Ada Dot Net ?

Ada Dot Net ?

[WJT]

I know this is probably an old and sore subject but I have to ask. Does
anyone know if any of the Ada vendors (the few left) are developing an Ada95
compiler for the MS Dot Net Environment?

Before all you GNAT/Linux freaks start ganging up on me, save it.

The reason I want to know is that I have been developing in C# lately. I
started using the command line compiler that comes free with the SDK. And
believe it or not I was generating some impressive GUI applications using
nothing more than a text editor. I do not attribute this to the superiority
of the C# language, I attribute it to the Common Language Runtime (CLR) and
its class libraries that make Windows programming .... well, actually
enjoyable. As a matter of fact, because of the class libraries C#, VB and
any other 'OO' language all look pretty similar on the Dot Net platform (and
pretty simple as well).

For the first time I have to say that MS has actually come close to hitting
the mark, they have actually done something that has benefited the
developer. The entire OS (or 99% of it) is available as an OO class library!
I just can't help but think that having an Ada95 compiler in such an
environment could only benefit Ada as a whole. For the first time Ada would
be on a level playing field with all of the other languages on that
development platform.

The real reason that I would like to have an Ada95 compiler on such a
platform (and hopefully developed as a plug-in for MS Visual Studio) is
well...just because. Because I've dedicated over 20 years of my life to the
little lady (mostly using her in embedded systems) and would really like to
see here catch the boat this time.

It's just that I'm seeing Visual COBOL and Visual FORTRAN compilers
integrated into the Dot Net environment (and MS Visual Studio) and I have a
bad case of 'class envy'!!! For 20 years I've suffered from a bad case of
API envy, but now it's class envy!!!

And look, before all of  you start flaming me with a bunch of crap about
portable GUI interfaces, the evils of the Gates Empire, just stop before you
even get started. Your dealing with someone who has spent thousands of
dollars over the years purchasing all of the PC based Ada compilers that the
industry could conjure up, someone who has programmed GUI's at the X Windows
Tool Kit level (in Ada), all the way down to scrapping the metal of the
packed pixel frame buffers (via Ada 2D arrays mapped with 'for use' at
clauses). When it comes to Ada development I have no problem saying that I
have been there and I have done that, period. I know what I want, and I know
what is needed.

I have earned the right to say that as far as I'm concerned Ada has always
missed the mark, has always fallen short of the other languages when it
comes to MS Windows development environments and tools. I would just like to
see a vendor fully embrace what may be the last chance to 'get it right'
when it comes to providing an Ada solution for serious MS Windows
application development. If there is any vendor out their even attempting
such an undertaking then God Bless and God Speed !!!

William J. Thomas

P.S. Just for the record, when it comes to Cross Compilers for Embedded
Systems nothing in the industry even comes close to Ada, not even close. Ada
has made programming large real-time embedded systems an absolute pleasure,
a pure joy compared to any other language in existence !!! If you have to
take it from the metal up and make it fly, there's nothing like Ada!!!

Re: Ada Dot Net ?

[Jerry van Dijk]

"WJT" <bthomas@aisvt.bfg.com> writes:

> I know this is probably an old and sore subject but I have to ask. Does
> anyone know if any of the Ada vendors (the few left) are developing an Ada95
> compiler for the MS Dot Net Environment?

Alas I do not know if such a beast is somewhere being thought about, but I
agree Ada should jump on the .net (.gnu ?) bandwagon. Alas, looking at how
the Java ports have been doing, it does not seem commercially interesting.

In theory, of course, a volunteer effort could do it, starting with GNAT
(or even as a new gcc target). But last time I checked getting the info
needed from MS involved signing a lot of contracts with them, and paying
a handsome sum of serious money up front.

-- 
--  Jerry van Dijk   | email: jvandyk@attglobal.net
--  Leiden, Holland  | web:   users.ncrvnet.nl/gmvdijk

Re: Ada Dot Net ?

[Ingo Marks]

Jerry van Dijk wrote:

> 
> "WJT" <bthomas@aisvt.bfg.com> writes:
> 
>> I know this is probably an old and sore subject but I have to ask. Does
>> anyone know if any of the Ada vendors (the few left) are developing an
>> Ada95 compiler for the MS Dot Net Environment?
> 
> Alas I do not know if such a beast is somewhere being thought about, but I
> agree Ada should jump on the .net (.gnu ?) bandwagon. Alas, looking at how
> the Java ports have been doing, it does not seem commercially interesting.
> 
> In theory, of course, a volunteer effort could do it, starting with GNAT
> (or even as a new gcc target). But last time I checked getting the info
> needed from MS involved signing a lot of contracts with them, and paying
> a handsome sum of serious money up front.

Expanding GNAT to .NET would be a very hard job. Wouldn't it be much easier 
to write an Ada to C# compiler? Of course that wouldn't the real meat, but 
better than nothing.

I have experience in writing compilers and I am thinking seriously about 
starting such an Ada/C# project. Is there anyone able and interested to 
support this? 

What do you think about?

Re: Ada Dot Net ?

[Florian Weimer]

Ingo Marks <adv@region-nord.de> writes:

> Expanding GNAT to .NET would be a very hard job. Wouldn't it be much easier 
> to write an Ada to C# compiler? Of course that wouldn't the real meat, but 
> better than nothing.

An Ada compiler targeting the CLR seems to be more plausible.  You
could take JGNAT and replace the code generator.  Changing the
compiler this way is probably rather straightforward, but writing the
Ada run-time library certainly isn't.

Re: Ada Dot Net ?

[Marin David Condic]

That's a big problem with a lot of the other things in Ada - in theory,
volunteers could do it. But who is likely to dedicate the time needed
without any compensation? We all might do something reasonably small, but a
project of this magnitude might get in the way of our day jobs. :-) The
current vendors might be interested in it from a speculative standpoint, but
they are likely to be loathe to risk a large sum of money porting to a
platform with no clear market segment out there eager to buy the end
product.

There might be people willing to put in some sweat equity on a venture
targeting .NET, but I doubt it could get done as a volunteer effort.

MDC
--
Marin David Condic
Senior Software Engineer
Pace Micro Technology Americas    www.pacemicro.com
Enabling the digital revolution
e-Mail:    marin.condic@pacemicro.com


"Jerry van Dijk" <jvandyk@attglobal.net> wrote in message
news:wk8z8bg4ja.fsf@attglobal.net...
>
> In theory, of course, a volunteer effort could do it, starting with GNAT
> (or even as a new gcc target). But last time I checked getting the info
> needed from MS involved signing a lot of contracts with them, and paying
> a handsome sum of serious money up front.
>

Re: Ada Dot Net ?

[Wes Groleau]

> And look, before all of  you start flaming me with a bunch of crap about
> portable GUI interfaces, the evils of the Gates Empire, just stop before you
> even get started. .....
> 
> I have earned the right to say that as far as I'm concerned Ada has always
> missed the mark, has always fallen short of the other languages when it
> comes to MS Windows development environments and tools. I would just like to

This is true.  However, part of the reason is that Ada
folks generally HAVE been concerned with portability,
while Microsoft is concerned with preventing it!

Ada on .Net would be a Good Thing, but if it
would remain truly Ada, it would be a Better Thing.

.Net and C# are basically an attempt to make a
Microsoft-controlled alternative to Java.

I'd like to see Ada on .Net, but I know that
Gates & Co. will do everything they can get away
with to make it incompatible with any other Ada,
just like they did with Java.

If you consider this a flame, so be it.

-- 
Wes Groleau
http://freepages.rootsweb.com/~wgroleau

Re: Ada Dot Net ?

[Jean-Marc Bourguet]

Wes Groleau <wesgroleau@despammed.com> writes:

> Ada on .Net would be a Good Thing, but if it
> would remain truly Ada, it would be a Better Thing.

I seem to remember that Eiffel changed its definition to better fit
the model of .NET.

A+

-- 
Jean-Marc

Re: Ada Dot Net ?

[WJT]

"Wes Groleau" <wesgroleau@despammed.com> wrote in message
news:3CA48637.C5C4DEF5@despammed.com...
>
> > And look, before all of  you start flaming me with a bunch of crap about
> > portable GUI interfaces, the evils of the Gates Empire, just stop before
you
> > even get started. .....
> >
> > I have earned the right to say that as far as I'm concerned Ada has
always
> > missed the mark, has always fallen short of the other languages when it
> > comes to MS Windows development environments and tools. I would just
like to
>
> This is true.  However, part of the reason is that Ada
> folks generally HAVE been concerned with portability,
> while Microsoft is concerned with preventing it!
>
> Ada on .Net would be a Good Thing, but if it
> would remain truly Ada, it would be a Better Thing.

The Ada95 language definition has enough flexibility to fit on the .Net
platform through the use of  'allowable and documented restrictions',
implementation defined attributes, pragmas and of course there is always
implementation specific packages.

> .Net and C# are basically an attempt to make a
> Microsoft-controlled alternative to Java.
>
> I'd like to see Ada on .Net, but I know that
> Gates & Co. will do everything they can get away
> with to make it incompatible with any other Ada,
> just like they did with Java.

Your paranoid, its flattering to think that Bill would even care enough to
make Ada his own, besides that I believe that MS is starting to understand
the benefits of ISO standardization. They have submitted C# and the Common
Language Infrastructure to the ECMA (which is what I'm told the fast track
to ISO standardization).

> If you consider this a flame, so be it.
>
> --
> Wes Groleau
> http://freepages.rootsweb.com/~wgroleau

Re: Ada Dot Net ?

[Preben Randhol]

On Fri, 29 Mar 2002 16:27:45 GMT, WJT wrote:

> make Ada his own, besides that I believe that MS is starting to understand
> the benefits of ISO standardization. They have submitted C# and the Common
> Language Infrastructure to the ECMA (which is what I'm told the fast track
> to ISO standardization).

So what? If Mircosoft controls the ISO then it doesn't matter. Mickysoft
is going to whatever they can to prevent portability.

-- 
Preben Randhol         �For me, Ada95 puts back the joy in programming.�

Re: Ada Dot Net ?

[WJT]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1001 bytes --]


"Preben Randhol" <randhol+abuse@pvv.org> wrote in message
news:slrnaa97pv.s2.randhol+abuse@kiuk0156.chembio.ntnu.no...
> On Fri, 29 Mar 2002 16:27:45 GMT, WJT wrote:
>
> > make Ada his own, besides that I believe that MS is starting to
understand
> > the benefits of ISO standardization. They have submitted C# and the
Common
> > Language Infrastructure to the ECMA (which is what I'm told the fast
track
> > to ISO standardization).
>
> So what? If Mircosoft controls the ISO then it doesn't matter. Mickysoft
> is going to whatever they can to prevent portability.

After twenty years in the Ada scheme of things I'm willing to sacrifice a
little portability in order to get the power I need on the Windows platform.
Besides this is Ada, if you architect you system correctly you should be
isolating you 'dependent things' into layers and subsystems and keeping the
rest of you application true and ...Green.

> --
> Preben Randhol         �For me, Ada95 puts back the joy in programming.�

Re: Ada Dot Net ?

[Preben Randhol]

On Fri, 29 Mar 2002 17:10:55 GMT, WJT wrote:
> 
> After twenty years in the Ada scheme of things I'm willing to sacrifice a
> little portability in order to get the power I need on the Windows platform.

Which power is this?

> Besides this is Ada, if you architect you system correctly you should be
> isolating you 'dependent things' into layers and subsystems and keeping the
> rest of you application true and ...Green.

?

-- 
Preben Randhol         �For me, Ada95 puts back the joy in programming.�

Re: Ada Dot Net ?

[WJT]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1172 bytes --]


"Preben Randhol" <randhol+abuse@pvv.org> wrote in message
news:slrnaa98qb.s2.randhol+abuse@kiuk0156.chembio.ntnu.no...
> On Fri, 29 Mar 2002 17:10:55 GMT, WJT wrote:
> >
> > After twenty years in the Ada scheme of things I'm willing to sacrifice
a
> > little portability in order to get the power I need on the Windows
platform.
>
> Which power is this?

The power that is offered by the underlying OS (of which a big part of is
the GUI), the same power achieved when I was willing to sacrifice
portability when using all of the features of the VMS operating system
(anyone remember VAX Ada, Ada was a first class citizen under VMS, it lacked
nothing)?

>
> > Besides this is Ada, if you architect you system correctly you should be
> > isolating you 'dependent things' into layers and subsystems and keeping
the
> > rest of you application true and ...Green.

Green, Green ... before Ada was Ada the baseline for what we now come to
know and love was referred to as the 'Green' version of the language (during
the 'design competition' phase of the language's development).

>
> ?
>
> --
> Preben Randhol         �For me, Ada95 puts back the joy in programming.�

Re: Ada Dot Net ?

[Pascal Obry]

Jean-Marc Bourguet <jm@bourguet.org> writes:

> Wes Groleau <wesgroleau@despammed.com> writes:
> 
> > Ada on .Net would be a Good Thing, but if it
> > would remain truly Ada, it would be a Better Thing.
> 
> I seem to remember that Eiffel changed its definition to better fit
> the model of .NET.

Well they do not have changed Eiffel definition, but they have created an
Eiffel .NET language. For example on .NET there is no multiple inheritance...

Pascal.

-- 

--|------------------------------------------------------
--| Pascal Obry                           Team-Ada Member
--| 45, rue Gabriel Peri - 78114 Magny Les Hameaux FRANCE
--|------------------------------------------------------
--|         http://perso.wanadoo.fr/pascal.obry
--|
--| "The best way to travel is by means of imagination"

Re: Ada Dot Net ?

[Wes Groleau]

> Your paranoid, its flattering to think that Bill would even care enough to
> make Ada his own, besides that I believe that MS is starting to understand
> the benefits of ISO standardization. They have submitted C# and the Common
> Language Infrastructure to the ECMA (which is what I'm told the fast track
> to ISO standardization).

It's not paranoia, it's an accurate extrapolation
of past performance.  Now, if they're changing their
ways, that's cool.  It would certainly be the smart
thing to do.

-- 
Wes Groleau
http://freepages.rootsweb.com/~wgroleau

Re: Ada Dot Net ?

[Adrian Hoe]

Ingo Marks wrote:
> 
> Jerry van Dijk wrote:
> 
> >
> > "WJT" <bthomas@aisvt.bfg.com> writes:
> >
> >> I know this is probably an old and sore subject but I have to ask. Does
> >> anyone know if any of the Ada vendors (the few left) are developing an
> >> Ada95 compiler for the MS Dot Net Environment?
> >
> > Alas I do not know if such a beast is somewhere being thought about, but I
> > agree Ada should jump on the .net (.gnu ?) bandwagon. Alas, looking at how
> > the Java ports have been doing, it does not seem commercially interesting.
> >
> > In theory, of course, a volunteer effort could do it, starting with GNAT
> > (or even as a new gcc target). But last time I checked getting the info
> > needed from MS involved signing a lot of contracts with them, and paying
> > a handsome sum of serious money up front.
> 
> Expanding GNAT to .NET would be a very hard job. Wouldn't it be much easier
> to write an Ada to C# compiler? Of course that wouldn't the real meat, but
> better than nothing.
> 
> I have experience in writing compilers and I am thinking seriously about
> starting such an Ada/C# project. Is there anyone able and interested to
> support this?
> 
> What do you think about?


Most of the people use Ada for one compelling reason: non-proprietary
open system technology! C# is M$ proprietary language and is just
another attempt to outrage Java. I think Ada has moved along many OS
including Windows in a very steady and unique direction. I haven't been
programming in C# and will never do but I have taken a look at C# and I
personally don't like it.

Ada has interface to some major programming languages (I don't have to
mention the names). My question is ... Why bother to mess with a
programming language like C# which is not the mainstream? C# has not
even completed the adoption life cycle (I may be wrong)!
-- 
Remove *nospam* to email.              -- Adrian Hoe
                                       -- http://adrianhoe.com

Re: Ada Dot Net ?

[Alexandre E. Kopilovitch]

"WJT" <bthomas@aisvt.bfg.com> wrote:
>The reason I want to know is that I have been developing in C# lately. I
>started using the command line compiler that comes free with the SDK. And
>believe it or not I was generating some impressive GUI applications using
>nothing more than a text editor. I do not attribute this to the superiority
>of the C# language, I attribute it to the Common Language Runtime (CLR) and
>its class libraries that make Windows programming .... well, actually
>enjoyable.

That your excitement clearly shows that you have no previous experience with
Borland Delphi. Well. it isn't suprising that the chief architect of Delphi,
and the author of C#, and probably one of the main architectcs of the .NET is
the same person.

>The entire OS (or 99% of it) is available as an OO class library!

This is a huge overstatement, I think. It may be so for major part of GUI and
some broad classes of services, but very far from "entire OS".

>The real reason that I would like to have an Ada95 compiler on such a
>platform (and hopefully developed as a plug-in for MS Visual Studio) is
>well...just because. Because I've dedicated over 20 years of my life to the
>little lady (mostly using her in embedded systems) and would really like to
>see here catch the boat this time.

This isn't a technical reason. And Ada tries to follow the technical reasons.

>I know what I want, and I know what is needed.

Well, then, perhaps, you can express that in technical terms?

>P.S. Just for the record, when it comes to Cross Compilers for Embedded
>Systems nothing in the industry even comes close to Ada, not even close. Ada
>has made programming large real-time embedded systems an absolute pleasure,
>a pure joy compared to any other language in existence !!!

Well, good. Every programming language has its "virtual machine", sometimes
more or less specified explicitly, sometimes hidden. Ada is unique (at least
among the standartized languages) in that respect: there are TWO different
virtual machines - one for a host, and another for a target. And when you
speak about CLR you touch the target virtual machine, while when you speak
about Visual Studio plug-in, you touch the host virtual machine.
  It isn't clear from your speech, which one you consider most important to be
plugged to the .NET .

Wes Groleau <wesgroleau@despammed.com> wrote:
>.Net and C# are basically an attempt to make a
>Microsoft-controlled alternative to Java.

That may be partially true, but notice, that "Microsoft-controlled" is
secondary, while "alternative to Java" is primary, and very desirable thing
- not for Microsoft only.

>I'd like to see Ada on .Net, but I know that
>Gates & Co. will do everything they can get away
>with to make it incompatible with any other Ada,
>just like they did with Java.

This surely will be so for Microsoft Visual Ada++, but not for an independent
vendor, For example,

"WJT" <bthomas@aisvt.bfg.com> wrote:
>The Ada95 language definition has enough flexibility to fit on the .Net
>platform through the use of  'allowable and documented restrictions',
>implementation defined attributes, pragmas and of course there is always
>implementation specific packages.

Ada95 does not have enough flexibility to fit immediately on a platform which
have many vaguely specified but still powerful features. Don't forget, that
Microsoft always stands for "freedom for innovation", which effectively means
"no guarantee for any feature".


Alexander Kopilovitch                      aek@vib.usr.pu.ru
Saint-Petersburg
Russia

Re: Ada Dot Net ?

[Preben Randhol]

On Sat, 30 Mar 2002 05:08:32 +0300 (MSK), Alexandre E. Kopilovitch wrote:

> Microsoft always stands for "freedom for innovation", which effectively means
> "no guarantee for any feature".

I thought they stood for: 
"I'm the king of the hill, so when I say bugs are features, bugs are features!"

-- 
Preben Randhol         �For me, Ada95 puts back the joy in programming.�

Re: Ada Dot Net ?

[tony]

WJT wrote:
> 
> "Preben Randhol" <randhol+abuse@pvv.org> wrote in message
> news:slrnaa97pv.s2.randhol+abuse@kiuk0156.chembio.ntnu.no...
> > On Fri, 29 Mar 2002 16:27:45 GMT, WJT wrote:
> >
> > > make Ada his own, besides that I believe that MS is starting to
> understand
> > > the benefits of ISO standardization. They have submitted C# and 
> Besides this is Ada, if you architect you system correctly you should be
> isolating you 'dependent things' into layers and subsystems and keeping the
> rest of you application true and ...Green.
> 
> > --
> > Preben Randhol         �For me, Ada95 puts back the joy in programming.�

Agreed but isn't it fairly easy (courtesy of David Botton) to produce
internet server applications for windows clients, already without
touching .NET.

still such an effort to produce .NET for ada (or is it the other way
round) can only progress both our community and enhance the reliability
of using .NET. 

This is to be at least gently admired...........

Re: Ada Dot Net ?

[Preben Randhol]

On Sat, 30 Mar 2002 12:48:54 +0000 (UTC), tony wrote:
> WJT wrote:
>> 
>> "Preben Randhol" <randhol+abuse@pvv.org> wrote in message
>> news:slrnaa97pv.s2.randhol+abuse@kiuk0156.chembio.ntnu.no...
>> > On Fri, 29 Mar 2002 16:27:45 GMT, WJT wrote:
>> >
>> > > make Ada his own, besides that I believe that MS is starting to
>> understand
>> > > the benefits of ISO standardization. They have submitted C# and 
>> Besides this is Ada, if you architect you system correctly you should be
>> isolating you 'dependent things' into layers and subsystems and keeping the
>> rest of you application true and ...Green.
>> 
>> > --
>> > Preben Randhol         �For me, Ada95 puts back the joy in programming.�
> 
> Agreed but isn't it fairly easy (courtesy of David Botton) to produce
> internet server applications for windows clients, already without
> touching .NET.
> 
> still such an effort to produce .NET for ada (or is it the other way
> round) can only progress both our community and enhance the reliability
> of using .NET. 
> 
> This is to be at least gently admired...........


I didn't write that you quoted above. Pleace quote correctly. 

-- 
Preben Randhol         �For me, Ada95 puts back the joy in programming.�

Re: Ada Dot Net ?

[Steve Doiel]

> Microsoft always stands for "freedom for innovation", which effectively
means
> "no guarantee for any feature".

Gee.  I always thought Micro$oft stood for "freedom from competative
innovation" ;-)

Re: Ada Dot Net ?

[William J. Thomas]

We'll I would like to thank all of the Ada vendors for replying, I can see
that the future of Ada on the PC is in good hands (as it's always been).

Thank God I earn my paycheck using Ada on embedded systems.  I will just
have to be content knowing that Ada has found her nitch market in safety
critical systems, large complex weapon systems, and a few mission critical
commercial efforts.

As for serious PC Windows development (for things like lab support and test
software) I'm afraid that C# is rapidly winning my heart.

The company I work for is undergoing a serious process improvement effort,
we are standardizing on a great many things (thank God one of the things is
Ada95 for embedded systems, C comes in a rare second and only when Ada is
not available). We are also trying to standardize on the languages used for
future test equipment software development (this type of software is GUI
intensive, and makes extensive use of fancy I/O cards).

Many of the software engineers involved in the standardization effort would
love to use Ada even for this class of applications, but when push comes to
shove Ada can't hold a candle to any of the other languages coupled with
their GUI environments. The MS Visual Studio supported languages can't be
beat for their level of integration into the underlying OS and almost every
I/O board on the market comes with drivers/APIs for their products.

You can't ask an Ada zealot (in his right mind) to strap on one of the
existing PC native Ada compilers (with its inferior Windows GUI development
environment), have him walk into the current market place and expect him to
compete against the likes of ordinance the competition will bring to bear on
the same class of problems. That is the reality of the situation, believe me
I love Ada (spent my 1st honeymoon at the 2nd International Ada Tech
conference, gee maybe that's one of the reasons I had to have a second
honeymoon), but in the area of serious PC Windows based development the
vendors have fallen short of the mark. And from the responses I received on
this posting I can see that the vendors are not doing much (if anything at
all) to rectify this situation.

You'll have to excuse me, I thing I hear the System.Windows.Forms namespace
calling me.


WJT

Re: Ada Dot Net ?

[tmoran]

> but in the area of serious PC Windows based development the vendors
> have fallen short of the mark.  And from the responses I received on
> this posting I can see that the vendors are not doing much (if anything
> at all) to rectify this situation.
  Have you tried CLAW (Class Library Ada Windows) or its GUI Builder,
or Object Ada's GUI Builder, or GWindows, or GtkAda or the other tools
for building Windows apps using Ada?  Or are you assuming that if no
vendors responded to your post then there must be no vendors.  I'm not
a vendor responding, but I am one of the developers of CLAW, and I find
it easy to write Windows programs of various sorts using those tools.

Re: Ada Dot Net ?

[Al Christians]

I like Ada, but this GUI thing is one where it's not that strong.

moran@acm.org wrote:
> 
>   Have you tried CLAW (Class Library Ada Windows) or its GUI Builder,
You have admitted, haven't you, that you wished that you could add more 
to CLAW, that the set of widgets that CLAW supports is way less than
what comes with some of the more popular products for Windows (eg VB, 
VC++, Delphi)?

> or Object Ada's GUI Builder?

Similarly limited, isn't it? The demo I saw for Object Ada's GUI setup
(some years back) gave a simple little demo that it boasted was 
developed in something like 1/2 day.  Very nice. But it was the kind 
of demo that someone would throw together with the alternative tools 
I mentioned above in about 5 minutes.

> or GWindows

No GUI builder AFAIK.  Doesn't match the other tools (above) for 
number of widgets or ease of use.

> GtkAda or the other tools

Last time I looked at GtkAda, the Windows version of GtkAda depended on
the Windows version of Gtk, and the Windows version of Gtk said that it
was beta or alpha or developers' code, not intended for production use.

My favorite of the Ada GUI tools was RAPID, which gave a very small set
of Tk widgets.  It was real simple and easy to use.  Fine for a program
where you could get by with the very limited set of widgets, ie only
some simple programs, but very usable and competitive in time required 
to make an app. But not a match for Delphi if you want to do a high-
powered UI.   

The last version of Delphi I bought came with something like 150
widgets/
controls builtin.  I don't think that any of the Ada alternatives you 
mention can come near that.  Then, when I look to buy more, there are 
many many available for Delphi to add on that can be bought cheap or 
downloaded free that install right into the IDE and work just fine right 
away by point and click.  You can find the same for VB or VC++.

This is not a battle that Ada is going to win.  David Botton has the
right idea with Gnatcom.  Use Ada to write the backend logic where Ada
is not at a disadvantage.  Use the high-volume mass market UI tools to
do the UI.  Hook them together.

Al

Re: Ada Dot Net ?

[tmoran]

> You have admitted, haven't you, that you wished that you could add more
> to CLAW, that the set of widgets that CLAW supports is way less than
> what comes with some of the more popular products for Windows (eg VB,
> VC++, Delphi)?
  To admit that one can think of new nice things is hardly a
condemnation of what does exist.  The relevant question is how many
things are left out that a) are regularly useful and b) are difficult
to roll your own.

> This is not a battle that Ada is going to win.
  When we first started on CLAW the intent was to use the strengths of Ada
to make it easy for Ada programmers to write robust Windows programs.  For
instance Ada tasking is used to handle the Windows message loop, objects
like brushes are Controlled so they don't leak, errors are handled with
exceptions instead of return codes so you don't "forget" to test for error
returns, strong typing helps in the usual ways, and the developers of CLAW
widgets take care of a lot of learning of those little details MS forgot
to mention, or mentioned wrong.  If the important thing is how many exotic
widgets you need, then by all means grab the latest widgets hot off the
press in whatever language they present themselves.  If that language is
C* then the downside is having to switch to think and work in C terms and
get all the language help you usually get from C*.  If you are just
writing small amounts of glue to hook together various widgets that
probably isn't a problem.  That is not a battlefield where Ada can really
bring its advantages to bear.  There are, even in the Windows world, other
battlefields, where other aspects are more important.

Re: Ada Dot Net ?

[Preben Randhol]

On Sun, 31 Mar 2002 00:09:03 -0800, Al Christians wrote:
>> GtkAda or the other tools
> 
> Last time I looked at GtkAda, the Windows version of GtkAda depended on
> the Windows version of Gtk, and the Windows version of Gtk said that it
> was beta or alpha or developers' code, not intended for production use.

Are you talking about the 2.0 series? It was just released as was Gtk+
2.0, so you must wait a bit for the windows and linux binaries.
Especially for the windows Gtk+ 2.0 as this is a port.


-- 
Preben Randhol         �For me, Ada95 puts back the joy in programming.�

Re: Ada Dot Net ?

[Ingo Marks]

WJT wrote:

> The real reason that I would like to have an Ada95 compiler on such a
> platform (and hopefully developed as a plug-in for MS Visual Studio) is
> well...just because. Because I've dedicated over 20 years of my life to
> the little lady (mostly using her in embedded systems) and would really
> like to see here catch the boat this time.
 
> I have earned the right to say that as far as I'm concerned Ada has always
> missed the mark, has always fallen short of the other languages when it
> comes to MS Windows development environments and tools. I would just like
> to see a vendor fully embrace what may be the last chance to 'get it
> right' when it comes to providing an Ada solution for serious MS Windows
> application development. If there is any vendor out their even attempting
> such an undertaking then God Bless and God Speed !!!

I don't know if any vendor plans to provide .NET support for Ada, but note 
the following statements:

1) DotGNU project http://www.gnu.org/projects/dotgnu/faq.html:

"1.16 What programming languages are supported in DotGNU?

  We want to support Java in the same way as C#/CLR (as per Microsoft's
  ECMA specs) will also be supported.  As soon as Parrot (the bytecode
  system of Perl6) is available, it will probably be added to the list
  of bytecode systems that we want to support.

  Note that with the support for Java bytecode, all other programming
  languages that can be compiled to Java bytecode will also be
  available, for example Ada95.  Also, all programs that are written
  for .NET (for example in Visual Basic) should work with DotGNU."

2) GNU GCC 3.1 will integrate the former Ada frontend into its core.
See http://www.gnu.org/software/gcc/gcc.html:

"Currently GCC contains front ends for C, C++, Objective C, Chill, Fortran, 
and Java as well as libraries for these languages (libstdc++, libgcj,...). 
The next major release, GCC 3.1, will also include an Ada front end."

3) Mono project (http://www.go-mono.org/faq.html):

"Question 62: Are you working on a GCC front-end to C#? A GCC backend that 
will generate CIL images? What about making a front-end to GCC that takes 
CIL images and generates native code? 

We are currently seeking volunteers for those projects."

Re: Ada Dot Net ?

[Al Christians]

tmoran@acm.org wrote:
> 
> If the important thing is how many exotic
> widgets you need

I wasn't typing about 'exotic'.  I was typing about things that 
a windows user wouldn't call 'exotic':  grids (as in spreadsheets 
and similar tables to browse and update a database with data-
aware picklists or fields with edit masks in the columns),  
reports, charts and graphs, multi-column virtual lists (as in the 
windows event log), interfaces to the midi and media player, web 
browsers, printer setup and print dialogs, search and replace 
dialogs, multi-page wizard forms, forms that give an app a UI like 
IE or Outlook, calculators, data-entry widgets that work with edit 
masks, date entry widgets with pop-up calendars, numeric entry 
widgets with pop-up calculators,  etc, etc, etc, etc.

Al

Re: Ada Dot Net ?

[William J. Thomas]

I've tried them all. I've crashed them all. I've cursed them all. I've
deleted them all from my system shortly after they were installed. The only
things that has lasted, and not because it is any good, but because I was
forced to work with it was the GUI builder in Object Ada. And even that tool
you had to tippy toe around it, pay it compliments on a regular basis,
create all of you visuals, stop using the GUI builder and they do the rest
by 'hand'. Thanks but no thanks.


<tmoran@acm.org> wrote in message
news:umyp8.198$7s2.138773555@newssvr14.news.prodigy.com...
> > but in the area of serious PC Windows based development the vendors
> > have fallen short of the mark.  And from the responses I received on
> > this posting I can see that the vendors are not doing much (if anything
> > at all) to rectify this situation.
>   Have you tried CLAW (Class Library Ada Windows) or its GUI Builder,
> or Object Ada's GUI Builder, or GWindows, or GtkAda or the other tools
> for building Windows apps using Ada?  Or are you assuming that if no
> vendors responded to your post then there must be no vendors.  I'm not
> a vendor responding, but I am one of the developers of CLAW, and I find
> it easy to write Windows programs of various sorts using those tools.

Re: Ada Dot Net ?

[William J. Thomas]

Thanks for the intelligent reply Ingo. Looks like there may be some hope.

Re: Ada Dot Net ?

[John R. Strohm]

For test equipment, using fancy I/O cards, you need to be looking at other
things besides general-purpose programming languages, with or without GUI
toolkits.  Specifically, you need to be looking REAL HARD at National
Instruments down in Austin, with their LabView and LabWindows tools (as I
recall).  They are about as far ahead of MSVC++ as MSVC++ is ahead of the
various PC native Ada compilers.

"William J. Thomas" <WJThomas@WCVT.COM> wrote in message
news:uadamv9fot6vda@corp.supernews.com...
> The company I work for is undergoing a serious process improvement effort,
> we are standardizing on a great many things (thank God one of the things
is
> Ada95 for embedded systems, C comes in a rare second and only when Ada is
> not available). We are also trying to standardize on the languages used
for
> future test equipment software development (this type of software is GUI
> intensive, and makes extensive use of fancy I/O cards).
>
> Many of the software engineers involved in the standardization effort
would
> love to use Ada even for this class of applications, but when push comes
to
> shove Ada can't hold a candle to any of the other languages coupled with
> their GUI environments. The MS Visual Studio supported languages can't be
> beat for their level of integration into the underlying OS and almost
every
> I/O board on the market comes with drivers/APIs for their products.
>
> You can't ask an Ada zealot (in his right mind) to strap on one of the
> existing PC native Ada compilers (with its inferior Windows GUI
development
> environment), have him walk into the current market place and expect him
to
> compete against the likes of ordinance the competition will bring to bear
on
> the same class of problems. That is the reality of the situation, believe
me
> I love Ada (spent my 1st honeymoon at the 2nd International Ada Tech
> conference, gee maybe that's one of the reasons I had to have a second
> honeymoon), but in the area of serious PC Windows based development the
> vendors have fallen short of the mark. And from the responses I received
on
> this posting I can see that the vendors are not doing much (if anything at
> all) to rectify this situation.

Re: Ada Dot Net ?

[Al Christians]

Preben Randhol wrote:
> 
> On Sun, 31 Mar 2002 00:09:03 -0800, Al Christians wrote:
> >> GtkAda or the other tools
> >
> > Last time I looked at GtkAda, the Windows version of GtkAda depended on
> > the Windows version of Gtk, and the Windows version of Gtk said that it
> > was beta or alpha or developers' code, not intended for production use.
> 
> Are you talking about the 2.0 series? It was just released as was Gtk+
> 2.0, so you must wait a bit for the windows and linux binaries.
> Especially for the windows Gtk+ 2.0 as this is a port.
> 

I go to the GtkAda main page.  From there, there is a link to the Gtk+
home page.  (It looks like GtkAda uses Gtk+) From there, there is a link 
to the Gtk+ and Gimp for Windows page.  That page says:

	"Warnings

        The program(s) might crash unexpectedly or behave otherwise 			
strangely. ...   The stability seems to depend a lot on the 		machine,
display drivers, other software installed,
        and whatnot. ... Many people with ATI Rage Pro cards have 		
complained about display errors.  ... If you have NT 4, and a 			
Microsoft Wheel Mouse, some people tell me GIMP 
	hangs unless you remove the mouse icon from the taskbar, or 			update
the mouse software to the latest version. Go figure." 

I'll admit that GtkAda has made plenty of progress over recent years and
that some of the demos now look very nice.  But these kinds of warnings
and possible mysterious incompatibilities make it not-usable for much
of my work.

No offense.  I know this stuff is hard.  But if they warn me about what
might happen, I have to believe that what they say might happen might 
happen.  My applications are not important enough for me to tell my
users that they have to start deleting other software and tracking down
new drivers.  


Al

Re: Ada Dot Net ?

[Wes Groleau]

> 2) GNU GCC 3.1 will integrate the former Ada frontend into its core.
> See http://www.gnu.org/software/gcc/gcc.html:

What do they mean by "former" ? .....

-- 
Wes Groleau
http://freepages.rootsweb.com/~wgroleau

Re: Ada Dot Net ?

