ISO/IEC DIS 8652 and ISO/IEC DIS 14519-1

ISO/IEC DIS 8652 and ISO/IEC DIS 14519-1

[Erik Naggum]

INTRODUCTORY REMARKS

as a consultant to the Norwegian Standards Organization on matters of
ISO/IEC JTC 1/SC 22 (although my "home committee" is SC 18), I receive some
funny things in the mail.  in this case, it may be very serious.

voting on ISO/IEC DIS 8652, ISO-speak for "Ada9X", terminates 1994-10-30,
and it is expected that progression to IS will be expeditious.

meanwhile, IEEE standard 1003.5-1992 is being fast-tracked as ISO/IEC DIS
14519-1, on which voting began 1994-08-25 and terminates 1995-02-25.  IEEE
1003.5 is better known as POSIX.5, its full name being

    IEEE Standard for Information Technology --
    POSIX Ada Language Interfaces --
    Part 1: Binding for System Application Program Interface (API)

it was approved by the IEEE Standards Board 1992-06-18.  the "fast-track
procedure" takes an existing standard from some member body or category A
liaison and asks ISO to rubber-stamp it by vote from its member bodies.
IEEE is such a category A liaison, and IEEE 1003 is generally being
fast-tracked into various parts of ISO/IEC 9945, with Ada bindings
apparently findings its place as ISO/IEC 14519.

1003.5 is a binding to Ada 83 (ISO 8652:1987).  it is not unlikely that the
second edition of ISO 8652 (Ada9X) will be published or at least approved
for publication before voting on this standard terminates.


PROBLEM

I am in a difficult position, as I think most other SC 22 members and
consultants are, whether I shall recommend to disapprove this DIS on
grounds of impending revision of one of its base standard, or to proceed
with the rubber-stamping procedure in the hopes that a revised version will
eventually come along, and that an Ada 83 binding is more important than no
binding.  that this is an IEEE standard already diminishes the importance
of the latter point to near zero in my eyes.

what does the Ada community think?  I am not in position to appreciate the
consequences of either choice, and do not know whether this draft standard
should be progressed, and would like to avoid an embarrassing mistake of
helping to approve a standard that will be obsolete by the time its ink
dries.

I do not understand why IEEE decides to fast-track this standard now that
Ada is in the final stages of its revision, so if anybody knows this,
please let me know.

your advice is greatly appreciated.

#<Erik>
--
Microsoft is not the answer.  Microsoft is the question.  NO is the answer.

Re: ISO/IEC DIS 8652 and ISO/IEC DIS 14519-1

[David Emery]

Erik raises some interesting questions.  As a member of the POSIX/Ada
committee (and also WG9), let me answer them.  First, some history:

	1.  Because of politics relating to 'language independence',
ISO would NOT accept the submission of the POSIX/Ada binding (IEEE
Project P1003.5) or POSIX/Fortran binding (IEEE Project P1003.9) under
the same terms as the C binding/POSIX interface definition (IEEE
Project P1003.1).  
	2.  Therefore, the C binding became ISO 9945.1:1989
(relatively) concurrent with its approval by IEEE as IEEE Std
1003.1-1990.  
	3.  POSIX/Ada became IEEE Std 1003.5-1992 (and
FORTRAN IEEE Std 1003.9-1992) independent of any ISO actoin.
	4.  After approval of IEEE Std 1003.5-1992 (in 1992 :-), we
then proposed ISO JTC1 fast-track of the IEEE POSIX/Ada standard.
After some coordination and a consultative letter ballot by ISO/IEC
JTC1/SC22 which concurred (with comments), we then forwarded the
document to the JTC1 secretariat for fast-track standardization.  At
the same time, the IEEE started work on a amendment project for
POSIX/Ada, both to fix some known defects and also to handle any
comments from the ISO ballot.
	5.  Somewhere in ISO, this document got lost.  We got it back
on track this year.  (Thus, we lost a full year in the balloting
cycle.)

Now, the real technical question is: Does POSIX/Ada have value for
Ada9X?  The answer (IMHO) is "Yes".  The POSIX/Ada standard is upwards
compatable with Ada9X.  The only 9X issue is the use of interrupt
entries for signal handlers, a feature declared obsolete in Ada9X.
Other than that, the Ada83 binding is completely consistent with
Ada9x.  

Furthermore, there has been a lot of interest in this standard, and I
think it's important that we get the I.S. out ASAP.

Finally, we are currently working on the POSIX/Ada Real-Time binding.
The way POSIX standards work, this is a revision to the existing
POSIX/Ada binding.  We're explicitly considering Ada9X in this work.  

There are some places in the current POSIX/Ada (IEEE Std 1003.5-1992)
standard that I would like to 'change' to take advantage of Ada9X
features.  It's possible that we'll propose a revision of the entire
POSIX/Ada document to take better advantage of Ada9X in the future,
but for now, the POSIX/Ada binding is very usable from Ada9X.

In short, vote "yes".

