General
>instruction set!
>


More liley its DEC's own compiler he's complaing
about. Their Fortran compilers had lots of extensions
that were far from ANSI standard Fortran.

My company had to move a lot of this stuff from
VMS to Unix and then NT

The DEC extensions were a major headache

Keith

Re: [Fwd: F22 completes 11% of its Flight tests]

[Marin D. Condic]

Keith Willshaw wrote:
> 
> More liley its DEC's own compiler he's complaing
> about. Their Fortran compilers had lots of extensions
> that were far from ANSI standard Fortran.
> 
But DEC's Ada compiler was validated and, in my experience, very
portable. You had to stay away from Starlet and other similar compiler
provided packages if you wanted to port, but if you did so, the port was
usually just a matter of recompilation. DEC had one of the better Ada
compilers out there and I am unaware of any extensions to the syntax of
the language within their compiler.

DEC's Fortran compilers often had language extensions (anybody remember
Fortran-10 for the DECSystem 10?) which enabled them to utilize system
dependent features and I know that porting their Fortran somewhere else
could be painful. But I don't recall that they ever did anything similar
with Ada.

MDC
-- 
=============================================================
Marin David Condic   - Quadrus Corporation -   1.800.555.3393
1015-116 Atlantic Boulevard, Atlantic Beach, FL 32233
http://www.quadruscorp.com/
m c o n d i c @ q u a d r u s c o r p . c o m

Visit my web site at:  http://www.mcondic.com/

"Capitalism without failure is like religion without sin." 
        --  Allan Meltzer, Economist 
=============================================================

Re: [Fwd: F22 completes 11% of its Flight tests]

[Keith Willshaw]

Marin D. Condic wrote in message <3880CCC7.261957BC@quadruscorp.com>...
>Keith Willshaw wrote:
>>
>> More liley its DEC's own compiler he's complaing
>> about. Their Fortran compilers had lots of extensions
>> that were far from ANSI standard Fortran.
>>
>But DEC's Ada compiler was validated and, in my experience, very
>portable. You had to stay away from Starlet and other similar compiler
>provided packages if you wanted to port, but if you did so, the port was
>usually just a matter of recompilation. DEC had one of the better Ada
>compilers out there and I am unaware of any extensions to the syntax of
>the language within their compiler.
>

OK I don't know ADA so can't comment.

>DEC's Fortran compilers often had language extensions (anybody remember
>Fortran-10 for the DECSystem 10?) which enabled them to utilize system
>dependent features and I know that porting their Fortran somewhere else
>could be painful. But I don't recall that they ever did anything similar
>with Ada.
>


Fair enough its just having been bitten by DEC Fortran
I tend to be a little suspicious of VAX developed software.

Keith

Re: [Fwd: F22 completes 11% of its Flight tests]

[Gautier]

> Fair enough its just having been bitten by DEC Fortran
> I tend to be a little suspicious of VAX developed software.

(About DEC Ada) You can write perfectly portable code with it;
just ignore the extensions and you have `casher' Ada 83...

Of course I guess it is simpler in my case (solving partial differential
equations) than for technical computing which is made easier with
Ada 95 standard features - the Interfaces packages etc.

Note that I'm using the same finite element code to solve the
equations (on an AXP/OpenVMS) and for the plottings on PC
with GNAT - not a comma to change!

-- 
Gautier

_____\\________________\_______\
http://members.xoom.com/gdemont/

Re: [Fwd: F22 completes 11% of its Flight tests]

[Gautier]

Keith Willshaw wrote:

> >(About DEC Ada) You can write perfectly portable code with it;
> >just ignore the extensions and you have `casher' Ada 83...

> >Of course I guess it is simpler in my case (solving partial differential
> >equations) than for technical computing which is made easier with
> >Ada 95 standard features - the Interfaces packages etc.

> >Note that I'm using the same finite element code to solve the
> >equations (on an AXP/OpenVMS) and for the plottings on PC
> >with GNAT - not a comma to change!

> Its of course perfectly possible to that with any language

With many, surely! My remark was not to `sell' Ada but to clarify
that DEC Ada is not incompatible with the standard - it covers
a large portion of what can be programmed...

Just curious:
 * Do you have also _one_ makefile ?
 * Is there no conditional defines in the source code files that make differences
   between the Windows NT/9x ,SG Irix, Solaris, Dec Unix, RS6000 and HP Unix
   platforms ?

I ask it because many resources (in C) on the Web claim to be portable
but in fact are only ported on many platforms, with conditional defines
that switch even basic type definitions, memory management etc.
e.g. Info-Zip stuff.

-- 
Gautier

_____\\________________\_______\_________
http://members.xoom.com/gdemont/gsoft.htm

Re: [Fwd: F22 completes 11% of its Flight tests]

[Keith Willshaw]

Gautier wrote in message <38826BCE.6FB484B3@maths.unine.ch>...

>
>Just curious:
> * Do you have also _one_ makefile ?

Yes

> * Is there no conditional defines in the source code files that make
differences
>   between the Windows NT/9x ,SG Irix, Solaris, Dec Unix, RS6000 and HP
Unix
>   platforms ?
>

We use machine dependent libraries to handle the
differences between implementations. This is mainly
for issues like  FileIO Graphics etc.

The idea is that the calls to those machine dependent libraries are
the same what ever implementation is required. Once a set
of these libraries is stable they rarely change

In reality there are two main graphics libraries
Windows and Motif and the  file I/O parts are
also essentially split between UNIX and Win9x
with same differences in the RS implementation

>I ask it because many resources (in C) on the Web claim to be portable
>but in fact are only ported on many platforms, with conditional defines
>that switch even basic type definitions, memory management etc.
>e.g. Info-Zip stuff.
>


No thats a pain

We actually have 4 develiopers sharing the same source
code mounted off our main Sun Server.

Each has a Unix Box of different type and a PC

They simply compile for each target from the same code.
There's a script which checks the machine ID and
links the correct machine libraries.

Keith

Re: [Fwd: F22 completes 11% of its Flight tests]

[Keith Thompson]

"jtarver" <jtarver@tminet.com> writes:
> Nothing like having the rules change to deliver that Boston pork.  DEC had
> extensions when their computer became the defacto standard for Ada.  At the
> time any extensions to Ada were forbidden.  Data General was rightiously
> upset when this was done.  I was working SDI in the 80s when Ada became the
> perfered controls language.

Please be specific.  What extensions are you referring to?  (BTW, I
was working on Ada compilers in the 80s.)

> Ada suxs.

Yeah, whatever.

-- 
Keith Thompson (The_Other_Keith) kst@cts.com  <http://www.ghoti.net/~kst>
San Diego Supercomputer Center           <*>  <http://www.sdsc.edu/~kst>
Welcome to the last year of the 20th century.

Re: [Fwd: F22 completes 11% of its Flight tests]

[jtarver]

Keith Thompson <kst@cts.com> wrote in message
news:yecaem64cny.fsf@king.cts.com...
> "jtarver" <jtarver@tminet.com> writes:
> > Nothing like having the rules change to deliver that Boston pork.  DEC
had
> > extensions when their computer became the defacto standard for Ada.  At
the
> > time any extensions to Ada were forbidden.  Data General was rightiously
> > upset when this was done.  I was working SDI in the 80s when Ada became
the
> > perfered controls language.
>
> Please be specific.  What extensions are you referring to?  (BTW, I
> was working on Ada compilers in the 80s.)

I am refering to VAX Ada compilers with extended instruction sets.  Do you
have a reading comprehension problem?