[Marin David Condic]

"William J. Thomas" <WJThomas@WCVT.COM> wrote in message
news:uadamv9fot6vda@corp.supernews.com...
> We'll I would like to thank all of the Ada vendors for replying, I can see
> that the future of Ada on the PC is in good hands (as it's always been).
>
I take that to be sarcasm. Generally, you won't hear the Ada vendors respond
to much in this forum. *Maybe* if you make a bogus claim about their
products or gratuitously insult the quality of their compilers, you might
see a post correcting that perception, but we just don't see a lot of
"Speaking For Compilers-R-Us" sorts of posts here when the topic is "Where
should Ada be heading?". Too bad. It would often be nice to hear the
vendor's positions on various subjects that come up here.


> Thank God I earn my paycheck using Ada on embedded systems.  I will just
> have to be content knowing that Ada has found her nitch market in safety
> critical systems, large complex weapon systems, and a few mission critical
> commercial efforts.
>
Ada has an easier time fitting into that niche. Unfortunately, many embedded
developers pay little attention to it.


>
> The company I work for is undergoing a serious process improvement effort,
> we are standardizing on a great many things (thank God one of the things
is
> Ada95 for embedded systems, C comes in a rare second and only when Ada is
> not available). We are also trying to standardize on the languages used
for
> future test equipment software development (this type of software is GUI
> intensive, and makes extensive use of fancy I/O cards).
>
Maybe you wouldn't mind telling us who you work for? Its interesting (and
admirable) to notice that some company is making sound decisions about
software development and including Ada in their best practices.


> Many of the software engineers involved in the standardization effort
would
> love to use Ada even for this class of applications, but when push comes
to
> shove Ada can't hold a candle to any of the other languages coupled with
> their GUI environments. The MS Visual Studio supported languages can't be
> beat for their level of integration into the underlying OS and almost
every
> I/O board on the market comes with drivers/APIs for their products.
>
I wouldn't expect Ada to attempt to compete head-to-head with an MS built
product. It would lose - if for no other reason than MS will always be able
to keep changing the playing field. Ada's strength (one of them, at least)
lies in its ability to be operating-system-agnostic. You might try looking
at GtkAda as a means of developing apps with some level of OS independence -
well, as long as its one of two kinds of OS. :-) Ada would benefit from
defining *something* as a kind of standard GUI interface or GUI
development/execution environment so it could be used more readily to build
portable apps on a variety of platforms. But that, of course, requires some
consensus in the community - and in particular from the vendors, and, well,
we don't hear much from them. :-)



MDC
--
Marin David Condic
Senior Software Engineer
Pace Micro Technology Americas    www.pacemicro.com
Enabling the digital revolution
e-Mail:    marin.condic@pacemicro.com

Re: Ada Dot Net ?

[WJT]

"John R. Strohm" <strohm@airmail.net> wrote in message
> For test equipment, using fancy I/O cards, you need to be looking at other
> things besides general-purpose programming languages, with or without GUI
> toolkits.  Specifically, you need to be looking REAL HARD at National
> Instruments down in Austin, with their LabView and LabWindows tools (as I
> recall).  They are about as far ahead of MSVC++ as MSVC++ is ahead of the
> various PC native Ada compilers.

Right you are, there is a whole camp here that is pushing LabView.

Re: Ada Dot Net ?

[Ted Dennison]

Al Christians <achrist@easystreet.com> wrote in message news:<3CA7DB49.31633D7@easystreet.com>...
> Preben Randhol wrote:
> 	"Warnings
> 
>         The program(s) might crash unexpectedly or behave otherwise 			
> strangely. ...   The stability seems to depend a lot on the 		machine,
> display drivers, other software installed,
...
> I'll admit that GtkAda has made plenty of progress over recent years and
> that some of the demos now look very nice.  But these kinds of warnings
> and possible mysterious incompatibilities make it not-usable for much
> of my work.

How is this substantially different from any other CYA warnings you
see on software today (for example, the famous "not for use in
safety-critical systems" warning that comes with Java)?

-- 
T.E.D.
Home     -  mailto:dennison@telepath.com (Yahoo: Ted_Dennison)
Homepage -  http://www.telepath.com/dennison/Ted/TED.html

Re: Ada Dot Net ?

[Al Christians]

Ted Dennison wrote:
> 
> How is this substantially different from any other CYA warnings you
> see on software today (for example, the famous "not for use in
> safety-critical systems" warning that comes with Java)?
> 

As I understand it, the restriction on use of Java arises 
because the JVM has a garbage collector that may or may not 
run when it wants to, so real-time processing cannot be 
guaranteed.  IAC, the restriction on use of java applies only 
to 'any nuclear facility', which does not include me or my
likely users.

But I'm brave, so I'll tentatively go along with your logic and
assume that there is no difference. This would mean that if there 
were as many users of GtkAda programs as users of java programs, 
and, if developers using GtkAda were enjoying commercial success 
to the same extent as java developers,  and if I had found as 
many good GtkAda applications usable on my machine as I have 
java apps running on my machine, and if a firm as large and 
successful as Sun was as committed to the success of GtkAda as
Sun is committed to the success of java, then I would consider 
using GtkAda.  

Lacking such evidence for high usability of GtkAda, I've got to 
take the warnings on the Gtk for Windows website seriously.


Al

Re: Ada Dot Net ?

[Greg C]

Pascal Obry <p.obry@wanadoo.fr> wrote in message news:<u3cyjfbn9.fsf@wanadoo.fr>...
> Jean-Marc Bourguet <jm@bourguet.org> writes:
> 
> > Wes Groleau <wesgroleau@despammed.com> writes:
> > 
> > > Ada on .Net would be a Good Thing, but if it
> > > would remain truly Ada, it would be a Better Thing.
> > 
> > I seem to remember that Eiffel changed its definition to better fit
> > the model of .NET.
> 
> Well they do not have changed Eiffel definition, but they have created an
> Eiffel .NET language. For example on .NET there is no multiple inheritance...

Not true. The Beta version had some restrictions, but no MI was not
one of them. There are limitations on using MI when integrating Eiffel
components with other languages, but that's really because of
limitations in the other languages.

Greg

Re: Ada Dot Net ?

[Randy Brukardt]

William J. Thomas wrote in message ...
>I've tried them all. I've crashed them all.

I very much doubt that you crashed the Claw Builder, and if you did, you
didn't report it to us. We've only had one bug report on the Builder
about a real crash (where the Builder exits and the project can't be
saved), and we've only had a few where the Builder refused to do
something. (Most of the reported bugs are cases where the resulting
program works differently from the behavior specified).

If you did crash the Claw Builder, we would have very much appreciated
sending in a bug report. Otherwise, the odds of a problem being
encountered and getting fixed is slim-to-none.

I can certainly imagine finding it too limited for your purposes, but
that's very different than crashing...

>I've cursed them all. I've
>deleted them all from my system shortly after they were installed. The
only
>things that has lasted, and not because it is any good, but because I
was
>forced to work with it was the GUI builder in Object Ada. And even that
tool
>you had to tippy toe around it, pay it compliments on a regular basis,
>create all of you visuals, stop using the GUI builder and they do the
rest
>by 'hand'. Thanks but no thanks.


I don't understand this remark at all. You have to build the logic part
of your program "by hand" -- how else can you write code?  If you expect
the GUI Builder to write your entire application, why in the world would
you care what the programming language is? Ada brings no benefits to a
GUI Builder; it's advantages exist only in the code that you write "by
hand".


