GNAT 3.03 for OS/2

GNAT 3.03 for OS/2

[Dale Pontius]

Is there an estimated date for this?
Is it being held up by porting to emx 0.9b?
Are the changes from 3.01 to 3.01 significant?
What does it take for the AIX port to be updated?

Thanks,
Dale Pontius
(NOT speaking for IBM)

Re: GNAT 3.03 for OS/2

[Robert Dewar]

Answering Dale's queries"

   Is there an estimated date for this?
   (this = 3.03 for OS/2)

No, there is no estimated date. In general we make no advance announcements
of planned availability of public releases of GNAT.

   Is it being held up by porting to emx 0.9b?

No, it is just not done yet. We have very few customers for the OS/2
version, so it is not high priority. Indeed it would be even lower in
priority if I did not use OS/2 and nothing else in my work :-) I am
currently running OS/2 GNAT vrsion 3.04 on my machine, so likely we will
jump straight to this version and skip 3.03.

   Are the changes from 3.01 to 3.01 significant?

Not very! :-) but the changes from 3.01 to 3.03 are quite significant,
and the changes from 3.01 to 3.04 quite significant. Many important
performance enhancements, bug fixes, and some useful new functionality.

   What does it take for the AIX port to be updated?

This is not a port that we currently support. We have announced Q4
availability for support of AIX on RS6000 and Power PC. Of course
this could accelerate if there is suddenly some interest! A volunteer
(David Bacon) has recently completed the 3.03 port, so this should
be available some time soon.

GNAT 3.03 for RS/6000 AIX 3.2 (was: GNAT 3.03 for OS/2)

[David J. Bacon]

>   What does it take for the AIX port to be updated?

The binary release of GNAT 3.03 for RS/6000 AIX 3.2 is available
under the heading of Detritus on my Web home page at

  http://galt.cs.nyu.edu/~bacon

I've tried to be careful, but please be warned that this release
hasn't been checked by a member of Ada Core Technologies yet.
Okay, maybe it never will be, but hopefully that will happen soon
and then it will appear in the cs.nyu.edu ftp archive as well.

Don't worry if your browser insists on displaying the gzipped tar
image as strange-looking binary-type spam, by the way:  you can
still Save it locally via File.

Please address questions to

  bacon@cs.nyu.edu

as I do not usually receive comp.lang.ada articles (except for
those by Robert Dewar or Tucker Taft, due to a special filter).

    dB

Re: GNAT 3.03 for OS/2

[Norman H. Cohen]

In article <dewar.827529052@schonberg>, dewar@cs.nyu.edu (Robert Dewar) writes: 

|>                                               We have announced Q4
|> availability for support of AIX on RS6000 and Power PC.

Great news!  (As is David Bacon's announcement of a 3.03 port!)

I presume the ACT port will include run-time support for tasking.
Will ACT be supporting both AIX 3.2.5 and AIX 4.1?

--
Norman H. Cohen    ncohen@watson.ibm.com

Re: GNAT 3.03 for OS/2

[Robert Dewar]

Norm asks

"Great news!  (As is David Bacon's announcement of a 3.03 port!)

I presume the ACT port will include run-time support for tasking.
Will ACT be supporting both AIX 3.2.5 and AIX 4.1?"

definitely yes on the first, to be decided on the second.

Re: GNAT 3.03 for OS/2

[Dale Pontius]

In article <dewar.827529052@schonberg>, dewar@cs.nyu.edu (Robert Dewar) writes:

> No, it is just not done yet. We have very few customers for the OS/2
> version, so it is not high priority. Indeed it would be even lower in

How do you measure 'very few customers'? Is it in downloads of the
compiler? hits on some web site? or Paying Customers?

How does one become a Paying Customer of ACT? As I've said before, I'm
in a bit of a Catch-22 regarding buying a compiler, but it may be
worth a try. Perhaps after I get my current project running better I
can pursuade someone.

