[Tools-discuss] Re: [abnf-discuss] Re: [art] Re: Exploring ABNF extracts from RFCs

[Paul Kyzivat <[e-mail]>]

I'm not wed to any particular solution. But I think we need to work to 
come to some consensus on one. I'm looking for something that allows 
automated verification of language source embedded in drafts and RFCs, 
including cross references to other documents.

It would be *nice* if we could find an approach that works for at least 
the languages most commonly embedded in RFCs. That seems challenging 
given the variety of language syntax/semantics. Embedding some parts in 
comments is interesting, but of course the comment syntax varies.

ISTM the namespace issues are more challenging. Discussing this from 
multiple language perspectives could provide some insights.

What sort of organization makes sense to work on this? A WG? Task force?

	Thanks,
	Paul

On 7/29/24 10:59 AM, Dave Crocker wrote:
> On 7/29/2024 7:54 AM, Carsten Bormann wrote:
>> On 2024-07-29, at 16:47, Dave Crocker <[e-mail]> wrote:
>>> As for namespaces, something like organization-document probably 
>>> suffice to take care of the global uniqueness part.  So, for example, 
>>> ietf-rfc-nnn-…
>> Generally, it is advantageous to separate the document reference 
>> (which may need to carry a lot of information to properly select the 
>> document, its revision, provenance, ...) from how the referenced item 
>> is used in the actual specification.
>>
>> So in CDDL, we do something like
>>
>>     ;# import cose.label, cose.values from rfc9052 as cose
>>
>> Which makes cose.label (cose--label when using ABNF name characters 
>> only) the rule name to be used in the CDDL (ABNF) and keeps the 
>> detailed document reference (in this example rfc9052) in the import 
>> statement, which acts as the glue.
> 
> Unless I'm misreading, I think we are agreeing on the major point I was 
> making, although differing on what I think is an important detail.
> 
> Your example uses rfc9052 as the external reference.  Mine is meant to 
> allow more variation in source and indentification of publication.  An I 
> meant the ellipsis in my example to refer to lower-level detail withn 
> the document.
> 
> Your 'as' construct is interesting.  I hadn't thought about mapping 
> between rule names.  fwiw, it makes me uncomfortable, though I 
> understand the motivation and possibly even the need...
> 
> d/
> 
>