John

Re: [Fwd: F22 completes 11% of its Flight tests]

[Gautier]

> > Please be specific.  What extensions are you referring to?  (BTW, I
> > was working on Ada compilers in the 80s.)
> 
> I am refering to VAX Ada compilers with extended instruction sets.  Do you
> have a reading comprehension problem?

Interesting - I'm using DEC Ada for 3 years and I've never seen a single
instruction more than standard Ada!
Which one(s) ?!

-- 
Gautier

_____\\________________\_______\
http://members.xoom.com/gdemont/

Re: [Fwd: F22 completes 11% of its Flight tests]

[Keith Thompson]

I should really just give up on this thread; it's showing no signs of
going anywhere.

"jtarver" <jtarver@tminet.com> writes:
> Keith Thompson <kst@cts.com> wrote in message
> news:yecaem64cny.fsf@king.cts.com...
> > "jtarver" <jtarver@tminet.com> writes:
> > > Nothing like having the rules change to deliver that Boston pork.
> > > DEC had extensions when their computer became the defacto standard
> > > for Ada.  At the time any extensions to Ada were forbidden.  Data
> > > General was rightiously upset when this was done.  I was working
> > > SDI in the 80s when Ada became the perfered controls language.
> >
> > Please be specific.  What extensions are you referring to?  (BTW, I
> > was working on Ada compilers in the 80s.)
> 
> I am refering to VAX Ada compilers with extended instruction sets.  Do you
> have a reading comprehension problem?

Not usually, but some writers do give me some trouble.  Let's try again.

I don't know what you mean by "extended instruction sets" in this
context.  The phrase "instruction set" usually refers to the set of
instructions that can be executed by a specific kind of CPU.  In this
sense, an extended instruction set would be relevant to users of an
Ada compiler only if they're using machine code insertions.  MCIs
aren't generally considered to be a language extension, though I
suppose you could look at them that way.  They're a facility for
writing deliberately non-portable code, and in my experience they're
rarely used.  (My experience may be atypical; the Ada code I wrote had
to run on multiple platforms, so of course it couldn't use MCIs.)

If you're referring to the instruction set the compiler uses for
generated code, of course that's going to be non-portable, but that
has nothing to do with extensions to Ada.

If you're referring to something else, you're not using the term
"instruction set" in the same sense most of us do.

You refer to extensions to Ada implemented in VAX Ada compilers.
Again, what extensions are you talking about?  A snippet of sample
code that depends on such an extension would be helpful.

(Note that implementation-defined pragmas and attributes have always
been specifically allowed by the language standard, and
implementation-defined library units are not extensions to the
language.)

-- 
Keith Thompson (The_Other_Keith) kst@cts.com  <http://www.ghoti.net/~kst>
San Diego Supercomputer Center           <*>  <http://www.sdsc.edu/~kst>
Welcome to the last year of the 20th century.

Re: [Fwd: F22 completes 11% of its Flight tests]

[David Starner]

On Sun, 16 Jan 2000 09:26:46 -0800, jtarver <jtarver@tminet.com> wrote:
>
>Keith Thompson <kst@cts.com> wrote in message
>news:yecaem64cny.fsf@king.cts.com...
>> "jtarver" <jtarver@tminet.com> writes:
>> > Nothing like having the rules change to deliver that Boston pork.  DEC
>had
>> > extensions when their computer became the defacto standard for Ada.  At
>the
>> > time any extensions to Ada were forbidden.  Data General was rightiously
>> > upset when this was done.  I was working SDI in the 80s when Ada became
>the
>> > perfered controls language.
>>
>> Please be specific.  What extensions are you referring to?  (BTW, I
>> was working on Ada compilers in the 80s.)
>
>I am refering to VAX Ada compilers with extended instruction sets.  Do you
>have a reading comprehension problem?

Do you have a reading comprehension problem? No one knows what you mean by
that, so repeating it won't do any good. What do you mean by 
"extended instruction sets"? 

-- 
David Starner - dstarner98@aasaa.ofe.org
If you wish to strive for peace of soul then believe; 
if you wish to be a devotee of truth, then inquire.
   -- Friedrich Nietzsche

Re: [Fwd: F22 completes 11% of its Flight tests]

[Marin D. Condic]

jtarver wrote:
> I am refering to VAX Ada compilers with extended instruction sets.  Do you
> have a reading comprehension problem?
> 

Everyone here has been trying to politely address your concerns so it
would be nice if you were to reciprocate the courtesy.

We have already stated that the DEC Ada compiler for the VAX did not
implement any syntactic features outside of the language standard. The
compiler did provide some system dependent packages and pragmas as
permitted by the standard and the documentation did highlight any
"implementation defined" behaviors as required by the standard, but
there were *no* extensions to the syntax of the language.

We have asked what you mean by "extended instruction sets" because
clearly you are using the terminology in some way that is not the common
understanding. "VAX" itself means "Virtual Address eXtended" because the
VAX architecture was conceived as an "extention" to the "instruction
set" of the PDP11 computer. That clearly has nothing whatsoever to do
with Ada or portability. The instruction set for the VAX architecture
was never contracted for by the government to be some sort of "portable"
machine instruction set, nor did DEC ever mean for Data General to be
able to interpret VAX executable code on their machines or whatever
confusion may exist about this.

"Instruction Sets" in most common usage refer to machine language
instructions - the actual zeros and ones that the processor interprets.
Ada permits an implementation to define a mechanism for inserting
machine language instructions into an Ada program. This is necessary for
those of us who do embedded programming and need to actually get at the
hardware and manipulate it. It is, of necessity, machine specific, just
as similar machine code insertions in C or Fortran or whatever are going
to be. If this is what you are referring to, then of course it will be
non portable. Name me a language that permits machine code insertions
which is?

If by "extended instruction sets" you mean to imply that DEC Ada would
parse out tokens from an Ada program which were not defined in the LRM
and interpret those tokens in some DEC defined manner, then you need to
show us an example, because I know that I am unaware of any such
extensions to the language in that compiler and I have used it
extensively. DEC did not in any way extend the Ada language by, for
example, providing a "do...while" looping construct, or a "case"
statement with fallthrough semantics or any such thing.

Your claim that Ada (in particular DEC Ada) is not portable runs counter
to the experience of those of us who do it every day for a living. If
you believe this to be the case, you should post an example of what you
mean.

MDC
-- 
=============================================================
Marin David Condic   - Quadrus Corporation -   1.800.555.3393
1015-116 Atlantic Boulevard, Atlantic Beach, FL 32233
http://www.quadruscorp.com/
m c o n d i c @ q u a d r u s c o r p . c o m

Visit my web site at:  http://www.mcondic.com/

"Capitalism without failure is like religion without sin." 
        --  Allan Meltzer, Economist 
=============================================================

Re: [Fwd: F22 completes 11% of its Flight tests]

[jtarver]

Marin D. Condic <mcondic-nospam@quadruscorp.com> wrote in message
news:38835B50.B2EF88C6@quadruscorp.com...
> jtarver wrote:
> > I am refering to VAX Ada compilers with extended instruction sets.  Do
you
> > have a reading comprehension problem?
> >
>
> Everyone here has been trying to politely address your concerns so it
> would be nice if you were to reciprocate the courtesy.

The person I was responding to does not belong on this thread.

> We have already stated that the DEC Ada compiler for the VAX did not
> implement any syntactic features outside of the language standard.

And having been there at the time I know for a fact that the VAX had an
extended instruction set.  I made no comment about syntax.

John

Re: [Fwd: F22 completes 11% of its Flight tests]

[Gautier]

> And having been there at the time I know for a fact that the VAX had an
> extended instruction set.  I made no comment about syntax.