The OS/2 world is rather language-poor. Outside of C/C++, there are a
few commercial Pascal extensions, a few Modula-2's, a Modula-3, maybe
Eiffel soon, etc. As for workstation (AIX) <--> OS/2 portability,
there are only a few options, and Ada may well be the best of them.

Thanks,
Dale Pontius
(NOT speaking for IBM)

Re: GNAT 3.03 for OS/2

[Tore Joergensen]

Dale Pontius (pontius@twonky.btv.ibm.com) wrote:
: The OS/2 world is rather language-poor. Outside of C/C++, there are a
: few commercial Pascal extensions, a few Modula-2's, a Modula-3, maybe
: Eiffel soon, etc. As for workstation (AIX) <--> OS/2 portability,
: there are only a few options, and Ada may well be the best of them.

There are some compilers in addition to the C/C++ and Pascal/Modula-x :-).
After a quick look in the Indelible-Blue catalog, I found:
IBM VisualAge Cobol, IBM VisualAge Smalltalk, PL/1, APL2, Fortran 77,
CA-Realizer and of course several REXX implementations. If you make
GUI programs, I think VisualAge Smalltalk gives you best portability
between OS/2 and AIX (and Windows) for the moment. Of course I would
like to see faster releases of GNAT for OS/2 as well :-).
-- 
+-------------------------+-------------------------------------------+
| Tore B. Joergensen      | e-mail : tore@lis.pitt.edu                |
| Centre Court Villa      | web    : http://www.pitt.edu/~tojst1      |
| 5535 Centre Avenue # 6  |                                           |
| Pgh, PA 15232, USA      | Norwegian MSIS-student at Univ. of Pgh.   |
+-------------------------+-------------------------------------------+

Re: GNAT 3.03 for OS/2

[Norman L. Reitzel]

In article <4j6l1i$3jk@usenet.srv.cis.pitt.edu> tojst1+@pitt.edu (Tore Joergensen) writes:
>Dale Pontius (pontius@twonky.btv.ibm.com) wrote:
>: The OS/2 world is rather language-poor. Outside of C/C++, there are a
>: few commercial Pascal extensions, a few Modula-2's, a Modula-3, maybe
>: Eiffel soon, etc. As for workstation (AIX) <--> OS/2 portability,
>: there are only a few options, and Ada may well be the best of them.
>
>There are some compilers in addition to the C/C++ and Pascal/Modula-x :-).
>After a quick look in the Indelible-Blue catalog, I found:
>IBM VisualAge Cobol, IBM VisualAge Smalltalk, PL/1, APL2, Fortran 77,
>CA-Realizer and of course several REXX implementations. If you make
>GUI programs, I think VisualAge Smalltalk gives you best portability
>between OS/2 and AIX (and Windows) for the moment. Of course I would
>like to see faster releases of GNAT for OS/2 as well :-).

Don't forget Java.  The version 1.0 early release is available (upon 
(free) registry) from ncc.hursley.ibm.com.  The compiler is fast (reminds 
me of UCSD Pascal, actually), and though the runtime has bugs, it's still 
in it's infancy.


-- 
   Norman L. Reitzel, Jr.       |    "When you live beside the graveyard,
   nreitzel@lonestar.utsa.edu   |     you can't cry for every funeral."
   Blue Water Ventures, dba.    |                     Russian Proverb

Re: GNAT 3.03 for OS/2

[Ted Dennison]

Tore Joergensen wrote:
> 
> Dale Pontius (pontius@twonky.btv.ibm.com) wrote:
> : The OS/2 world is rather language-poor. Outside of C/C++, there are a
> : few commercial Pascal extensions, a few Modula-2's, a Modula-3, maybe
> : Eiffel soon, etc. As for workstation (AIX) <--> OS/2 portability,
> : there are only a few options, and Ada may well be the best of them.
...
> There are some compilers in addition to the C/C++ and Pascal/Modula-x :-).
> After a quick look in the Indelible-Blue catalog, I found:
> IBM VisualAge Cobol, IBM VisualAge Smalltalk, PL/1, APL2, Fortran 77,
> CA-Realizer and of course several REXX implementations. If you make
> GUI programs, I think VisualAge Smalltalk gives you best portability
...

