Received-SPF: pass (google.com: domain of abnf-discuss-bounces@ietf.org designates 2001:1890:126c::1:1e as permitted sender) client-ip=2001:1890:126c::1:1e;
MIME-Version: 1.0
In-Reply-To: <50E46CC7.3080700@gmail.com>
References: <50DCD850.3020106@alum.mit.edu>
	<alpine.LSU.2.00.1301021022570.15409@hermes-1.csi.cam.ac.uk>
	<50E45623.9060607@gmail.com>
	<alpine.LSU.2.00.1301021552190.15409@hermes-1.csi.cam.ac.uk>
	<50E45B21.20509@gmail.com>
	<alpine.LSU.2.00.1301021611070.15409@hermes-1.csi.cam.ac.uk>
	<50E460D0.1020308@gmail.com>
	<CABSMSPXthVjLT1e6MgoAwK+3W6-kSimE+XJ_UnKA4ZQgm3vjqA@mail.gmail.com>
	<50E46CC7.3080700@gmail.com>
Date: Wed, 2 Jan 2013 18:44:34 +0100
Message-ID: <CABSMSPVCOBrUTPiaUm9o-eTmfE3t=OQ-Y8DZaZvtztNUPfbyNg@mail.gmail.com>
From: Riccardo Bernardini <framefritti@gmail.com>
To: abnf-discuss@ietf.org
Subject: Re: [abnf-discuss] Some thoughts about how to make ABNF more
 convenient to use
Precedence: list
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Sender: abnf-discuss-bounces@ietf.org
Errors-To: abnf-discuss-bounces@ietf.org

On Wed, Jan 2, 2013 at 6:22 PM, Dave Crocker <dcrocker@gmail.com> wrote:
>
> On 1/2/2013 9:14 AM, Riccardo Bernardini wrote:
>>
>> I speak from experience: I worked both with languages (like C) without
>> namespaces where, to avoid clashes, you must resort to the "prefix
>> trick" and languages (like Ada) where every package has its own
>> namespace and I must say that it is much easier to work with the
>> latter...
>
>
> And to play devil's advocate, I'll note that the difference in the relati=
ve
> successes of the two languages might be relevant...

Alas, yes :-)

>
> Anything like this needs to balance ease of use (non-complexity) with pow=
er.
> There must be a compelling demand for any added complexity.
>
> And any enhanced naming scheme must be downward compatible.  Old ruleset
> need to continue to work.
>

I am not seeing any special problem with this.  Currently, I saw that
when people  need definitions from other documents they (i) do c'n'p
(cut and paste) from the old document into the new one or (ii) they
say something like "Definitions for ALPHA, DIGIT, ... are the same
given in Section B.1 Core Rules of RFC 5234."  The possibility of
writing RFC5234.core.ALPHA would not invalidate the two approaches
above; actually, it would be equivalent to (ii), but with the
advantage of being more formal and prone to automatic checking.

>
>
>> BTW, inspired by Ada, you could use the RFC number as the namespace
>> itself, e.g., RFC4566.proto-version.
>
>
> Only rules from RFCs would be permitted? Not from external documents, such
> as other standards groups?

Of course! Don't you know that something is the Truth if and only if
it is written in an RFC?!? :-) (which implies that we create the Truth
and that something switch from false to true in the instant the RFC is
published, interesting... :-)

Seriously, demi-touch=E9... While writing, I had my mind turned to RFCs,
but I guess the same convention can apply to other document too, for
example, you could use ISO8652-2007.foo or ISO.8652-2007.foo to import
an ANBF definition from the Ada 2005 standard :-) :-)  Of course you
would need an algorithmic way to map the document to a fixed prefix
and, maybe, to shorten prefixes that are too long.
At the very worst, you can resort to the solutions (i) and (ii) above.

Riccardo

>
>
> d/
> --
>  Dave Crocker
>  bbiw.net
_______________________________________________
abnf-discuss mailing list
abnf-discuss@ietf.org
https://www.ietf.org/mailman/listinfo/abnf-discuss