Annex I?

Annex I?

[Mattias Lindblad]

Hi,

In IEEE 1003.5, even in the 1999 edition, there are some references to
Annex I of the LRM (for example in 3.3.17.2). I have noticed that there
is no such annex in the current revisions of the LRM, but the
references seem to map directly onto Annex J instead.

I'm curious about why this annex has changed its name from I to J. I
can't find any references to this event. My best guess is that "I" was
to easily confused with "1". But I am not able to find a copy of the
standard where Annex I exists, so I can't see if there are any
differences in the content that would explain the name change.

Is there anyone who knows the true story about this?

//Mattias

Re: Annex I?

[Robert A Duff]

"Mattias Lindblad" <ml_news@matli.net> writes:

> Hi,
> 
> In IEEE 1003.5, even in the 1999 edition, there are some references to
> Annex I of the LRM (for example in 3.3.17.2).

What's IEEE 1003.5?

>... I have noticed that there
> is no such annex in the current revisions of the LRM, but the
> references seem to map directly onto Annex J instead.
> 
> I'm curious about why this annex has changed its name from I to J. I
> can't find any references to this event. My best guess is that "I" was
> to easily confused with "1". But I am not able to find a copy of the
> standard where Annex I exists, so I can't see if there are any
> differences in the content that would explain the name change.
> 
> Is there anyone who knows the true story about this?

As I recall, the ISO standard for standards forbids Annexes I and O,
probably for the reason you guessed.  There might have been a
preliminary version with an Annex I, which would have been produced
before I had read the ISO standard for standards and obeyed all the
nitpicking rules and regulations -- I don't remember.  But if so, nobody
should be referring to such a preliminary, unapproved version.

The Ada RM also has Annexes N and P, but not O.

The ISO rules also forbid paragraph numbers.  So we put paragraph
numbers in the version that most folks look at -- but if you buy it from
ISO, it will be sans paragraph numbers.

- Bob

Re: Annex I?

[Mattias Lindblad]

Thanks for a fast and authoritative answer!

> What's IEEE 1003.5?

"POSIX Ada Language Interfaces- Part 1: Binding for System
Application Program Interface (API)", i.e. the bindings between
Ada and POSIX.

> There might have been a preliminary version with an Annex I [...]
> But if so, nobody should be referring to such a preliminary,
> unapproved version.

I agree. The IEEE standard refers to a version of the Ada standard
issued 15 February 1995. However, the first edition of the IEEE
standard was issued 1996, so one could guess that they started with
a preliminary version of the Ada standard while writing the standard
and didn't notice the change in annex numbering in the final version.

> The ISO rules also forbid paragraph numbers.

Isn't it possible to get some kind of exemption from those rules? My C
and C++ standards (from 1999) both have paragraph numbering, and I
believe they are bought from ISO. Or maybe those standards are simply
not standard compliant. (In addition to the paragraph numbering,
the C standard actually has an "Annex I".)

//Mattias

Re: Annex I?

[Robert A Duff]

"Mattias Lindblad" <ml_news@matli.net> writes:

> Thanks for a fast and authoritative answer!

You're welcome.

> > What's IEEE 1003.5?
> 
> "POSIX Ada Language Interfaces- Part 1: Binding for System
> Application Program Interface (API)", i.e. the bindings between
> Ada and POSIX.

Ah, yes.

> > There might have been a preliminary version with an Annex I [...]
> > But if so, nobody should be referring to such a preliminary,
> > unapproved version.
> 
> I agree. The IEEE standard refers to a version of the Ada standard
> issued 15 February 1995. However, the first edition of the IEEE
> standard was issued 1996, so one could guess that they started with
> a preliminary version of the Ada standard while writing the standard
> and didn't notice the change in annex numbering in the final version.

Could be.

I actually printed out the final version of the Ada 95 standard in early
December, 1994.  This is the version that was given to ISO.  It took a
couple-or-few months to get through the ISO beaurocracy.  I'm still a
little bit annoyed that everybody calls it "Ada 95", making it look like
we were 2 years late, when we were actually only one year late.  ;-)

> > The ISO rules also forbid paragraph numbers.
> 
> Isn't it possible to get some kind of exemption from those rules? My C
> and C++ standards (from 1999) both have paragraph numbering, and I
> believe they are bought from ISO. Or maybe those standards are simply
> not standard compliant. (In addition to the paragraph numbering,
> the C standard actually has an "Annex I".)

I don't know.  The last time I read the ISO standard for standards was
in 1994.  It has probably changed since then.  In practise, the real
rule is that you have to do what some guy at ISO tells you.  I tried to
obey the standard for standards, and sent a copy to this guy, who marked
it up in red.  Then I obeyed the red marks, and *that* was good enough.
He was very concerned about the fonts and other details on the title
page and the first few pages, but he obviously didn't read the bulk of
the RM.

I think they allowed *line* numbers, as in numbering every fifth line:
5, 10, 15, 20, or something like that.  But the Ada-style paragraph
numbers are much more aesthetically pleasing.  I think we tried to get
ISO to agree, but they refused.  And the reviewers were very concerned
that the Ada 9X standard have the same look and feel as the Ada 83 one.

I believe the index contains paragraph numbers even in the official ISO
version.  We could get away with that because the index is "informative"
rather than "normative".  I've always found that terminology amusing --
in order to be "normative" one must write incomprehensible
(uninformative) gibberish.  ;-)

You should see the ugly hackery that was necessary to get the paragraph
numbers to look like the Ada 83 ones.  Perl scripts for both
preprocessing the input to Scribe, and postprocessing the output.
Randy has since switched to a completely different system.

- Bob

Re: Annex I?

[Randy Brukardt]

"Robert A Duff" <bobduff@shell01.TheWorld.com> wrote in message
news:wccoecyzfar.fsf@shell01.TheWorld.com...
...
> You should see the ugly hackery that was necessary to get the paragraph
> numbers to look like the Ada 83 ones.  Perl scripts for both
> preprocessing the input to Scribe, and postprocessing the output.
> Randy has since switched to a completely different system.

...which is an Ada program that directly generates HTML and RTF from input
very similar to the old Scribe. (I only changed it when I couldn't figure
out how to implement the original code.)

But you should see the ugly hackery that is necessary in the RTF output to
get the paragraph numbers to show up in the right place automatically. It
took me a lot of attempts to find something that would work *and* not crash
Word when it is read. (Word still has problems displaying the RTF properly,
although simply saving the RTF as a DOC file and reloading it causes it to
work just fine. Bizarre.)

The HTML doesn't suffer from *those* problems, but the paragraph numbers
disappear on some older browsers (hopefully no new ones). The more things
change, the more they stay the same, I guess.

                     Randy Brukardt.