			Dave Emery
			(P1003.5 Technical Editor and
			 Interpretations Vice-Chair)
--
--The preceeding opinions do not necessarily reflect the opinions of
--The MITRE Corporation or its sponsors. 
-- "A good plan violently executed -NOW- is better than a perfect plan
--  next week"                                      George Patton
-- "Any damn fool can write a plan.  It's the execution that gets you
--  all screwed up"                              James Hollingsworth
-------------------------------------------------------------------------

Re: ISO/IEC DIS 8652 and ISO/IEC DIS 14519-1

[Mats Weber]

In article <EMERY.94Sep15084810@goldfinger.mitre.org>
emery@goldfinger.mitre.org (David Emery) writes:

> Now, the real technical question is: Does POSIX/Ada have value for
> Ada9X?  The answer (IMHO) is "Yes".  The POSIX/Ada standard is upwards
> compatable with Ada9X.  The only 9X issue is the use of interrupt
> entries for signal handlers, a feature declared obsolete in Ada9X.
> Other than that, the Ada83 binding is completely consistent with
> Ada9x.  

I second that. I read the draft five years ago and made some comments
on it. By that time, I thougth that it was very nearly ready for
approval.

I see no major Ada 9X features except child library units that could
make this standard significantly better. I would prefer having this
standard approved now than having it delayed (for another 5 years ? :-)
just for a few easthetic fixes.

Besides, I think 1003.5 is, with the Ada-SQL binding, the most
important standard for Ada.

Re: ISO/IEC DIS 8652 and ISO/IEC DIS 14519-1

[Erik Naggum]

[David Emery]

|   In short, vote "yes".

thanks for many informative replies!  it was obviously right to post here.

I don't, strictly speaking, determine the vote, but will make my
recommendation in line with the unanimous recommendations I have received.

thank you to you all!  (and best of luck to ISO/IEC 14519-1!)

#<Erik>
--
Microsoft is not the answer.  Microsoft is the question.  NO is the answer.

Re: ISO/IEC DIS 8652 and ISO/IEC DIS 14519-1

[Tucker Taft]

In article <19940915.5134@naggum.no>, Erik Naggum  <erik@naggum.no> wrote:

> ...
>1003.5 is a binding to Ada 83 (ISO 8652:1987).  it is not unlikely that the
>second edition of ISO 8652 (Ada9X) will be published or at least approved
>for publication before voting on this standard terminates.

Actually, since Ada 9X is upward compatible with Ada 83,
1003.5 is also a binding to Ada 9X.

> ...
>PROBLEM
>
>I am in a difficult position, as I think most other SC 22 members and
>consultants are, whether I shall recommend to disapprove this DIS on
>grounds of impending revision of one of its base standard, or to proceed
>with the rubber-stamping procedure in the hopes that a revised version will
>eventually come along, and that an Ada 83 binding is more important than no
>binding.  that this is an IEEE standard already diminishes the importance
>of the latter point to near zero in my eyes.
>
>what does the Ada community think?  I am not in position to appreciate the
>consequences of either choice, and do not know whether this draft standard
>should be progressed, and would like to avoid an embarrassing mistake of
>helping to approve a standard that will be obsolete by the time its ink
>dries.

I believe you should support 1003.5.  Even though it was
designed with Ada 83 in mind, it is quite adequate for use
with Ada 9X, and will fill an important need that is just
as important with Ada 9X as with Ada 83, namely portable
access to Unix/Posix system calls.

>I do not understand why IEEE decides to fast-track this standard now that
>Ada is in the final stages of its revision, so if anybody knows this,
>please let me know.

There is no requirement to update every binding every time
a language is extended.  I am sure when and if new features
are added to C (e.g. the proposed numeric extensions), there 
will not be a sudden rush to revise every Posix/C binding
(or every other C binding).

In general, ISO standards that make references to other standards
implicitly refer to the latest revision of those standards --
here are words from the standard ISO "Normative References" prologue:

   "All standards are subject to revision, and parties to agreements
    based on this International Standard are encouraged to investigate
    the possibility of applying the most recent editions of the standards
    indicated below."

Presuming 1003.5 has similar wording, there should be no problem
using the Ada 9X standard in conjunction with 1003.5.
Upward compatibility allows essentially all Ada 83 bindings to
be used with Ada 9X.

>your advice is greatly appreciated.
>
>#<Erik>

S. Tucker Taft   stt@inmet.com
Ada 9X Mapping/Revision Team
Intermetrics, Inc.
Cambridge, MA  02138

Re: ISO/IEC DIS 8652 and ISO/IEC DIS 14519-1

[Erik Naggum]

[Tucker Taft]

|   In general, ISO standards that make references to other standards
|   implicitly refer to the latest revision of those standards --
|   here are words from the standard ISO "Normative References" prologue:
|   
|      "All standards are subject to revision, and parties to agreements
|       based on this International Standard are encouraged to investigate
|       the possibility of applying the most recent editions of the standards
|       indicated below."
|   
|   Presuming 1003.5 has similar wording, there should be no problem
|   using the Ada 9X standard in conjunction with 1003.5.
|   Upward compatibility allows essentially all Ada 83 bindings to
|   be used with Ada 9X.

I'll recommend a comment that such a clause should be added to the standard
before it is published, possibly updating the reference to point to the new
edition of ISO 8652 (if available in time), or to remove the edition-
specific notation for the existing reference.  these are merely editorial
changes, and should have no impact on its adoption or the speed thereof.

#<Erik>
--
Microsoft is not the answer.  Microsoft is the question.  NO is the answer.