On 14-Jul-20 02:58, Paul Kyzivat wrote:
> Brian,
>
> On 7/12/20 6:40 PM, Brian Carpenter via Datatracker wrote:
>
>> Nits:
>> -----
>>
>>>> 4.1. MSRP URI
>> ....
>>>> transport /= "dc"
>>
>> I see that RFC7977 takes a slightly different approach to updating the ABNF:
>>
>>>> transport = "tcp" / "ws" / 1*ALPHANUM
>>
>> The advantage of listing out
>>
>> transport = "tcp" / "ws" / "dc" / 1*ALPHANUM
>>
>> would be that the reader sees the full list.
>
> While it might be nice to see the full list, it can't be counted on to
> remain the full list. If each update did this, then every draft that
> updates the list should formally Update every other document that
> updates the list.
>
> The *point* of /= in abnf is that it decouples the particular addition
> from all others.
Sure, I realise that. But...
> This sort of thing should really have an IANA registry. But often the
> expectation is that changes aren't made often enough to justify that.
Indeed, that occurred to me as an alternative approach but a registry
with 3 entries is a bit light (and also, readers starting at RFC4975
wouldn't know to look for the registry).
> Lacking that, IMO /= is the preferred way to go.
It was only a nit...
Brian
>
> Just kibitzing,
> Paul
>
>
>
--
last-call mailing list
last-call@xxxxxxxx
https://www.ietf.org/mailman/listinfo/last-call
