Hi Pawel,

I think we as WG have been discussing 2 separate points lately:

  1.  To what extent should a bare identifier be allowed? Never, always, under 
certain well-defined conditions (say, when only one query path, query 
parameter, JSON name, object class, and/or another extension point (HTTP 
header, etc) needs naming), or when a technical solution cannot be defined 
otherwise. The next version of the RDAP Extensions draft would present these 
options so that we can settle this matter.
  2.  Should a media type parameter, as part of an RDAP extension, also be 
prefixed? For prefixing consistency, the answer seems to be yes, per earlier 
discussion with James.

So far, from naming conflict prevention angle, we have identified query paths, 
query parameters, JSON member names, object class names, HTTP headers, and 
media type parameters as prefixing candidates.

Andy and I were discussing one more. :) Should a new relation type defined for 
a web link used in an RDAP extension also be prefixed? Looks like we should 
since the Media Types reviewers recommend greater specificity for relation 
types. Though we haven’t yet tried registering “foobar123-xyz” as a new 
relation type. ;) The closest prefixing example for regext purposes has been 
“rdap-up”, etc for the RIR search, whereas “geofeed” was allowed for the RDAP 
Geofeed from specificity angle.

Thanks,
Jasdip

From: Pawel Kowalik <[email protected]>
Date: Tuesday, June 24, 2025 at 1:42 AM
To: Jasdip Singh <[email protected]>, Gould, James <[email protected]>, 
[email protected] <[email protected]>, Hollenbeck, Scott <[email protected]>, 
[email protected] <[email protected]>
Cc: [email protected] <[email protected]>
Subject: Re: [regext] Re: On bare identifiers in Extensions draft

Hi Jasdip,
On 18.06.25 19:43, Jasdip Singh wrote:
Hi Pawel,

Totally agree with being consistent prefixing-wise for query requests and 
responses in RDAP extensions!

But afraid, we might be overlaying this prefixing concept on to the media type 
space unnecessarily. If another RDAP extension ends up adding another parameter 
to “application/rdap+json” or another media type, the IANA registration process 
shall identify any conflict with existing parameters when updating that media 
type’s specification.

Correct. Being consistent is also my point but not necessarily in a twist that 
everything shall be always prefixed. I argument, that "IANA registration 
process shall identify any conflict" also applies to all other RDAP cases, thus 
rendering bare identifiers no issue for the interoperability or conflict 
potential between extensions.
 Thanks,
Jasdip
[...]
_______________________________________________
regext mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to