inline...

On 1/16/26 9:58 AM, Gould, James wrote:
> The other factor for EPP extension I-Ds is the need for URIs to change during 
> the maturity of the draft to encourage implementation.  This has been the 
> pattern for the latest set of EPP extensions.  Can RFC 7120 support a 
> wildcard namespace URI early allocation?  
> 
> Consider the latest EPP extension I-D draft-ietf-regext-balance that has the 
> XML namespace URI urn:ietf:params:xml:ns:epp:balance-0.1, which is close to 
> the Maturity Versioning defined for RDAP, where when the XML schema is change 
> in the I-D the namespace will be bumped and changed to finally changed to 
> urn:ietf:params:xml:ns:epp:balance-1.0 once the I-D passes WGLC.  The ABNF 
> format for the XML namespace is "urn:ietf:params:xml:ns:epp:balance-" DIGIT 
> "." DIGIT.  There is the risk of material changes occurring after WGLC, but 
> that's a risk an implementer would need to take in implementing ahead of the 
> EPP extension becoming an RFC.  

James,

The text I offered applies 7120 process to the EPP Extension registry, not the 
XML namespace registry. To avoid the problem you outline, the text I offered 
exempts Internet Drafts from the requirement to register XML namespace IDs and 
schemas.

>> On 1/15/26 2:46 PM, Hollenbeck, Scott wrote:
>>> [SAH] Before I comment on anything below, we need to discuss the
>> applicability of RFC 7120. It's focused on "Early IANA Allocation of 
>> Standards
>> Track Code Points". I want to emphasize the "Code Points" part of the title. 
>> I
>> didn't find a concise definition in 7120 that describes what a code point 
>> is, but
>> the acceptability of your replacement text hinges on whether or not an
>> Internet-Draft can be described as a code point. It seems like a stretch to 
>> me
>> since an I-D isn't a value that needs to be implemented in code where
>> interoperability depends on value agreement. It's not exactly a TCP port
>> number, for example.
>>> That's just my take. What do others think?
>> Do we really want to rathole on what is and what is not a "code point"? Is it
>> only a value of no more than one octet? Can a 1 million bit number be a code
>> point? Are OIDs code points, because they too are sequences of octets, just
>> like a URI. And OIDs are definitely used in IETF early registrations.
>>
>> The point is that 7120 describes the process for early registrations. It is 
>> well
>> used in the IETF and we would not be doing anything unusual. You stated
>> during the interim that it would be difficult to write the instructions for 
>> doing
>> early registrations, and the text on offer shows that you do not need to do
>> so... it already exists in an RFC.
> 
> [SAH] 7120 describes a process for early registration of _code points_, not 
> specifications. I agree that it could be used for early registration of URIs, 
> but I'm not convinced that it applies for registration of Internet-Draft 
> specifications.
> 

Scott,

If you want to take this line of reason to its hair-splitting end, the EPP 
extension registry is a registry of URI references to EPP extension 
specifications. The specifications themselves are not in the EPP extensions 
registry. And URIs are a sequence of octets, just like OIDs and other "code 
points".

What I have put forward is a compromise between "NO I-Ds AT ALL" and "I-Ds 
ALWAYS AND FOREVER". And, IMHO, without this compromise I believe the default 
is what you have in the current text. I am not gonna die on a hill over this. 
If the working group wants "NO I-Ds AT ALL" then so be it.

-andy, as an individual

_______________________________________________
regext mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to