Fix normative vs informative references

[handrews]

Addresses #471. Note that no changes were needed in the core spec, and the commented-out RFC reference that was left in hyper-schema will be addressed by the next PR.

I left the HTTP RFCs (including PATCH) as informative because there is guidance on how they are used but I don't think anything rises to the level of a testable requirement, even an optional one. Implementations, for instance, are not obligated to look for "accept-patch" in "targetHints", and if they do, the general directive that the meta-data is treated according to whatever protocol is in use is sufficient. Additional information is present only to reduce confusion.

The one way in which this might not be correct is the requirement that API implementors MUST NOT define POST semantics for a collection other than collection item creation semantics. That hinges on the use of POST. But it's not actually a requirement for Hyper-Schema implementations, it's a requirement for hyper-schema users. I'm not sure how that's supposed to work (or if that's improperly written to start with).

[awwright]

This looks good at a glance

…

On Nov 16, 2017 19:52, "Henry Andrews" ***@***.***> wrote: Addresses #471 <#471>. Note that no changes were needed in the core spec, and the commented-out RFC reference that was left in hyper-schema will be addressed by the next PR. I left the HTTP RFCs (including PATCH) as informative because there is guidance on how they are used but I don't think anything rises to the level of a testable requirement, even an optional one. Implementations, for instance, are not obligated to look for "accept-patch" in "targetHints", and if they do, the general directive that the meta-data is treated according to whatever protocol is in use is sufficient. Additional information is present only to reduce confusion. The one way in which this might not be correct is the requirement that API implementors MUST NOT define POST semantics for a collection other than collection item creation semantics. That hinges on the use of POST. But it's not actually a requirement for Hyper-Schema implementations, it's a requirement for hyper-schema users. I'm not sure how that's supposed to work (or if that's improperly written to start with). ------------------------------ You can view, comment on, or merge this pull request online at: #496 Commit Summary - Validation spec normative reference audit - Hyper-Schema spec normative reference audit File Changes - *M* jsonschema-hyperschema.xml <https://github.com/json-schema-org/json-schema-spec/pull/496/files#diff-0> (8) - *M* jsonschema-validation.xml <https://github.com/json-schema-org/json-schema-spec/pull/496/files#diff-1> (30) Patch Links: - https://github.com/json-schema-org/json-schema-spec/pull/496.patch - https://github.com/json-schema-org/json-schema-spec/pull/496.diff — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#496>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAatDdwEUi8bO1li7X0QbOhIH0BNvEwWks5s3PTbgaJpZM4Qhdlr> .