It arouses my curiosity. Just read the release notes of DEC Ada 3.5 (May 1999),
found details about options for generating machine instructions for
such or such target processor, instructions for the debugger, but - damn -
no extended instruction for the Ada language itself. Really interesting.
Is it a hidden feature ? Or has it disappeared ? Do you remember an example ?

-- 
Gautier

_____\\________________\_______\_________
http://members.xoom.com/gdemont/gsoft.htm

Re: [Fwd: F22 completes 11% of its Flight tests]

[jtarver]

Gautier <gautier.demontmollin@maths.unine.ch> wrote in message
news:38838D5A.9EB9F452@maths.unine.ch...
> > And having been there at the time I know for a fact that the VAX had an
> > extended instruction set.  I made no comment about syntax.
>
> It arouses my curiosity. Just read the release notes of DEC Ada 3.5 (May
1999),
> found details about options for generating machine instructions for
> such or such target processor, instructions for the debugger, but - damn -
> no extended instruction for the Ada language itself. Really interesting.
> Is it a hidden feature ?

The Ada language itself does not allow extended instruction sets, as I
plainly pointed out at the onset of this discussion.

> Or has it disappeared ? Do you remember an example ?

Another poster has already answered your question.  I suggest you go to
Deja-News and read the thread.  What the poster implied to me is that the
extensions are part of a tool set now.

John

Re: [Fwd: F22 completes 11% of its Flight tests]

[Bill Greene]

jtarver wrote:
> 
> The Ada language itself does not allow extended instruction sets, as I
> plainly pointed out at the onset of this discussion.

The reason I and several other posters wore out your patience asking
what you meant by this statement was so that we could determine whether
your statement was false or was just badly stated.

If you are referring to the instruction sets of computers, your
statement is false.  The Ada standard says nothing of the sort.

If you are referring to extensions to the Ada programming language, they
were indeed prohibited.  This prohibition has nothing to do with the
extended instruction set of a particular computer such as the VAX.
-- 
William R. Greene                              1100 Perimeter Park Drive
Ganymede Software, Inc.                                        Suite 104
http://www.ganymede.com                       Morrisville, NC  27560 USA
Phone: (919) 469-0997, ext. 280                      Fax: (919) 469-5553

Re: [Fwd: F22 completes 11% of its Flight tests]

[Robert S. White]

In article <nhMg4.3374$iy5.215390@typ12.deja.bcandid.com>, jtarver@tminet.com says...
>
>...  What the poster implied to me is that the
>extensions are part of a tool set now.

  And back in 1985 when DEC Ada came out the "extensions" (aka
Starlet package) were part of a tool set then.  There never were
any "DEC Ada extensions".  Why don't you just admit that you
are/were wrong?

-- 
Robert S. White         -- An embedded systems software engineer
e-mail reply to reverse of ( add .'s ): com home shift2 rswhite
            or do a "Reply To All" for direct eMailed cc'd followups.

Re: [Fwd: F22 completes 11% of its Flight tests]

[Ted Dennison]

In article <hx9h4.304$IF5.13456@news.rdc1.il.home.com>,
  RSWhite@nospam.somewhere.ia.us (Robert S. White) wrote:
> In article <nhMg4.3374$iy5.215390@typ12.deja.bcandid.com>,
jtarver@tminet.com says...
> >
> >...  What the poster implied to me is that the
> >extensions are part of a tool set now.
>
>   And back in 1985 when DEC Ada came out the "extensions" (aka
> Starlet package) were part of a tool set then.  There never were
> any "DEC Ada extensions".  Why don't you just admit that you
> are/were wrong?

That's silly. You can't call Starlet "extensions to Ada" any more than
you can call the Motif library "extensions to C". Its just a (big
honking) package someone wrote, in Ada.

--
T.E.D.

http://www.telepath.com/~dennison/Ted/TED.html


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: [Fwd: F22 completes 11% of its Flight tests]

[Robert S. White]

In article <864hrd$546$1@nnrp1.deja.com>, dennison@telepath.com says...
>
>In article <hx9h4.304$IF5.13456@news.rdc1.il.home.com>,
>  RSWhite@nospam.somewhere.ia.us (Robert S. White) wrote:
>> In article <nhMg4.3374$iy5.215390@typ12.deja.bcandid.com>,
>jtarver@tminet.com says...
>> >
>> >...  What the poster implied to me is that the
>> >extensions are part of a tool set now.
>>
>>   And back in 1985 when DEC Ada came out the "extensions" (aka
>> Starlet package) were part of a tool set then.  There never were
>> any "DEC Ada extensions".  Why don't you just admit that you
>> are/were wrong?
>
>That's silly. You can't call Starlet "extensions to Ada" any more than
>you can call the Motif library "extensions to C". Its just a (big
>honking) package someone wrote, in Ada.

  Exactly!

-- 
Robert S. White         -- An embedded systems software engineer
e-mail reply to reverse of ( add .'s ): com home shift2 rswhite
            or do a "Reply To All" for direct eMailed cc'd followups.

Re: [Fwd: F22 completes 11% of its Flight tests]

[Stefan Skoglund]

Robert S. White wrote:
> >That's silly. You can't call Starlet "extensions to Ada" any more than
> >you can call the Motif library "extensions to C". Its just a (big
> >honking) package someone wrote, in Ada.
> 
>   Exactly!

Can we agree on one thing ?
Every time someone uses a package provided by the vendor
they do so of their own volution ? ie if i need a LALR parser in
my program i probably picks up aflex ?

If someone is doing Ada with a VAX as target they could
very well use some package provided by Dec in their Ada offerings.

Every time someone decides that it is worthwhile using some package
from someone else and running the risk that that package could be
unavailable some day ie probably causing a re-write.

Like in the case there you go from VMS to UNIX to NT (maybe 2 rewrites
and maybe not.)

This is one of the differences between Freeware projects like
GNAT and vendor-specific products ie the customer is usually out on
a limb if the vendor decides to drop some product.

Re: [Fwd: F22 completes 11% of its Flight tests]

[Richard D Riehle]

In article <388E014A.B8F87F5@ebox.tninet.se>,
	Stefan Skoglund <stetson@ebox.tninet.se> wrote:

>Can we agree on one thing ?
>Every time someone uses a package provided by the vendor
>they do so of their own volution ? ie if i need a LALR parser in
>my program i probably picks up aflex ?

I would agree that _withing_ a vendor supplied package at the
specification level often reflects poor planning.  Ordinarily,
one should define the specification required for the problem
space then _with_ the vendor's package in the package body 
(or even in a separately compiled subpgram) to effect the actual
implementation.  This accomplishes several things: 

     1) The interface is independent of the implementation
     2) A different implementation can be used to effect the
        interface,
     3) Vendor dependencies are hidden from users,
     4) With Ada 95, we can take advantage of private child units,
     5) We are more platform-independent.

These are just a few reasons for hiding vendor supplied packages
when possible.  Granted, it is not always possible or even appropriate.

There is even a valid school-of-thought that suggests no
relying directly on standard Ada library packages at the specification
level.  How often does one need all of the resources of Text_IO, or
Direct_IO. Instead, a package such as Direct_IO can be _withed_ in 
the package body to provide primitives necessary for exporting more
powerful abstractions.  Examples,

         procedure Open_With_Lock ( ... );
         procedure Read_With_Lock ( ... );
         procedure Unlock_File   ( ... );

and other routines, all of which can be implemented through the
services of Direct_IO, for ensuring file security in a task based
product.

So all the fuss about _extensions_ is just that, "fuss."  Experienced
Ada software designers have been avoiding such dependencies for a 
long time. 

Richard Riehle
richard@adaworks.com
  

Re: [Fwd: F22 completes 11% of its Flight tests]

[Mark Lundquist]

Richard D Riehle wrote:

> I would agree that _withing_ a vendor supplied package at the
> specification level often reflects poor planning.  Ordinarily,
> one should define the specification required for the problem
> space then _with_ the vendor's package in the package body
> (or even in a separately compiled subpgram) to effect the actual
> implementation.  This accomplishes several things:
>
>      1) The interface is independent of the implementation
>      2) A different implementation can be used to effect the
>         interface,
>      3) Vendor dependencies are hidden from users,
>      4) With Ada 95, we can take advantage of private child units,
>      5) We are more platform-independent.
>
> These are just a few reasons for hiding vendor supplied packages
> when possible.  Granted, it is not always possible or even appropriate.
>
> [...]
>
> So all the fuss about _extensions_ is just that, "fuss."  Experienced
> Ada software designers have been avoiding such dependencies for a
> long time.

Richard, those are good points about using packages to minimize coupling.
But it's quite beside the point of the discussion!  You're talking about
the scope of dependencies, and how to reduce that scope through the use of
proper programming techniques.  The discussion (the "fuss"?) wasn't about
that, it was about the risk to a project of using components supplied by
someone else (e.g. a compiler vendor or a third party).  I.e., about using
them, not hiding vs. exposing them.

(I'm talking now about the last 2-3 messages on this sub-thread... before
that, there were a huge number of messages trying to make sense of the
babblings of some kind of deranged person.  The most fruitless exchange of
messages I've ever seen.  They all pretty much looked like:

            Deranged Person:    @#&$~!   grlb,  flgmrzz!
            Ada Person:            Excuse me?

That was where the "extension" thing came from, but nobody could figure
out what the guy meant, and he finally went away).

Re: [Fwd: F22 completes 11% of its Flight tests]

[Samuel T. Harris]

Mark Lundquist wrote:
> 
> Richard D Riehle wrote:
> 
> > I would agree that _withing_ a vendor supplied package at the
> > specification level often reflects poor planning.  Ordinarily,
> > one should define the specification required for the problem
> > space then _with_ the vendor's package in the package body
> > (or even in a separately compiled subpgram) to effect the actual
> > implementation.  This accomplishes several things:
> >
> >      1) The interface is independent of the implementation
> >      2) A different implementation can be used to effect the
> >         interface,
> >      3) Vendor dependencies are hidden from users,
> >      4) With Ada 95, we can take advantage of private child units,
> >      5) We are more platform-independent.
> >
> > These are just a few reasons for hiding vendor supplied packages
> > when possible.  Granted, it is not always possible or even appropriate.

Yes, less filling AND tastes great. Done this many time before.

> >
> > [...]
> >
> > So all the fuss about _extensions_ is just that, "fuss."  Experienced
> > Ada software designers have been avoiding such dependencies for a
> > long time.
> 
> Richard, those are good points about using packages to minimize coupling.
> But it's quite beside the point of the discussion!  You're talking about
> the scope of dependencies, and how to reduce that scope through the use of
> proper programming techniques.  The discussion (the "fuss"?) wasn't about
> that, it was about the risk to a project of using components supplied by
> someone else (e.g. a compiler vendor or a third party).  I.e., about using
> them, not hiding vs. exposing them.

I believe Richard's points are to the point. Risk is about
cost and benefit. I benefit from using a vendor supplied package
but I incur cost in the form of portability. Sometimes this cost
is a limitation when the vendor supplied package is something really
special doing things not readily available on other platforms.
Most times the cost is simply extra effort when porting since most
stuff is solved everywhere by someone or another. I use the example
of socket libraries. In practical terms, every platform I ever used
has/had them but the bindings differ. So I have extra cost using
a vendor package. Following Richard's advise, I minimize my cost.

Risks are not examined in a void. The cost and payoff of taking
the risky road is compared with the cost of taking the safe road
(which usually has not payoff which is why one would be considering
the risky road in the first place).

Now I can decide based on good cost/benefit analysis. Do I grow my
own solution with significant development cost concerning things
which are not directly related to the problem at hand, or do I
take advantage of someone else's work and use a vendor supplied
package?

As a parting note, please consider that todays vendor specific
package may be tomorrows standard. 
Take the Alsys generic_elementary_functions,
and Rational's ASIS stuff. Also consider that what many customers
rely upon become de-facto standards and competitors make their
own "knock-offs." For this I present the great efforts ACT expends
to provide all those VADS and VAX specific idiosynchracies.
ACT knows that to get folks to migrate, ACT has to pave the way
with these value-added packages.

> 
> (I'm talking now about the last 2-3 messages on this sub-thread... before
> that, there were a huge number of messages trying to make sense of the
> babblings of some kind of deranged person.  The most fruitless exchange of
> messages I've ever seen.  They all pretty much looked like:
> 
>             Deranged Person:    @#&$~!   grlb,  flgmrzz!
>             Ada Person:            Excuse me?
> 
> That was where the "extension" thing came from, but nobody could figure
> out what the guy meant, and he finally went away).

-- 
Samuel T. Harris, Principal Engineer
Raytheon, Aerospace Engineering Services
"If you can make it, We can fake it!"

Re: [Fwd: F22 completes 11% of its Flight tests]

[Keith Thompson]

"jtarver" <jtarver@tminet.com> writes:
> Marin D. Condic <mcondic-nospam@quadruscorp.com> wrote in message
> news:38835B50.B2EF88C6@quadruscorp.com...
> > jtarver wrote:
> > > I am refering to VAX Ada compilers with extended instruction sets.
> > > Do you have a reading comprehension problem?
> >
> > Everyone here has been trying to politely address your concerns so it
> > would be nice if you were to reciprocate the courtesy.
> 
> The person I was responding to does not belong on this thread.

I believe this refers to me.  This looks like a gratuitous personal
insult, but as you've pointed out I sometimes have problems
understanding what you write.  Please clarify.

> > We have already stated that the DEC Ada compiler for the VAX did not
> > implement any syntactic features outside of the language standard.
> 
> And having been there at the time I know for a fact that the VAX had an
> extended instruction set.  I made no comment about syntax.

If you're talking about the VAX instruction set, you're not saying
anything particularly relevant to Ada.  You originally mentioned
this in the context of the Ada standard, implying that the VAX Ada
compiler(s) violated the Ada standard by implementing language
extensions.  Please clarify.  It might be helpful if you would
re-phrase your comments rather than repeating them.  Concrete examples
would be especially helpful.

I've nearly lost hope that this thread will lead to anything useful,
but I'm prepared to be pleasantly surprised.

-- 
Keith Thompson (The_Other_Keith) kst@cts.com  <http://www.ghoti.net/~kst>
San Diego Supercomputer Center           <*>  <http://www.sdsc.edu/~kst>
Welcome to the last year of the 20th century.

Re: [Fwd: F22 completes 11% of its Flight tests]

[K. Devlin]

>On 16 Jan 2000 00:57:05 -0800, Keith Thompson <kst@cts.com> wrote:

>"jtarver" <jtarver@tminet.com> writes:
>> Nothing like having the rules change to deliver that Boston pork.  DEC had
>> extensions when their computer became the defacto standard for Ada.  At the
>> time any extensions to Ada were forbidden.  Data General was rightiously
>> upset when this was done.  I was working SDI in the 80s when Ada became the
>> perfered controls language.
>
>Please be specific.  What extensions are you referring to?  (BTW, I
>was working on Ada compilers in the 80s.)
>
>> Ada suxs.
>
>Yeah, whatever.
>
>-- 
>Keith Thompson (The_Other_Keith) kst@cts.com  <http://www.ghoti.net/~kst>
>San Diego Supercomputer Center           <*>  <http://www.sdsc.edu/~kst>
>Welcome to the last year of the 20th century.

I would suggest that attempting to reason with jtarver is a waste of
time.  He clearly hates Ada, appears to hate Lockheed Martin and
NOTHING you can "say" nor can any documented proof, change his
opinions.

I hope he doesn't fly on Boeing 777 A/C as they are as close to 100%
Ada based as possible (see http://adaic.org/docs/flyers/boe-777.shtml)

He doesn't seem to, or want to, understand that (particularly when DoD
"owned" Ada) there was (and still is) only one validation suite. A
compiler either met this and received a validation certificate or it
didn't.  If DEC Ada became the "defacto standard" it was because they
had an excellent product and their computers were widely used in the
area where Ada was mandated.  


Of course there were/are differences between compilers. Among others,
I've used DEC Ada and also a version of Telesoft Ada ported to the
VAX.  In the Telesoft version one had Integer (32 bits) and
Short_Integer (16 bits). DEC Ada had Integer (16 bits) and
Long_Integer (32 Bits).  Both compilers were validated and these types
were clearly defined in the appropriate place of their respect RM.
This particular example wasn't a big deal (nor was it an "extension)
as a simple global replace of integer for long_integer and then
short_integer to integer or vica-versa was sufficient and of course
either compiler would point out the unknown type name. 

My original cut at the program was in DEC Ada as much due to the
excellent error messages as anything else, and then I moved it to
Telesoft which is where the one issue (the integer types) surfaced.
Granted both compilers targeted the VAX but each was validated and the
differences were minor, and those differences were properly
documented.

IIRC, to do anything other than simple text_io required using package
STARLET (which might be jtarver's extension's).  This package was/is
"legal" and was clearly documented in the appropriate section of the
RM in accordance with the standard. 

 

kbd

Re: [Fwd: F22 completes 11% of its Flight tests]

[Chris Douglas]

K. Devlin wrote:

> ...
> I would suggest that attempting to reason with jtarver is a waste of
> time.  He clearly hates Ada, appears to hate Lockheed Martin and
> NOTHING you can "say" nor can any documented proof, change his
> opinions.

I think you've hit the nail on the head there.  Right or wrong, he's a "true
believer" that the F-22 and anything related to it is bad.  Examining fact doesn't
play much of a role in the equation--there's a bit too much emotion at play for
that.  Don't ask me why.

Looking forward to the inevitable reprisal,
Chris Douglas

Re: [Fwd: F22 completes 11% of its Flight tests]

[P. S. Norby]

Chris Douglas wrote:
> 
> K. Devlin wrote:
> 
> > ...
> > I would suggest that attempting to reason with jtarver is a waste of
> > time.  He clearly hates Ada, appears to hate Lockheed Martin and
> > NOTHING you can "say" nor can any documented proof, change his
> > opinions.
> 
> I think you've hit the nail on the head there.  Right or wrong, he's a "true
> believer" that the F-22 and anything related to it is bad.  Examining fact doesn't
> play much of a role in the equation--there's a bit too much emotion at play for
> that.  Don't ask me why.
> 
> Looking forward to the inevitable reprisal,
> Chris Douglas

Well, as someone said:

Never try to teach a pig to sing.  It wastes your time, and it
annoys the pig.
 
P. S. Norby

 "Software engineers are, in many ways, similar to normal people"

        --  Scott Adams

"No excuses.  No embarrassment.  No apologies...
 Ada -- the most trusted and powerful programming language
 on earth, or in space." -- S. Tucker Taft
 
\\\    \\\    \\\    \\\    \\\    \\\    \\\    \\\    \\\ 
( :)   ( :)   ( :)   ( :)   ( :)   ( :)   ( :)   ( :)   ( :)
///    ///    ///    ///    ///    ///    ///    ///    /// 
(Speaking only for myself)

Re: [Fwd: F22 completes 11% of its Flight tests]

[David Starner]

On Sun, 16 Jan 2000 19:01:53 -0800, jtarver <jtarver@tminet.com> wrote:
>K. Devlin <kdevlin@mpinet.net> wrote in message
>news:388274ac.42926465@news.mpinet.net...

>> I hope he doesn't fly on Boeing 777 A/C as they are as close to 100%
>> Ada based as possible (see http://adaic.org/docs/flyers/boe-777.shtml)
>
>In fact Boeining studied using Ada and then rejected it along with the ARPA
>funding using the language would have provided.  In order to fly on a civil
>airplane running under Ada you would have to fly on a 717.  Of course I only
>built 777s; Bees love C.

Comparing the trustworthness of someone who uses the word "sux", ad homidea
attacks and doesn't use his real name on posts, to a well documented,
highly detailed website . . . I think I know who I will believe.

-- 
David Starner - dstarner98@aasaa.ofe.org
If you wish to strive for peace of soul then believe; 
if you wish to be a devotee of truth, then inquire.
   -- Friedrich Nietzsche

Re: [Fwd: F22 completes 11% of its Flight tests]

[Mark Lundquist]

David Starner wrote:

> On Sun, 16 Jan 2000 19:01:53 -0800, jtarver <jtarver@tminet.com> wrote:
> >K. Devlin <kdevlin@mpinet.net> wrote in message
> >news:388274ac.42926465@news.mpinet.net...
>
> >> I hope he doesn't fly on Boeing 777 A/C as they are as close to 100%
> >> Ada based as possible (see http://adaic.org/docs/flyers/boe-777.shtml)
> >
> >In fact Boeining studied using Ada and then rejected it along with the ARPA
> >funding using the language would have provided.  In order to fly on a civil
> >airplane running under Ada you would have to fly on a 717.  Of course I only
> >built 777s; Bees love C.
>
> Comparing the trustworthness of someone who uses the word "sux",

Hey, come on... I use that word all the time!  :-)

Of course, it's much more effective when there is some content to back it up...

Re: [Fwd: F22 completes 11% of its Flight tests]

[Stuart Palin]

jtarver wrote:
> K. Devlin <kdevlin@mpinet.net> wrote in message
> news:388274ac.42926465@news.mpinet.net...

<snip>

> > I hope he doesn't fly on Boeing 777 A/C as they are as close to 100%
> > Ada based as possible (see http://adaic.org/docs/flyers/boe-777.shtml)
> 
> In fact Boeining studied using Ada and then rejected it along with the ARPA
> funding using the language would have provided.  In order to fly on a civil
> airplane running under Ada you would have to fly on a 717.  Of course I only
> built 777s; Bees love C.

The 777 Flight Control Computer software is written Ada.

It is compiled to run on three different processors using three
differently compilers.

Regards
--
Stuart Palin

Re: [Fwd: F22 completes 11% of its Flight tests]

[jtarver]

Stuart Palin <stuart.palin@no.spam.gecm.com> wrote in message
news:3882DC89.AB2831A9@no.spam.gecm.com...
> jtarver wrote:
> > K. Devlin <kdevlin@mpinet.net> wrote in message
> > news:388274ac.42926465@news.mpinet.net...
>
> <snip>
>
> > > I hope he doesn't fly on Boeing 777 A/C as they are as close to 100%
> > > Ada based as possible (see http://adaic.org/docs/flyers/boe-777.shtml)
> >
> > In fact Boeining studied using Ada and then rejected it along with the
ARPA
> > funding using the language would have provided.  In order to fly on a
civil
> > airplane running under Ada you would have to fly on a 717.  Of course I
only
> > built 777s; Bees love C.
>
> The 777 Flight Control Computer software is written Ada.

My lead when I was at Boeing developed the FCC system for the 777 and claims
they chose to not use the Ada.  Do you know Dave Mc Laughlin?

