Open rfc2822 opened 2 days ago
ArnyminerZ
commented
23 hours ago Well, technically it can be whatever unique URI we choose. If we are ever going to have a hostname for push, I think we could use it. Otherwise, we can directly use the repo URL, or even host some kind of schema on Github pages or similar and use it, even though it's not necessary by any means.
If we need one right now, I'd go for the repo URL.
rfc2822
commented
21 hours ago Hm a good idea. https://www.rfc-editor.org/rfc/rfc3470#section-4.9 says:
Protocol definitions should use existing XML namespaces where appropriate.
By the way, can we just put it into DAV:, as for instance RFC 6578 does? But it feels a bit wrong for me …
When a new namespace is needed, the "namespace name" is a URI that is used to identify the namespace; it's also useful for that URI to point to a description of the namespace. Typically (and recommended practice in W3C) is to assign namespace names using persistent http URIs.
The repo URL would be unique and also point to the description (because the repo contains or links to the spec).
The disadvantage is that the fate of the namespace is then bound to the bitfireAT Github organization.
RFC 4918 says:
Creation of identifiers in the "DAV:" namespace is controlled by the IETF. Note that defining new URI schemes for XML namespaces is now discouraged. "DAV:" was defined before standard best practices emerged.
So I think DAV:Push is a bad choice. The question is what the best practices are :)
https://www.w3.org/TR/xml-names11/ talks about IRIs and URNs.
What other specs use:
Choose appropriate XML NS
Anyone knows what would be appropriate?