IBM also has a port of Java for OS/2.


-- 
T.E.D.          
                |  Work - mailto:dennison@escmail.orl.mmc.com  |
                |  Home - mailto:dennison@iag.net              |
                |  URL  - http://www.iag.net/~dennison         |

Re: GNAT 3.03 for OS/2

[Robert Dewar]

Dale asked

-- How do you measure 'very few customers'? Is it in downloads of the
-- compiler? hits on some web site? or Paying Customers?

  I was talking about paying customers, the ones who pay our bills. They
  tend to set our priorities :-)

-- How does one become a Paying Customer of ACT? As I've said before, I'm
-- in a bit of a Catch-22 regarding buying a compiler, but it may be
-- worth a try. Perhaps after I get my current project running better I
-- can pursuade someone.

  Send mail to info@gnat.com for an automatic reply giving full details
  of our support services, or if you have specific questions, you can
  send them to support@gnat.com

-- Perhaps after I get my current project running better I
-- can pursuade someone


  That certainly makes sense if your current project is non-critical!

Re: GNAT 3.03 for OS/2

[John Howard]

On 25 Mar 1996, Dale Pontius wrote:
> In article <dewar.827529052@schonberg>, dewar@cs.nyu.edu (Robert Dewar) 
> writes:
> > No, it is just not done yet. We have very few customers for the OS/2
> > version, so it is not high priority. Indeed it would be even lower in
>
> How do you measure 'very few customers'? Is it in downloads of the
> compiler? hits on some web site? or Paying Customers?
>
> How does one become a Paying Customer of ACT? As I've said before, I'm
> in a bit of a Catch-22 regarding buying a compiler, but it may be
> worth a try. Perhaps after I get my current project running better I
> can pursuade someone.
>
> The OS/2 world is rather language-poor. Outside of C/C++, there are a
> few commercial Pascal extensions, a few Modula-2's, a Modula-3, maybe
> Eiffel soon, etc. As for workstation (AIX) <--> OS/2 portability,
> there are only a few options, and Ada may well be the best of them.
>
> Thanks,
> Dale Pontius
> (NOT speaking for IBM)

I have a few complaints about the tough situation with OS/2 and GNAT.  My 
impression is that ACT is foot-dragging in order for IBM to be persuaded 
to fund development of GNAT for OS/2 and AIX.  I don't blame ACT for 
giving priority to paying customers.  But I do blame them for not 
effectively promoting the use of GNAT to the bulk of professional OS/2 
developers.

In June of '95, I attempted to have GNAT distributed on the CD-ROM's with 
the IBM Developer Connection for OS/2.  I contacted the people at IBM and 
hooked them up with Robert Dewar who officially took over from there.  
IBM was very interested in getting GNAT distributed on their CD's and 
even contacted me several more times until September.  IBM wants to see 
an OS/2 environment that can be used for either PowerPC or Intel 
machines.  I suggested GNAT Ada 95 to them.

There was one catch: the owner of the copyrights to GNAT had to sign a 
standard release form.  At that time, the copyrights were held by several 
University entities.

Now how tough is it for ACT to sign the IBM forms and send a recent copy 
of GNAT for OS/2 to the Developer Connection staff?  In my opinion, 
Robert Dewar is being too picky by holding-off on the OS/2 releases.  The 
subscribers to IBM DEVCON are exactly the people to reach in order to 
promote Ada 95 for OS/2 and AIX.  If ACT really wanted OS/2 and AIX 
customers other than IBM, then ACT would've moved by now to access the 
subscribers of DEVCON.

