Re: [Uri-review] mt: scheme documentation (was: Matter "mt" URI permanent registration review request)

[Ted Hardie <[e-mail]>]

Hi Graham,

I've snipped most of this to comment on one particular matter, covered in a
couple of sections of your response:
<snip>
On Wed, Jan 4, 2023 at 6:18 PM Graham Klyne <[e-mail]> wrote:

>
>
>> The broad intention with any IETF technical specification is that there
>> is
>> sufficient information freely available to allow interoperable
>> implementations,
>> preferably between any two agents on the World Wide Web.  So what I'm
>> trying to
>> do here is interpret the explicit and implicit requirements in a way that
>> respects that intention.
>>
>
>> I agree with this as a statement of broad intention, but my reading of
the URI registration procedure history is that it has gone from a mechanism
with a fairly strong gatekeeping approach to one primarily focused on
ensuring that collisions do not occur.  The provisional registration
procedures are the primary signal for that, but even within permanent
registrations, I feel the same is true (witness the shift from "ietf" tree
to permanent).


<snip>
>
>

> OR (2) add a technical section to the registration template covering the
> above information.  This is probably the easier option, but it goes against
> the intent of the latest URI registration procedure, which was to defer all
> technical specifics a to a defining document.


 I don't disagree with this reading of the registration procedures, but I
feel the collision avoidance issue is even larger than the deferral to
external documents presumption.  That means that I personally would
support the inclusion of the relevant technical details in the template
if that turns out to be the easiest method to provide them to the IETF
and broader technical community.  The template could also include a
pointer to the broader matter specifications for those with the interest
 in getting the additional context, perhaps with a note that at the time
 of registration a EULA-style agreement was required to retrieve them.
But the key thing is that the registry would contain the scheme along
with sufficient information to identify the change controller and the
formal syntax; that's truly the *sine qua non* for me.

best regards,

Ted Hardie