docs(api): clarify graphic data validation (Issue #75) - #79
Conversation
ficosta
left a comment
There was a problem hiding this comment.
I went through this against the Working Group conclusion recorded in #75 and it
lines up. The issue asks three things: whether the response should be a 500 or a 550,
whether type mismatches deserve a dedicated response body, and whether this is a hard
fail condition. The added text declines all three, which matches the conclusion that
"it's up to the implementers to decide on how to handle this case".
One question, asked genuinely rather than as a suggested change.
The first bullet, "The Server MAY validate params.data sent to load() or
updateAction() against the Graphic Manifest's schema field", is not part of the
recorded conclusion itself, and in the issue thread @didikunz argued the opposite
case: that the schema exists so a client can present appropriate controls, and that
servers should not enforce it a second time. A permissive MAY is consistent with
leaving the decision to implementers, so it does not contradict anything.
Is the intent to explicitly grant that permission, or would the section rather stay
silent on whether the Server validates at all, and only address what happens when
data does not conform? The two readings come apart for anyone writing a conformance
test later, which is why I ask.
Clarifies that the OGraf Server may validate params.data against the Graphic Manifest's schema field.
Closes #75.