I'm a non-paying customer of ACT and I must admit I'm unable to 
effectively do anything more except complain in order to get GNAT into 
the hands of OS/2 developers.

------------------------------------------------------------------------
-- John Howard <jhoward@sky.net>               -- Team Ada  Team OS/2 --
------------------------------------------------------------------------

Re: GNAT 3.03 for OS/2

[Robert Dewar]

I have a few complaints about the tough situation with OS/2 and GNAT.  My
impression is that ACT is foot-dragging in order for IBM to be persuaded
to fund development of GNAT for OS/2 and AIX.  I don't blame ACT for
giving priority to paying customers.  But I do blame them for not
effectively promoting the use of GNAT to the bulk of professional OS/2
developers.

  We are not foot dragging at all. We estimate the probability of IBM
  funding development of GNAT for OS/2 (AIX is another matter entirely)
  to be zero. We keep the OS/2 version alive really only because I use
  it, and historically it was the version required by the DoD contract.
  A company spends its promotional energy where it thinks results will
  be obtained. We have very limited resources.

In June of '95, I attempted to have GNAT distributed on the CD-ROM's with
the IBM Developer Connection for OS/2.  I contacted the people at IBM and
hooked them up with Robert Dewar who officially took over from there.
IBM was very interested in getting GNAT distributed on their CD's and
even contacted me several more times until September.  IBM wants to see
an OS/2 environment that can be used for either PowerPC or Intel
machines.  I suggested GNAT Ada 95 to them.

  Actually we persued this, but mutually agreed that we could not make the
  very tight release schedule for some particular version of the IBM
  Developer Connection CD ROM. We also agreed that they would contact
  us again for the subsequent release. We have not heard a thing from
  them. Yes I suppose we could go and track them down, but frankly
  it is not clear that IBM is really energetically promoting OS/2 at
  this stage themselves!

There was one catch: the owner of the copyrights to GNAT had to sign a
standard release form.  At that time, the copyrights were held by several
University entities.

  This is uninformed speculation. This was not, and has never been a problem!

Now how tough is it for ACT to sign the IBM forms and send a recent copy
of GNAT for OS/2 to the Developer Connection staff?  In my opinion,
Robert Dewar is being too picky by holding-off on the OS/2 releases.  The
subscribers to IBM DEVCON are exactly the people to reach in order to
promote Ada 95 for OS/2 and AIX.  If ACT really wanted OS/2 and AIX
customers other than IBM, then ACT would've moved by now to access the
subscribers of DEVCON.

  Not tough at all. If IBM gets in touch with us as they promised to do,
  we will be glad to cooperate with them. We are not "holding off" on
  the OS/2 releases, it is just that the OS/2 release is highly non-standard,
  requires a lot of special fiddling each time, and is rather low on the
  priority list, so it takes a while to get around to it. We are hoping
  to get a 3.04 out in the forseeable future, but we can't promise anything.
  Please don't assume that we are deliberately holding back OS/2 releases,
  far from it! Since I use no other version, I would certainly be as happy
  as anyone to have my version properly sychronized with the official release.

I'm a non-paying customer of ACT and I must admit I'm unable to
effectively do anything more except complain in order to get GNAT into
the hands of OS/2 developers.

  Actually, you could perfectly well take on the packaging up of new releases
  if you want to, no one has a monopoly on this. Many of the unsupported ports
  are maintained by volunteers for that matter. You could also send the stuff
  off to IBM for that matter. Remember that we do not own the copyright
  on GNAT any more than you do, so it is no easier for us to deal with
  the copyright release than it would be for you (the copyright is assigned
  to the FSF).

  One thing to understand is that are first priority is to get releases to
  our customers. Sometimes people are under the false impression that the
  GPL requires us to make general releases. It does not, the GPL only says
  that if we release, we have to release with sources. 

  ACT chooses to make public binary versions availale of all the ports we
  can, but these public releases are not our highest priority item, and
  this is especially true for ports in which there is very little
  commercial interest.

