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

[Paul Kyzivat <[e-mail]>]

[This message apparently failed to be sent, so I'm trying again.]

On 7/26/24 12:36 PM, Carsten Bormann wrote:
> Hi Paul,
> 
>> There are interesting open issues around abnf dependencies:
> 
> Let me talk about the directions draft-ietf-cbor-cddl-modules <https://www.ietf.org/archive/id/draft-ietf-cbor-cddl-modules-02.html> [1] has taken on these.
> Most of these decisions should transfer to ABNF right away.
> 
> [1]: https://www.ietf.org/archive/id/draft-ietf-cbor-cddl-modules-02.html

I agree the embedding of directives into comments could also work for abnf.

>> * If I want to use a rule defined in another document, do I add to my abnf *all* of the foreign abnf that contains the rule definitions I need? Or do I only import the specific rules I need and any other rules those depend on?
> 
> Just the needed rules and dependencies.
> (We do have an “import *”.)

It seems you allow both.

>> * What to do if the imported abnf has rule names that duplicate ones I have defined in my own abnf? Should the abnf of each document live in a distinct namespace, linked only for the cross referencing rule names? Or should the writer of abnf with dependencies be responsible for avoiding naming conflicts?
> 
> Namespacing can help.
> https://www.ietf.org/archive/id/draft-ietf-cbor-cddl-modules-02.html#name-namespacing

While you embedded much in comments, minimizing impact on formal cddl 
syntax, qualifying modules by module name involves the basic syntax of 
cddl. It seems you qualify imported stuff by module name by prefixing 
names with <module-name>".". This works because "." is legal in cddl 
identifiers.

I think there must be some issues with scoping, multiple imports of the 
same module in transitive closures, etc. Most could probably be resolved 
if namespaces were defined as a native feature of the base syntax.

Doing something similar in abnf is limited by its restrictive syntax. 
The only identifier character that could act as a module name separator 
is "-". Maybe that would be sufficient, but it certainly needs further 
consideration.

Potentially abnf could be extended, either to simply allow ".' as an 
identifier character, or to add module semantics with module names 
separated by ".". I believe this would be an unambiguous extension, but 
of course it wouldn't be backward compatible.

Bottom line: it will take work to get consensus for it in abnf.

>> * How to deal with the different ways abnf is used in documents. E.g.,
>>
>> - the simple case: the document contains a single block of abnf defining the syntax for the protocol that is the subject of the document.
> 
> Easy.
> 
>> - the doc has several blocks of abnf, each defining a separate syntax. Rule names may be reused, with different syntax, in each.
> 
> The sourcecode block should have a name= attribute to help with this.
> (Eventually, this should be checked by tools/the RFC editor.)
> 
>> - the doc contains (possibly incomplete) fragments of abnf scattered through the doc text, plus a complete consolidated abnf syntax that likely duplicates some or all of the fragments.
> 
> Again, a name= help; the consolidated grammar will be what you want to reference.
> 
>> - the doc contains a complete abnf syntax that is broken into fragments interspersed with text descriptions.
> 
> No problem, if the name= is equal for all the snippets, kramdown-rfc-extract-sourcecode assembles the fragments for you (in document order, but that doesn’t make a difference for ABNF).

I agree with all of the above. But AFAIK these are all simply suggested 
conventions. Can we get them formalized and enforced during document 
review or by the editor?

I also wish these things were accessible to document readers without 
consulting the xml. And can this stuff be specified in markdown?

	Thanks,
	Paul