John

Re: [Fwd: F22 completes 11% of its Flight tests]

[David Tannen]

John,

>> The 777 Flight Control Computer software is written Ada.
>
>My lead when I was at Boeing developed the FCC system for the
>777 and claims they chose to not use the Ada.  Do you know
>Dave Mc Laughlin?


I just finished working with about 8 people who worked at Honeywell
in Phoenix on the 777 flight software.  They all used Ada.  I have
another friend who has been at Honeywell for a long time (about
8 or 9 years) who has worked on various Boeing projects - all
in Ada.

I think it is getting harder and harder to believe that you actually
know anything about the Ada market.

An ambassador for Christ
David Tannen (tannen@jcdisciples.org)
John 14:21 "Whoever has my commands and obeys them, he is the one who
            loves me.  He who loves me will be loved by my Father,
            and I too will love him and show myself to him."
Check Out:   http://www.jcdisciples.org/davidtannen/index.html

Re: [Fwd: F22 completes 11% of its Flight tests]

[K. Devlin]

On Mon, 17 Jan 2000 09:48:13 -0800, "jtarver" <jtarver@tminet.com>
wrote:

>
>Stuart Palin <stuart.palin@no.spam.gecm.com> wrote in message
>news:3882DC89.AB2831A9@no.spam.gecm.com...
>> jtarver wrote:
>> > K. Devlin <kdevlin@mpinet.net> wrote in message
>> > news:388274ac.42926465@news.mpinet.net...
>>
>> <snip>
>>
>> > > I hope he doesn't fly on Boeing 777 A/C as they are as close to 100%
>> > > Ada based as possible (see http://adaic.org/docs/flyers/boe-777.shtml)
>> >
>> > In fact Boeining studied using Ada and then rejected it along with the
>ARPA
>> > funding using the language would have provided.  In order to fly on a
>civil
>> > airplane running under Ada you would have to fly on a 717.  Of course I
>only
>> > built 777s; Bees love C.
>>
>> The 777 Flight Control Computer software is written Ada.
>
>My lead when I was at Boeing developed the FCC system for the 777 and claims
>they chose to not use the Ada.  Do you know Dave Mc Laughlin?
>
>John
>
>
>
Claims????   

That doesnt seem to be a very definitive statement. If you were there
why hedge?

How 'bout proof??  Any documentation you can point us to??

kbd

Re: [Fwd: F22 completes 11% of its Flight tests]

[David Gillon]

jtarver wrote:

> > I hope he doesn't fly on Boeing 777 A/C as they are as close to 100%
> > Ada based as possible (see http://adaic.org/docs/flyers/boe-777.shtml)
> 
> In fact Boeining studied using Ada and then rejected it along with the ARPA
> funding using the language would have provided.  In order to fly on a civil
> airplane running under Ada you would have to fly on a 717.  Of course I only
> built 777s

I spent four years or so coding the 777 flight control system in Ada
(and as it happens was project point-of-contact for the DEC XD-Ada
compiler, which was probably the most trouble-free of the three
compilers the code was targetted against).

-- 

David Gillon

Re: [Fwd: F22 completes 11% of its Flight tests]

[K. Devlin]

On Sun, 16 Jan 2000 19:01:53 -0800, "jtarver" <jtarver@tminet.com> 
>> I would suggest that attempting to reason with jtarver is a waste of
>> time.  He clearly hates Ada, appears to hate Lockheed Martin and
>> NOTHING you can "say" nor can any documented proof, change his
>> opinions.
>
>You are so emotional Keith.
>

My first name is not Keith,... just one of many errors on your part.


>> I hope he doesn't fly on Boeing 777 A/C as they are as close to 100%
>> Ada based as possible (see http://adaic.org/docs/flyers/boe-777.shtml)
>
>In fact Boeining studied using Ada and then rejected it along with the ARPA
>funding using the language would have provided.  In order to fly on a civil
>airplane running under Ada you would have to fly on a 717.  Of course I only
>built 777s; Bees love C.
>
>John
>
>

Then I suggest you contact Boeing (and or the AdaIC) and inform them
of the false (according to you) assertion made by the AdaIC (follow
url I provided).

kbd

Re: [Fwd: F22 completes 11% of its Flight tests]

[Ed Falis]

In article <3883b041.3549326@news.mpinet.net>,
  kdevlin@mpinet.net (K. Devlin) wrote:
> On Sun, 16 Jan 2000 19:01:53 -0800, "jtarver" <jtarver@tminet.com>
> >In fact Boeining studied using Ada and then rejected it along with the ARPA
> >funding using the language would have provided.  In order to fly on a civil
> >airplane running under Ada you would have to fly on a 717.  Of course I only
> >built 777s; Bees love C.


> Then I suggest you contact Boeing (and or the AdaIC) and inform them
> of the false (according to you) assertion made by the AdaIC (follow
> url I provided).
>
> kbd

Yeah, I kind of wonder why my company got revenue for providing Ada compilers
and certification packages for fire control, navigation and braking systems
from several 777 subcontracters if this is true.  Maybe they just had a lot
of money to blow to market Ada. Man, I just love to see that kind of altruism
in American business.  Course, I could've been hallucinating the commissions
I got.

- Ed


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: [Fwd: F22 completes 11% of its Flight tests]

[Ted Dennison]

In article <860n70$cn4$1@nnrp1.deja.com>,
  Ed Falis <falis@ma.aonix.com> wrote:

> of money to blow to market Ada. Man, I just love to see that kind of
> altruism in American business.  Course, I could've been hallucinating
> the commissions I got.

Damn. How come *I* never have a hallucination that involves commisions?
That sounds like a blast! :-)

Mine generally involve my code having no bugs...

--
T.E.D.

http://www.telepath.com/~dennison/Ted/TED.html


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: [Fwd: F22 completes 11% of its Flight tests]

[Roga Danar]

Hi,

  I worked with a contractor fellow who actually carried around to project to
project the Ada source code he developed for the 777.

 Perhaps to convince people like jtarver ;->

jtarver wrote:

> Stuart Palin <stuart.palin@no.spam.gecm.com> wrote in message
> news:3882DC89.AB2831A9@no.spam.gecm.com...
> > jtarver wrote:
> > > K. Devlin <kdevlin@mpinet.net> wrote in message
> > > news:388274ac.42926465@news.mpinet.net...
> >
> > <snip>
> >
> > > > I hope he doesn't fly on Boeing 777 A/C as they are as close to 100%
> > > > Ada based as possible (see http://adaic.org/docs/flyers/boe-777.shtml)
> > >
> > > In fact Boeining studied using Ada and then rejected it along with the
> ARPA
> > > funding using the language would have provided.  In order to fly on a
> civil
> > > airplane running under Ada you would have to fly on a 717.  Of course I
> only
> > > built 777s; Bees love C.
> >
> > The 777 Flight Control Computer software is written Ada.
>
> My lead when I was at Boeing developed the FCC system for the 777 and claims
> they chose to not use the Ada.  Do you know Dave Mc Laughlin?
>
> John

Re: [Fwd: F22 completes 11% of its Flight tests]

[K. Devlin]

On Sat, 15 Jan 2000 02:04:58 -0000, "Keith Willshaw"
<keith_willshaw@compuserve.com> wrote:

>
>Bill Greene wrote in message <387FCA73.3A61@Ganymede.com>...
>>Ed Falis wrote:
>>
>>Perhaps he is really complaining about the instruction set of a CPU or
>>processor architecture (the VAX), and not about a programming language
>>(Ada) at all.  Note his assertion that "VAX" violated some alleged
>>design goal of Ada.  Having tried unsuccessfully to discover his meaning
>>by direct questions, all I can conclude is that he is blaming the Ada
>>programming language (which he apparently thought was designed by DEC)
>>for the fact that the VAX is non-compliant with the Data General
>>instruction set!
>>
>
>
>More liley its DEC's own compiler he's complaing
>about. Their Fortran compilers had lots of extensions
>that were far from ANSI standard Fortran.
>
>My company had to move a lot of this stuff from
>VMS to Unix and then NT
>
>The DEC extensions were a major headache
>
>Keith
>
>
Have to agree, DEC FORTAN has lots of non-standard extensions, in the
DEC FORTRAN manual I have these extensions are printed in blue to make
them easy to identify. If you stayed away from them, then the program
"should" be portable to any other "standard" FORTRAN.