  We will certainly try to make an OS/2 3.04 as soon as possible, but we
  cannot promise any definite time table.

  Incidentally, when we use the term customer, like any other business, we
  mean paying customers. The two kinds of GNAT users who catch our
  attention are commercial customers, and those who contribute to the
  GNAT effort by volunteer work. We are happy that many others can make
  good use of GNAT, but you have to remember that the continued development
  and improvement of GNAT depends on giving our customers good service!

  Finally, John, I appreciate you making the contact with IBM for the
  developers connection. Too bad that the time scale was just too tight
  when IBM first made contact. It is a pity that IBM has not followed
  up on sending us the material for a future release.

Re: GNAT 3.03 for OS/2

[Tore Joergensen]

Robert Dewar (dewar@cs.nyu.edu) wrote:
:   Not tough at all. If IBM gets in touch with us as they promised to do,
:   we will be glad to cooperate with them. We are not "holding off" on
:   the OS/2 releases, it is just that the OS/2 release is highly non-standard,
:   requires a lot of special fiddling each time, and is rather low on the
:   priority list, so it takes a while to get around to it. We are hoping
:   to get a 3.04 out in the forseeable future, but we can't promise anything.
:   Please don't assume that we are deliberately holding back OS/2 releases,
:   far from it! Since I use no other version, I would certainly be as happy
:   as anyone to have my version properly sychronized with the official release.

I am happy for the free compiler I get. I would not be suprised if the number
of paying customers are below ten (I wouldn't even be suprised if it is zero).
It is quite obvious that the OS/2 port require more fiddling than Unix ports
(even emx can't change that). 

: I'm a non-paying customer of ACT and I must admit I'm unable to
: effectively do anything more except complain in order to get GNAT into
: the hands of OS/2 developers.

:   Actually, you could perfectly well take on the packaging up of new releases
:   if you want to, no one has a monopoly on this. Many of the unsupported ports
:   are maintained by volunteers for that matter. You could also send the stuff
:   off to IBM for that matter. 

There are lots of things to do, even if you don't want to get directly
involved in GNAT. To make GNAT attractive to proffesional developers, we
need libraries! I think the thin bindings are starting to become complete
now, but we need high-level libraries as well. A nicer developer environment
wouldn't hurt either. If you are working or plan to start working on anything
that have something to do with Ada on OS/2, feel free to write to me. I might
have information that can be useful. If you have suggestions for my 'Ada for
OS/2' page (http://www.pitt.edu/~tojst1/adaos2), please e-mail me about that
too. It's still just a puppy, but I think it will start growing into a more
useful resource soon. 
-- 
+-------------------------+-------------------------------------------+
| Tore B. Joergensen      | e-mail : tore@lis.pitt.edu                |
| Centre Court Villa      | web    : http://www.pitt.edu/~tojst1      |
| 5535 Centre Avenue # 6  |                                           |
| Pgh, PA 15232, USA      | Norwegian MSIS-student at Univ. of Pgh.   |
+-------------------------+-------------------------------------------+

Re: GNAT 3.03 for OS/2

[Dale Pontius]

In article <4j8t6p$ae2@usenet.srv.cis.pitt.edu>, tojst1+@pitt.edu (Tore Joergensen) writes:
>
> I am happy for the free compiler I get. I would not be suprised if the number
> of paying customers are below ten (I wouldn't even be suprised if it is zero).
> It is quite obvious that the OS/2 port require more fiddling than Unix ports
> (even emx can't change that).
>
Up until now, I've been quite happy, too. I've had a little off-line
discussion with Bob Dewar on this, and the picture is not quite what
we had thought. Without his OK, I won't release the number of paying
OS/2 users of GNAT, but suffice it to say that GNAT gets OS/2 ports
largely out of Mr. Dewar's own personal interest. This is not meant
to accuse him of 'being fickle', or anything else, but circumstances
do change, and his personal interest may not be sufficient to keep
OS/2 revisions of GNAT coming in the long term.