As far as your original .NET question goes, we considered doing it. We
were contacted by a Microsoft rep who wanted to talk to us about it. I
wasn't in at the time, so someone got his number. I called him back many
times, but never got a response other than a short e-mail saying that he
was really busy and that he'd call back soon. So I never found out what
they wanted in order to do a .NET implementation, and thus we didn't
ever seriously think about doing one.

This has typically been the sort of response that we have gotten from
Microsoft. I think they really prefer to discourage "alternative"
development tools, and then once in a while, they try to get us to give
them some money. Kinda like sleeping with a gorilla.

              Randy.

Re: Ada Dot Net ?

[Eric G. Miller]

In <3CA87193.F6EEB08F@despammed.com>, Wes Groleau wrote:

> 
>> 2) GNU GCC 3.1 will integrate the former Ada frontend into its core.
>> See http://www.gnu.org/software/gcc/gcc.html:
> 
> What do they mean by "former" ? .....

I believe, in this context, they mean the "front end that was previously
a separate add-on product will be integrated into the compiler collection".

It's pretty clear from the gcc website that one shouldn't get there hopes
up too high about the quality of the Ada compiler that will ship with 3.1.
It's clearly not their highest priority (fixing their C++ ABI, again, seems
to be).  If I had to guess an order of importance, it'd be C, C++, Fortran,
Java, Objective C, Ada, Chill (maybe dead).  Might swap Java/Fortran...

One could hope that direct inclusion of Ada will increase interest in Ada
in the user community which will therefore spurn interest in the gcc
community.

Re: Ada Dot Net ?

[Pascal Obry]

gmc444@yahoo.com (Greg C) writes:

> Pascal Obry <p.obry@wanadoo.fr> wrote in message news:<u3cyjfbn9.fsf@wanadoo.fr>...
> > Jean-Marc Bourguet <jm@bourguet.org> writes:
> > 
> > > Wes Groleau <wesgroleau@despammed.com> writes:
> > > 
> > > > Ada on .Net would be a Good Thing, but if it
> > > > would remain truly Ada, it would be a Better Thing.
> > > 
> > > I seem to remember that Eiffel changed its definition to better fit
> > > the model of .NET.
> > 
> > Well they do not have changed Eiffel definition, but they have created an
> > Eiffel .NET language. For example on .NET there is no multiple inheritance...
> 
> Not true. The Beta version had some restrictions, but no MI was not
> one of them. There are limitations on using MI when integrating Eiffel
> components with other languages, but that's really because of
> limitations in the other languages.

Oh, well I'm sure I have read that somewhere... I should have checked myself.
Sorry for the misinformation.

Pascal.

-- 

--|------------------------------------------------------
--| Pascal Obry                           Team-Ada Member
--| 45, rue Gabriel Peri - 78114 Magny Les Hameaux FRANCE
--|------------------------------------------------------
--|         http://perso.wanadoo.fr/pascal.obry
--|
--| "The best way to travel is by means of imagination"

Re: Ada Dot Net ?

[Stephen Leake]

"Eric G. Miller" <egm2@jps-nospam.net> writes:

> It's pretty clear from the gcc website that one shouldn't get there hopes
> up too high about the quality of the Ada compiler that will ship with 3.1.
> It's clearly not their highest priority (fixing their C++ ABI, again, seems
> to be).  If I had to guess an order of importance, it'd be C, C++, Fortran,
> Java, Objective C, Ada, Chill (maybe dead).  Might swap Java/Fortran...

"Importance of integrating" does _not_ equate with "quality"! gcc Ada
is a _very_ high quality compiler.

The GNAT compiler has an extensive test suite. It's mostly
proprietary, consisting of ACT customer code sent in over the years as
bug reports. So it doesn't show up in the gcc tree. But the gcc tree
code is run against the test suite by ACT.

> One could hope that direct inclusion of Ada will increase interest
> in Ada in the user community which will therefore spurn interest in

Perhaps you meant "spur", not "spurn"? One of those nasty little
single-letter typos :).

> the gcc community.

-- 
-- Stephe

Re: Ada Dot Net ?

[Wes Groleau]

> One could hope that direct inclusion of Ada will increase interest in Ada
> in the user community which will therefore spurn interest in the gcc
> community.

Or spur it.  :-)

-- 
Wes Groleau
http://freepages.rootsweb.com/~wgroleau

Re: Ada Dot Net ?

[Robert Dewar]

Ingo Marks <adv@region-nord.de> wrote in message news:<a81gkp$f6j$04$1@news.t-online.com>...
> Expanding GNAT to .NET would be a very hard job.

Not particularly, JGNAT is a reasonable guide

> Wouldn't  it be much easier  to write an Ada to C# 
> compiler? 

No, much harder.

Re: Ada Dot Net ?

[Robert Dewar]

Al Christians <achrist@easystreet.com> wrote in message news:<3CA7DB49.31633D7@easystreet.com>...
> The program(s) might crash unexpectedly or behave 
> otherwise strangely. 


Hmm! That statement could be made about a *LOT* of software
but that does not deter people from using it :-)

In fact gtkada stability on Windows is improving very
rapidly, check out the latest release.

Re: Ada Dot Net ?

[William J. Thomas]

"Randy Brukardt" <randy@rrsoftware.com> wrote in message

> I very much doubt that you crashed the Claw Builder,

Please.... your starting to sound as aregent as me.

> and if you did, you didn't report it to us.

Your right. Besides that was many, many moons ago. Why don't I down load the
latest.... once bitten twice shy I guess.

Look I'm sorry if I got everybody's panties all twisted up, I just wanted to
know if anyone was working on an Ada .NET compiler (Robert !!!)? I guess the
answer is no.

The fact of the matter is I'm feeling kind of guilty. I went out and bought
a whole new PC, 1.8 GHz, XP the whole nine yards, just to run..., gulp,
Visual C# .NET.  And at this point in time I have not installed an Ada
compiler on it, ...and well, I feel like I'm cheating on her. (Boy do I miss
package specs and  array aggregates).

Anyway I don't feel too bad, tomorrow at work I'll be starting an evaluating
of the AdaMULTI product from Green Hills Software and I can't wait. That
aught to make up for me "seeing another" in the wee hours of the morning. I
hope she's not too sore at me.

WJT

Re: Ada Dot Net ?

[Eric G. Miller]

In <uofh26k1h.fsf@gsfc.nasa.gov>, Stephen Leake wrote:

> "Eric G. Miller" <egm2@jps-nospam.net> writes:
> 
>> It's pretty clear from the gcc website that one shouldn't get there hopes
>> up too high about the quality of the Ada compiler that will ship with 3.1.
>> It's clearly not their highest priority (fixing their C++ ABI, again, seems
>> to be).  If I had to guess an order of importance, it'd be C, C++, Fortran,
>> Java, Objective C, Ada, Chill (maybe dead).  Might swap Java/Fortran...
> 
> "Importance of integrating" does _not_ equate with "quality"! gcc Ada
> is a _very_ high quality compiler.

Well, then maybe the website needs to be fixed.  As I read it, the implication
is that the GCC community is willing to accept a higher number of defects
and/or suboptimal code generation compared with C or C++.  If they really
mean the depth of integration with the rest of the compiler, the website
should say so.  Anyway, good to hear the compiler works well -- supposed
to be released in a couple of weeks, no?

> The GNAT compiler has an extensive test suite. It's mostly
> proprietary, consisting of ACT customer code sent in over the years as
> bug reports. So it doesn't show up in the gcc tree. But the gcc tree
> code is run against the test suite by ACT.