Ada is much better in this regard. One of the nice features of Ada it
makes it easier (and as far as the compiler itself, requires it) to
compartmentalize the system specific "stuff". If done right (and I
realize thats a big IF) changing processors should be easier using Ada
than any other language. 

As far as processors themselves:

An option I havn't seen mentioned is the life-time buy option.  It has
been done in the past and could have been done for the F-22's i960
(but I don't know if it has) where the quantity of processors (or any
other part) determined to be needed to build some number of systems
and maintain same for a specified period of time, are bought before
the vendor stops production and are put into storage.

However I see that money is being spent to develop an upgrade for the
F18E/F (Military&AeroSpace Electronics Jan 2000 Vol 11 No 1) with a
PowerPC based board from DY 4. While the DY 4 board doesn't currently
appear to be one of the targets for an Ada compiler, rehosting isn't
THAT hard if you start with a compiler that meets the standard for the
processor.  (I count 7 different vendors of PowerPC based compilers
http://www.adaic.org/cgi-bin/vcl/report95.pl) 

For additional information on Ada:

The following URL is an online copy of the Ada reference Manual:

http://www.adapower.com/rm95/index.html

Section 1.1.3 Conformity of an Implementation with the Standard
discusses what's required to conform to the standard
(http://www.adapower.com/rm95/arm95_16.html#SEC16).

kbd

Re: [Fwd: F22 completes 11% of its Flight tests]

[Marin D. Condic]

Bill Greene wrote:
> Perhaps he is really complaining about the instruction set of a CPU or
> processor architecture (the VAX), and not about a programming language
> (Ada) at all.  Note his assertion that "VAX" violated some alleged
> design goal of Ada.  Having tried unsuccessfully to discover his meaning
> by direct questions, all I can conclude is that he is blaming the Ada
> programming language (which he apparently thought was designed by DEC)
> for the fact that the VAX is non-compliant with the Data General
> instruction set!
> 
My experience is that those who publicly abuse the Ada programming
language often do so out of ignorance or totally incorrect information
about the subject. I'm convinced that the negative views of Ada seen in
this thread must be for one of these reasons. We can hope that those
with some sort of axe to grind with Ada might be willing to take a look
at the complaints they have in light of whatever information we can
provide and perhaps come to see that the complaints are misguided.

In the mean time, keep up the good work of patiently explaining what Ada
is all about. I'm sure it will help others to see its usefulness and
perhaps we'll draw a few more users into the fold. :-)

MDC
-- 
=============================================================
Marin David Condic   - Quadrus Corporation -   1.800.555.3393
1015-116 Atlantic Boulevard, Atlantic Beach, FL 32233
http://www.quadruscorp.com/
m c o n d i c @ q u a d r u s c o r p . c o m

Visit my web site at:  http://www.mcondic.com/

"Capitalism without failure is like religion without sin." 
        --  Allan Meltzer, Economist 
=============================================================

Re: [Fwd: F22 completes 11% of its Flight tests]

[Marin D. Condic]

Ed Falis wrote:
> I can't think of any Ada compiler that went that far, though.  All
> "extensions" I'm aware of have either been domain-oriented packages
> (encouraged by the standard), or implementation-defined pragmas
> (allowed by the language).  DEC Ada featured mostly the former in its
> Starlet package that enabled access to OS facilities, much as the
> POSIX/Ada binding enabled access to POSIX compliant OS's.
> 
I don't think I've seen anything in recent years that actually provided
syntactic extensions to the language. Virtually all "extensions" have
been in the form of either pragmas or packages. However, I seem to
recall an early version of the RR compiler that provided some syntactic
extensions in the area of I/O because generic packages were not yet
implemented. If memory serves, early versions of the Telesoft Ada
compiler did something similar. (And I'm still holding a grudge against
that compiler for a whole variety of reasons. I'll bet I'm not the only
one! ;-)

But the important thing to note is that neither of those compilers had
been validated and they were very careful not to actually call
themselves "Ada" directly to keep from violating the trademark.

MDC
-- 
=============================================================
Marin David Condic   - Quadrus Corporation -   1.800.555.3393
1015-116 Atlantic Boulevard, Atlantic Beach, FL 32233
http://www.quadruscorp.com/
m c o n d i c @ q u a d r u s c o r p . c o m

Visit my web site at:  http://www.mcondic.com/

"Capitalism without failure is like religion without sin." 
        --  Allan Meltzer, Economist 
=============================================================

Re: [Fwd: F22 completes 11% of its Flight tests]

[David Starner]

On Sat, 15 Jan 2000 11:22:20 -0800, Marin D. Condic <mcondic-nospam@quadruscorp.com> wrote:
>I don't think I've seen anything in recent years that actually provided
>syntactic extensions to the language. 

GNAT has "with type" to allow two types to reference each other without being
in the same package. That's only enabled with -gnatX so it's not a big deal.
-- 
David Starner - dstarner98@aasaa.ofe.org
If you wish to strive for peace of soul then believe; 
if you wish to be a devotee of truth, then inquire.
   -- Friedrich Nietzsche

Re: [Fwd: F22 completes 11% of its Flight tests]

[Marin D. Condic]

David Starner wrote:
> GNAT has "with type" to allow two types to reference each other without being
> in the same package. That's only enabled with -gnatX so it's not a big deal.
> --
Thus demonstrating that I am not a knower-of-all once again! ;-)

As you point out, you have to throw a switch to make it happen. The
compiler still has a mode in which it will accept only standard
constructs, so if portability is your concern, you just always use that
mode. I think that's one of the best benefits of a validated compiler
with respect to portability. Go ahead and add features if you like, but
I can always verify that my code adheres to only standard features if
that is important to me.

MDC
-- 
=============================================================
Marin David Condic   - Quadrus Corporation -   1.800.555.3393
1015-116 Atlantic Boulevard, Atlantic Beach, FL 32233
http://www.quadruscorp.com/
m c o n d i c @ q u a d r u s c o r p . c o m

Visit my web site at:  http://www.mcondic.com/

"Capitalism without failure is like religion without sin." 
        --  Allan Meltzer, Economist 
=============================================================

Re: [Fwd: F22 completes 11% of its Flight tests]

[Keith Willshaw]

Marin D. Condic wrote in message <38816E17.8B24C111@quadruscorp.com>...
>David Starner wrote:
>> GNAT has "with type" to allow two types to reference each other without
being
>> in the same package. That's only enabled with -gnatX so it's not a big
deal.
>> --
>Thus demonstrating that I am not a knower-of-all once again! ;-)
>
>As you point out, you have to throw a switch to make it happen. The
>compiler still has a mode in which it will accept only standard
>constructs, so if portability is your concern, you just always use that
>mode. I think that's one of the best benefits of a validated compiler
>with respect to portability. Go ahead and add features if you like, but
>I can always verify that my code adheres to only standard features if
>that is important to me.
>
>MDC
>--


The trouble is too many people never think about it.