Therefore, if folks are really interested in OS/2 GNAT, we should get
either Paying Customers, or someone willing to assist with the OS/2
porting effort. Some of the GNAT ports are currently maintained by
volunteers, so there is precedent for this.

As for myself, I've found no problems so far with GNAT. All of my
problems have been learning the language with Lovelace, examples,
and an old pre-95 copy of Grady Booch. Still, at the appropriate
time, I plan to look into becoming a Paying Customer, if possible.

Also, Three Cheers to Mr. Dewar for keeping us running so far!

Dale Pontius
(NOT speaking for IBM)

Re: GNAT 3.03 for OS/2

[Tore Joergensen]

Dale Pontius (pontius@twonky.btv.ibm.com) wrote:
: Therefore, if folks are really interested in OS/2 GNAT, we should get
: either Paying Customers, or someone willing to assist with the OS/2
: porting effort. Some of the GNAT ports are currently maintained by
: volunteers, so there is precedent for this.

If there aren't any military projects or something like that made on
OS/2, I think it will be difficult to find people that will pay $1200
each year for support. I'm sure the support is fast and good, but at
least I can't afford to use that kind of money on a programming language.
Professional programmers that make general purpose programs are likely
to choose another language, since there aren't many nice libraries for
GNAT on OS/2. Sure Ada and GNAT are nice, but I think most people that
make programs for the non-critical market will prioritize simple GUI
and database connections, rather than the strong typing in Ada. If there
was a low-price alternative for non-professional programmers ($50 - 100),
it would be possible to get some paying customers. The priority for these
customers would of course be somewhere between non-paying and full-paying
customers. It's just that I think that even with a low price like this,
it would be difficult to get more than 10-30 customers. That wouldn't be
enough to make OS/2 ports a high priority, but maybe it is enough together
with Dewar's interest in OS/2 to keep the OS/2 ports coming?

It would be nice if somebody volunteered to do the OS/2 ports, but I think
that that job require some knowledge about GNAT, gcc and emx. I know that
I don't have the time to do something like that for the moment (partly
because I would have to spend time to get that knowledge), and I doubt that
I will get time to do it in the future (of course it is a question about
prioritzing). If anybody have a good idea about how to ensure continued
efforts to keep the OS/2 ports coming, speak out!

: As for myself, I've found no problems so far with GNAT.

I haven't found any problems with the last version either, but it would
be nice with a version that does some more optimizing. :-)

: Also, Three Cheers to Mr. Dewar for keeping us running so far!

 \       \       \
  \___    \___    \___
  / | \   / | \   / | \
 /  |  \ /  |  \ /  |  \ ...what... cheers... oh... I thought you said
                         chairs... well, three cheers for Dewar!!! :-)
-- 
+-------------------------+-------------------------------------------+
| Tore B. Joergensen      | e-mail : tore@lis.pitt.edu                |
| Centre Court Villa      | web    : http://www.pitt.edu/~tojst1      |
| 5535 Centre Avenue # 6  |                                           |
| Pgh, PA 15232, USA      | Norwegian MSIS-student at Univ. of Pgh.   |
+-------------------------+-------------------------------------------+

Re: GNAT 3.03 for OS/2

[Ted Dennison]

Tore Joergensen wrote:
> 
> Dale Pontius (pontius@twonky.btv.ibm.com) wrote:
> : Therefore, if folks are really interested in OS/2 GNAT, we should get
> : either Paying Customers, or someone willing to assist with the OS/2
> : porting effort. Some of the GNAT ports are currently maintained by
> : volunteers, so there is precedent for this.
> 
> If there aren't any military projects or something like that made on
> OS/2, I think it will be difficult to find people that will pay $1200
> each year for support. I'm sure the support is fast and good, but at
> least I can't afford to use that kind of money on a programming language.
> Professional programmers that make general purpose programs are likely
> to choose another language, since there aren't many nice libraries for