Will some kind of results be published with the release?

>> One could hope that direct inclusion of Ada will increase interest
>> in Ada in the user community which will therefore spurn interest in
> 
> Perhaps you meant "spur", not "spurn"? One of those nasty little
> single-letter typos :).

Yes, "generate" interest not "reject".

>> the gcc community.
>

Re: Ada Dot Net ?

[Steve Doiel]

"Eric G. Miller" <egm2@jps-nospam.net> wrote in message [snip]
> Well, then maybe the website needs to be fixed.  As I read it, the
implication
> is that the GCC community is willing to accept a higher number of defects
> and/or suboptimal code generation compared with C or C++.  If they really
> mean the depth of integration with the rest of the compiler, the website
> should say so.

You're probably looking at the GCC 3.1 release criteria.  It was agreed some
time ago that the integrety of the Ada support in the 3.1 release would not
be part of the general GCC release critera, since this is the first release
of GCC to include Ada.

I fully expect Ada will become part of the critera in some subsequent
release.  From the attention Ada has been getting from the gcc folks, I
suspect the Ada support in the GCC 3.1 release will be quite good.  I would
bet that the Ada support in GCC 3.1.1 will be excellent.

SteveD

>...  Anyway, good to hear the compiler works well -- supposed
> to be released in a couple of weeks, no?

Re: Ada Dot Net ?

[Robert Dewar]

"Steve Doiel" <nospam_steved94@attbi.com> wrote in message news:<aYvq8.161861$af7.79093@rwcrnsc53>...
> You're probably looking at the GCC 3.1 release criteria.  
> It was agreed some time ago that the integrety of the Ada 
> support in the 3.1 release would not
> be part of the general GCC release criteria

Relase criteria have nothing to do with quality. They have
to do with required tests to be performed by the gcc community before
the release is permitted. Since virtually all the testing is being
done by ACT at the moment, this
is really rather irrelevant.

It certainly is not the case that the GCC 3 based GNAT is
as reliable as the GCC 2 based GNAT yet, but it is in
pretty good shape (certainly well beyond 3.14p for instance

Re: Ada Dot Net ?

[Stephen Leake]

"Eric G. Miller" <egm2@jps-nospam.net> writes:

> In <uofh26k1h.fsf@gsfc.nasa.gov>, Stephen Leake wrote:
> 
> > "Eric G. Miller" <egm2@jps-nospam.net> writes:
> > 
> >> It's pretty clear from the gcc website that one shouldn't get
> >> there hopes up too high about the quality of the Ada compiler
> >> that will ship with 3.1. It's clearly not their highest priority
> >> (fixing their C++ ABI, again, seems to be). If I had to guess an
> >> order of importance, it'd be C, C++, Fortran, Java, Objective C,
> >> Ada, Chill (maybe dead). Might swap Java/Fortran...
> > 
> > "Importance of integrating" does _not_ equate with "quality"! gcc Ada
> > is a _very_ high quality compiler.
> 
> Well, then maybe the website needs to be fixed. 

Can you provide the url for this website? The closest thing I could
find to your discussion is
http://www.gnu.org/software/gcc/gcc-3.1/criteria.html, which defines
the release criteria for gcc 3.1. Quite sensibly, they are simply
stating that gcc 3.1 will support Ada, but they are not willing to
hold up the release for Ada quality issues. That just means they are
going slow with a new thing. In this case, it happens to be someone
else's job to ensure Ada quality!

> As I read it, the implication is that the GCC community is willing
> to accept a higher number of defects and/or suboptimal code
> generation compared with C or C++. 

"Willing to accept" and "actually present" are two different things. I
have no problem with the gcc steering committee not accepting
responsibility for Gnu Ada quality; ACT is doing a perfectly good job.

> If they really mean the depth of integration with the rest of the
> compiler, the website should say so. Anyway, good to hear the
> compiler works well -- supposed to be released in a couple of weeks,
> no?

I'm not clear precisely which compiler you are talking about here. The
Ada compilers released by ACT work well; I have not tried gcc 3.x yet.

> > The GNAT compiler has an extensive test suite. It's mostly
> > proprietary, consisting of ACT customer code sent in over the
> > years as bug reports. So it doesn't show up in the gcc tree. But
> > the gcc tree code is run against the test suite by ACT.
> 
> Will some kind of results be published with the release?

ACT has said they are working towards putting what tests they can in
the gcc tree; some of their tests are open source. I don't know what
the schedule is.


-- 
-- Stephe

Re: Ada Dot Net ?

[Eric G. Miller]

In <usn6c62up.fsf@gsfc.nasa.gov>, Stephen Leake wrote:

> "Eric G. Miller" <egm2@jps-nospam.net> writes:
> 
>> In <uofh26k1h.fsf@gsfc.nasa.gov>, Stephen Leake wrote:
>> 
>> > "Eric G. Miller" <egm2@jps-nospam.net> writes:
>> > 
>> >> It's pretty clear from the gcc website that one shouldn't get
>> >> there hopes up too high about the quality of the Ada compiler
>> >> that will ship with 3.1. It's clearly not their highest priority
>> >> (fixing their C++ ABI, again, seems to be). If I had to guess an
>> >> order of importance, it'd be C, C++, Fortran, Java, Objective C,
>> >> Ada, Chill (maybe dead). Might swap Java/Fortran...
>> > 
>> > "Importance of integrating" does _not_ equate with "quality"! gcc Ada
>> > is a _very_ high quality compiler.
>> 
>> Well, then maybe the website needs to be fixed. 
> 
> Can you provide the url for this website? The closest thing I could
> find to your discussion is
> http://www.gnu.org/software/gcc/gcc-3.1/criteria.html, which defines
> the release criteria for gcc 3.1. Quite sensibly, they are simply
> stating that gcc 3.1 will support Ada, but they are not willing to
> hold up the release for Ada quality issues. That just means they are
> going slow with a new thing. In this case, it happens to be someone
> else's job to ensure Ada quality!

You've got it.
 
It states GCC version 3.1 will support Ada, but that quality
issues will not hold up the release.  How different is that from what
I originally wrote?  If it sounded like I was implying the quality
wouldn't be good, that wasn't my intention.  All that I meant was,
the 3.1 Ada compiler might not be where people would like it when the
release occurs.  I'm no muckraker ;-)

>> As I read it, the implication is that the GCC community is willing
>> to accept a higher number of defects and/or suboptimal code
>> generation compared with C or C++. 
> 
> "Willing to accept" and "actually present" are two different things. I
> have no problem with the gcc steering committee not accepting
> responsibility for Gnu Ada quality; ACT is doing a perfectly good job.

Well, I was unaware that ACT was taking a lead role in the development.
Last I had read was they had submitted code, but the impression I had
was that ACT intended the GCC community at large would further the
development of what was supplied.

>> If they really mean the depth of integration with the rest of the
>> compiler, the website should say so. Anyway, good to hear the
>> compiler works well -- supposed to be released in a couple of weeks,
>> no?
> 
> I'm not clear precisely which compiler you are talking about here. The
> Ada compilers released by ACT work well; I have not tried gcc 3.x yet.

We've been talking about GCC 3.1, no?  I got the impression from your
previous post that you had some experience with CVS snapshots, otherwise
how do you make an assertion that the quality is very good?

Anyway, not much point in dragging this discussion on.  When it's
released, we'll see where it's at.