The people that wrote the DEC Fortran that I got stuck
with told me they couldn't imagine why they would ever
move off VAX, Theywere the most cost effective machines to
do it on.

Most of our customers are now on $1500 PC's and we
are back to Standard Fortran

Keith

Re: [Fwd: F22 completes 11% of its Flight tests]

[Paul J. Adam]

In article <85sbv7$6ea$1@ssauraab-i-1.production.compuserve.com>,
Keith Willshaw <keith_willshaw@compuserve.com> writes
>The people that wrote the DEC Fortran that I got stuck
>with told me they couldn't imagine why they would ever
>move off VAX, Theywere the most cost effective machines to
>do it on.
>
>Most of our customers are now on $1500 PC's and we
>are back to Standard Fortran

I remember writing Fortran 77 code on a PDP 11/70... (took two
hours to compile and link 25K of code) 

I'm getting old and I'm not even thirty yet :)

-- 
There are four kinds of homicide: felonious, excusable, justifiable and
praiseworthy...             

Paul J. Adam                                  paul@jrwlynch.demon.co.uk   

Re: [Fwd: F22 completes 11% of its Flight tests]

[John Keeney]

Paul J. Adam <Paul@jrwlynch.demon.co.uk> wrote in message
news:QDSgmSAAWbg4Ewqy@jrwlynch.demon.co.uk...
> In article <85sbv7$6ea$1@ssauraab-i-1.production.compuserve.com>,
> Keith Willshaw <keith_willshaw@compuserve.com> writes
> >The people that wrote the DEC Fortran that I got stuck
> >with told me they couldn't imagine why they would ever
> >move off VAX, Theywere the most cost effective machines to
> >do it on.
> >
> >Most of our customers are now on $1500 PC's and we
> >are back to Standard Fortran
>
> I remember writing Fortran 77 code on a PDP 11/70... (took two
> hours to compile and link 25K of code)
>
> I'm getting old and I'm not even thirty yet :)

You're not even thirty and you were working with 25K programs on an 11/70?
Doing what in God's name?

The last 11/70 I used was VERY early '80s and ten years later I knew people
who they were being given to in exchange for hauling them off. Turned down
one myself about '94.

> --
> There are four kinds of homicide: felonious, excusable, justifiable and
> praiseworthy...
>
> Paul J. Adam                                  paul@jrwlynch.demon.co.uk

Re: [Fwd: F22 completes 11% of its Flight tests]

[Paul J. Adam]

In article <38827ab1@news.iglou.com>, John Keeney
<jdkeeney@iglou.com> writes
>Paul J. Adam <Paul@jrwlynch.demon.co.uk> wrote in message
>> I remember writing Fortran 77 code on a PDP 11/70... (took two
>> hours to compile and link 25K of code)
>>
>> I'm getting old and I'm not even thirty yet :)
>
>You're not even thirty and you were working with 25K programs on an 11/70?
>Doing what in God's name?

Modelling weapon and countermeasure launches from a submarine. 

>The last 11/70 I used was VERY early '80s and ten years later I knew people
>who they were being given to in exchange for hauling them off. Turned 
>down
>one myself about '94.

Yep. This was 1988 and the 11/70 was available, while the VAX was
overloaded, so as junior team member I got volunteered to use the
PDP.

-- 
There are four kinds of homicide: felonious, excusable, justifiable and
praiseworthy...             

Paul J. Adam                                  paul@jrwlynch.demon.co.uk   

Re: [Fwd: F22 completes 11% of its Flight tests]

[Tarjei T. Jensen]

Keith Willshaw
>The trouble is too many people never think about it.
>
>The people that wrote the DEC Fortran that I got stuck
>with told me they couldn't imagine why they would ever
>move off VAX, Theywere the most cost effective machines to
>do it on.
>
>Most of our customers are now on $1500 PC's and we
>are back to Standard Fortran


I worked for a company that had a lot of DEC fortran code. They solved the
problem by buying compilers with DEC fortran extensions.

Perhaps your people choose the wrong extensions :-)


Greetings,

Re: [Fwd: F22 completes 11% of its Flight tests]

[Keith WIllshaw]

Tarjei T. Jensen <tarjei.jensen@kvaerner.com> wrote in message
news:861m2m$19d1@ftp.kvaerner.com...
>

>
>
> I worked for a company that had a lot of DEC fortran code. They solved the
> problem by buying compilers with DEC fortran extensions.
>
> Perhaps your people choose the wrong extensions :-)
>

We liiked at that option and could have gone that way for NT
but that only perpetuates the problem

Anyway we support SG , SUN , HP , RS6000 , DEC Ultrix
, NT and WIN9x

It was better to bite the bullet and go back to Ansi Standard
Fortran ( and we rewrote the Graphics stuff in C)

Keith

Re: [Fwd: F22 completes 11% of its Flight tests]

[Stefan Skoglund]

Keith WIllshaw wrote:
> We liiked at that option and could have gone that way for NT
> but that only perpetuates the problem
> 
> Anyway we support SG , SUN , HP , RS6000 , DEC Ultrix
> , NT and WIN9x
> 
> It was better to bite the bullet and go back to Ansi Standard
> Fortran ( and we rewrote the Graphics stuff in C)
                                  ^^^^^^^^^^^^^^^^
How did you get that portable ? conditional makefiles ?
or gnu autoconf ?

Re: [Fwd: F22 completes 11% of its Flight tests]

[Marin D. Condic]

jtarver wrote:
> Ada was originally envisioned to be platform transparent.  When Data General
> produced the only 100% compliant compliler and the Ada development program
> went for the noncompliant VAX as their standard platform that idea was out
> the window.
> 
It has always been known that no language can ever be 100% platform
transparent. Nobody who knew anything about language design ever
expected that Ada (or any language) could meet such an objective. The
objective was to minimize portability problems - much more likely to be
achieved.

I'm afraid I don't know about Data General building an Ada compiler and
I don't know why they would be using a VAX instead of a Data General
platform. But I do know that there are lots of "100% compliant" Ada
compilers available if, by compliant, you mean a compiler that has run
the Ada validation suite and been given a certificate stating that it
has met the standard.

You may be setting up a strawman here and then knocking him down. I
don't think anybody had any illusions about Ada being perfect in every
way. Certainly Ada had its problems in the early days because of
inefficient implementations, lack of understanding and acceptance by its
intended market and so on. But things have improved very considerably
since the first standard was issued and it is successfully used in a
wide variety of applications where it does, in fact, reduce errors,
improve productivity and enhance portability.

If there is some other language out there which is 100% platform
transparent and has no flaws at all, I'd sure like to hear about it. But
in the mean time, I think Ada does a better job than any other language
I know of for avionics and other complex systems with high reliability
requirements. If you have a better alternative, please mention what it
is.

MDC
-- 
=============================================================
Marin David Condic   - Quadrus Corporation -   1.800.555.3393
1015-116 Atlantic Boulevard, Atlantic Beach, FL 32233
http://www.quadruscorp.com/
m c o n d i c @ q u a d r u s c o r p . c o m

Visit my web site at:  http://www.mcondic.com/

"Capitalism without failure is like religion without sin." 
        --  Allan Meltzer, Economist 
=============================================================

Re: [Fwd: F22 completes 11% of its Flight tests]

[tmoran]

>Anyone that has ever done any serious programming under multiple platforms
>knows exactly what I mean by "extended instruction set".
  Do you mean the XFC instruction, generating a fault to help implement
user-written "instructions"?  Did some DEC Ada compiler rely on
some such "extended instructions"?  Why is that worse than any
other compiler on any other machine exploiting particular capabilities
of the target hardware?