$1,200 a year is about what support for most commercial development 
environments runs. An example from a study we did about 1.5 years 
ago showed the following support prices:
 For Ada compilers
   VADSSELF HP 9000     $1,356 /year/user
   VADSSELF Windows NT  $1,000 /year/user
   AdaWorld HP 9000     $1,600 /year 1 user (price per user goes down
                                             as more users are added)
   AdaWorld NT          $600 /year/user

 For GUI builders:
   Altia                $1900 /year
   XDesigner            $500  /year/user
   InTouch              $1000 /year

Support for Microsoft C++ is even more expensive, but it can also be 
had for $100 per phone call. Typically in the commercial world, any
development environment will cost you AT LEAST $600, and you still 
have to pay a grand for support (without which, updates cost extra 
too). Do you really find that preferable?

-- 
T.E.D.          
                |  Work - mailto:dennison@escmail.orl.mmc.com  |
                |  Home - mailto:dennison@iag.net              |
                |  URL  - http://www.iag.net/~dennison         |

Re: GNAT 3.03 for OS/2

[Robert Dewar]

"Up until now, I've been quite happy, too. I've had a little off-line
discussion with Bob Dewar on this, and the picture is not quite what
we had thought. Without his OK, I won't release the number of paying
OS/2 users of GNAT, but suffice it to say that GNAT gets OS/2 ports
largely out of Mr. Dewar's own personal interest. This is not meant
to accuse him of 'being fickle', or anything else, but circumstances
do change, and his personal interest may not be sufficient to keep
OS/2 revisions of GNAT coming in the long term."

Well don't worry too much. I am a BIG fan of OS/2, and I don't pan
to abandon it in the forseeable future as long as IBM keeps up
a reasonable level of support.

Just be patient, I think we wlil have a 3.04 of OS/2 out soon!

Re: GNAT 3.03 for OS/2

[John Norcross {83602}]

pontius@twonky.btv.ibm.com (Dale Pontius) writes:


>In article <4j8t6p$ae2@usenet.srv.cis.pitt.edu>, tojst1+@pitt.edu (Tore Joergensen) writes:
>>
>> I am happy for the free compiler I get. I would not be suprised if the number
>> of paying customers are below ten (I wouldn't even be suprised if it is zero).
>> It is quite obvious that the OS/2 port require more fiddling than Unix ports
>> (even emx can't change that).
>>
>Up until now, I've been quite happy, too. I've had a little off-line
>discussion with Bob Dewar on this, and the picture is not quite what
>we had thought. Without his OK, I won't release the number of paying
>OS/2 users of GNAT, but suffice it to say that GNAT gets OS/2 ports
>largely out of Mr. Dewar's own personal interest. This is not meant
>to accuse him of 'being fickle', or anything else, but circumstances
>do change, and his personal interest may not be sufficient to keep
>OS/2 revisions of GNAT coming in the long term.

>Therefore, if folks are really interested in OS/2 GNAT, we should get
>either Paying Customers, or someone willing to assist with the OS/2
>porting effort. Some of the GNAT ports are currently maintained by
>volunteers, so there is precedent for this.

>As for myself, I've found no problems so far with GNAT. All of my
>problems have been learning the language with Lovelace, examples,
>and an old pre-95 copy of Grady Booch. Still, at the appropriate
>time, I plan to look into becoming a Paying Customer, if possible.

>Also, Three Cheers to Mr. Dewar for keeping us running so far!

>Dale Pontius
>(NOT speaking for IBM)


O.K. Robert,

Get your ACT together (apologies for the pun) and offer a 
"Friend's of OS/2" support license for GNAT.  Charge $25 per 
student/hobbyist/home user and give us a bottom of the barrel 
support services - (Response to problems guaranteed within 
504 hours :) :)  

I realize that this won't generate much capital, but at least it might
keep the porting effort for an OS/2 version funded.  Who knows, you
might be surprised at how many OS/2 supporters/users there are out
here!  I for one would be happy to sign up.  (At $1200 a seat,
you won't find very many takers on the home front!)

