Hi Jim, On 09.06.26 16:32, Gould, James wrote:
PK> indeed, but it's not helpful to state that without any way forward. The WG would need to make call on this issue, I believe we would need an opinion poll of the WG. On the mailing list or in IETF 126 WG session. This issue is also expressed in "extensions" draft so it is an essential call for the WG. More on that below.We may get to agreeing to disagree on this.
RDAP is extensible, since there was no way that the working group could envision everything from the start.PK> we also agreed that extensions are not the same as core spec, therefore you can only do with extensions what the protocol framework allows you to do. A core thing about extensions is that you cannot make assumptions about clients and servers supporting them and same time break interoperability of either side not supporting it.
The extension identifier alone was agreed to not carry explicit versioning information years ago, which is what drove the work to address it in draft-ietf-regext-rdap-versioning.PK> this is true, therefore an extension which in a new version breaks behaviours of its predecessors is indistinguishable from an unrelated extension. This is what "extension" draft tells currently and this is the only interoperable interpretation.
PK> only if it is aimed to carry information about incompatible version of the same extension with the same extension identifier. draft-ietf-regext-rdap-versioning has an absolutely useful utility of signalling of back-compatible versions. I strongly believe that it would be in the best interest of extension developers to aim for it, as opposed to introduce breaking changes with each version.Now you’re arguing that an extension, such as draft-ietf-regext-rdap-versioning , that provides versioning information and that supports passing the versioning information is mandatory and needs to be baked into the core RDAP RFCs.
PK> the interoperability of the clients not supporting it is for me out of a greater good. Forcing clients to support draft-ietf-regext-rdap-versioning by breaking interoperability is no-go in my eyes.There is nothing forcing a client to implement the versioning extension other than wanting to receive real versioning information to help in interoperability. At a minimum the versioning extension will help in troubleshooting interoperability issues that may arise for clients that have a dependency to a version of an extension and choose not to implement the versioning extension.
Kind Regards, Pawel
smime.p7s
Description: S/MIME Cryptographic Signature
_______________________________________________ regext mailing list -- [email protected] To unsubscribe send an email to [email protected]
