comp.lang.ada
 help / color / mirror / Atom feed
From: "Randy Brukardt" <randy@rrsoftware.com>
Subject: Re: New Ada Standard
Date: Wed, 4 Apr 2007 17:44:15 -0500
Date: 2007-04-04T17:44:15-05:00	[thread overview]
Message-ID: <ev19l2$ujd$1@jacob-sparre.dk> (raw)
In-Reply-To: 1175687797.356962.209890@o5g2000hsb.googlegroups.com

<brian.b.mcguinness@lmco.com> wrote in message
news:1175687797.356962.209890@o5g2000hsb.googlegroups.com...
> On Apr 3, 7:37 pm, "Randy Brukardt" <r...@rrsoftware.com> wrote:
>
> > In any case, the best choice is to call it "Ada"; it's now the standard
and
> > should be the default on new compilers (it certainly will be on
Janus/Ada).
> > If you need to reference the standard itself, that is "Amendment 1"
(there
> > is no standard document that includes the complete language; you have to
> > merge three of them to get that, or use Ada Europe's consolidated
standard -
> > which isn't official anyway). You might need year numbers to talk about
> > obsolete Ada versions like Ada 95, but not about the current one.
> >
>
> To me, the term "Ada" could refer to any version of the language.  It is
> sometimes necessary to make it clear that you are referring to the most
> recent version of the language, and a term such as "Ada 2005" is a
> convenient way of doing so.

True enough, but you can only use that informally...

> For example, if someone wanted to inquire whether a certain compiler
> supported all of the latest language features, he might ask whether
> that compiler supported the full Ada 2005 standard.

...and by using the word "standard", you've left the realm of informal.
Indeed, there is no such thing as "the full Ada 2005 standard", and there
never will be (even ignoring the year number issue).

The current Ada standard is made up of three ISO/IEC published documents:
The International Standard "Programming Languages - Ada" (ISO/IEC
8652:1995); Technical Corrigendum 1 (ISO/IEC 8652:1995/COR.1:2001); and
Amendment 1 (ISO/IEC 8652:1995/AMD.1:2007). Note that only the first is even
called "International Standard". The others aren't "Standards" (even though
they're standardized, and go through similar approval processes as a
standard). There is no single document that you can reference, two of them
aren't Standards (although its probably OK to reference Amendment 1 as a
"standard", note the capitalization difference), and none of them are
published in 2005.

So it's OK to reference "full Ada 2005" as an informal reference to a
language; reference the Ada Europe consolidated reference manual (its cover
says "Ada 2005 Reference Manual"); [but that is most definitely not a
standard!], or reference the specific documents by their publication dates
("the Ada standards as of 2007" or "the 2007 Amendment to Ada"), but it
isn't OK to reference "standard" (singular) for Ada anymore, nor is it OK to
reference the standards documents by years other than those they are
published in. The possibility of confusion between "standard" and "Standard"
is so severe that think it is best to avoid the word altogether.

Admittedly, Ada users have a long history of being confused about what is a
standard and what is not (the RM you have on your desk or in your computer
is almost certainly not a standard, no matter which version of Ada you are
using, it just contains text fairly similar to that in the Standard [and
other documents]).

If it was up to me, I call it Ada 2007, as that doesn't lead to a confusion
between the standards and the other documents. But even that isn't ideal for
a number of reasons. But we're now in a situation that any name that you use
will be either misleading or wrong in some contexts or will be loudly
criticized by various self-appointed critics. I don't want to go there, so I
don't put any year number behind Ada unless I'm specifically talking about
old, obsolete versions.

In any case, "Ada" officially means the most recent set of standards; older
standards are "canceled and replaced". Using "Ada" when you mean only "Ada
83" is just plain wrong and is far too common (especially amongst those who
only used Ada in the 1980's). I understand the ambiguity problem, but I
prefer an combination of education (to get writers out of their bad habits)
and direct references to the actual Amendment document when absolutely
necessary.

                                          Randy.






      reply	other threads:[~2007-04-04 22:44 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-04-02 15:21 New Ada Standard Jeffrey D. Cherry
2007-04-02 16:15 ` Jeffrey R. Carter
2007-04-03  0:01   ` Randy Brukardt
2007-04-03 10:14 ` Stephen Leake
2007-04-03 23:37   ` Randy Brukardt
2007-04-04 11:56     ` brian.b.mcguinness
2007-04-04 22:44       ` Randy Brukardt [this message]
replies disabled

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox