Re: Comments on the Triple Patterns Fragments draft

> The MUST RFC2616 statement could be left out in any case. If HTTP URLs are used, a conformant server has obviously to conform to the HTTP specs.  Using the same reasoning, other normative statements in section 4 should be dropped (or at least made non-normative to prevent duplication of normative statements).

Agree. Following up on this in
https://github.com/HydraCG/Specifications/issues/63

> It is indeed and most of the things are already covered by the HTTP spec. Other things such as the
> 
>    ... a server SHOULD return a triple-based representation instead of
>    sending a response with status code 406 Not Acceptable
> 
> statement do not improve interoperability beyond what HTTP already does.

Well, it says the default should be sending a 406,
which is stronger than wat RFC7231 says.
But I could just leave it out.

> You mean Vary: Origin, right?

Err… yes.

> In the request it will be Origin: D1... Access-Control-Allow-Origin: D1 is in the response.

Yes I messed them up.

> Access-Control-Allow-Origin is not about authentication…

Oh yes, it can be:
"only Web applications on domain D1 can access this API".
It's primitive, but effective for browsers (only).

Best,

Ruben

Received on Thursday, 24 July 2014 07:34:48 UTC