Both Ada's and OS/2's popularity seem to be gaining mind share.  
I realize you are a big fan of OS/2 - why not consider this sugestion
and help promote GNAT's continued health on OS/2?

Ada's got a date with Merlin on my system this summer :)


Regards,

John Norcross
(Not speaking for Raytheon) ;)

Re: GNAT 3.03 for OS/2

[Robert Dewar]

Sorry, there is no way we could even come out ahead on such a proposition.
$25 would not even pay the expenses of getting something like that out
of the door.

If the friends of OS/2 want to help, follow Tore's suggestion and do useful
work and make useful technical contributions to the project.

You might also work on persuading IBM to provide more energetic support
for OS/2, we have not seen an updated version for a long time now, and
the file system is still a joke (with its 2meg limit on cache).

Robert Dewar
Ada Core Technologies

Re: GNAT 3.03 for OS/2

[Ted Dennison]

John Norcross {83602} wrote:
> Get your ACT together (apologies for the pun) and offer a
> "Friend's of OS/2" support license for GNAT.  Charge $25 per
> student/hobbyist/home user and give us a bottom of the barrel
> support services - (Response to problems guaranteed within
> 504 hours :) :)

How's this any different that what we already get for free?
When I had that gnat-emx problem, my mailbox practicaly filled 
with responses from the ACT folks.

On the other hand, I'm sure ACT will gladly except donations.

-- 
T.E.D.          
                |  Work - mailto:dennison@escmail.orl.mmc.com  |
                |  Home - mailto:dennison@iag.net              |
                |  URL  - http://www.iag.net/~dennison         |

Re: GNAT 3.03 for OS/2

[Robert Dewar]

T.E.D. said

""On the other hand, I'm sure ACT will gladly accept donations.""

Well we are certainly not soliciting donations, but a couple of people have
been eager to make donations. I didn't want these just to get folded into
ACT's operating expenses, so I decided to match them from ACT funds, and
create a special prize fund to be used to award prizes to students who
make valuable contributions to the free software world in general, and
GNAT in particular.

We already have a couple of hundred dollars in this fund, so we don't need
more donations, and we will figure out how to give this away when we get
a chance (we are not exactly idle around here). I will make an appropriate
CLA announcement (and acknolwedge those who have made donations) on CLA
at a later date.

Robert Dewar
Ada Core Technologies

Re: GNAT 3.03 for OS/2

[Ted Dennison]

Robert Dewar wrote:
> 
> Well we are certainly not soliciting donations, but a couple of people have
> been eager to make donations. I didn't want these just to get folded into
> ACT's operating expenses, so I decided to match them from ACT funds, and
> create a special prize fund to be used to award prizes to students who
> make valuable contributions to the free software world in general, and
> GNAT in particular.

Hmmm. Having been in acedemia yourself, I'm sure you have some special
insight as to why it is more important to encourage students than the
general population at large, but I'm confused. Why just students?

Also, if someone were to go out and enroll in a basket weaving course
at the local community college, whould they then qualify? :-)

-- 
T.E.D.          
                |  Work - mailto:dennison@escmail.orl.mmc.com  |
                |  Home - mailto:dennison@iag.net              |
                |  URL  - http://www.iag.net/~dennison         |

Re: GNAT 3.03 for OS/2

[Robert Dewar]

"Hmmm. Having been in acedemia yourself, I'm sure you have some special
insight as to why it is more important to encourage students than the
general population at large, but I'm confused. Why just students?
"

Well Ted, the ARA prizes seem to cover the more general case, and since
GNAT was originally conceived as being of primary use to students, it
seemed appropriate to encourage them to use it :-)