Hi Elwyn,
we tried one round at your comments.
We didn’t tackle the interaction with /.well-known/core — I continue to believe
what I said yesterday, but we still have to find a good way to put this into
the document.
We added a note that we need to decide whether there is an onus on a receiving
CBOR implementation to check for the absence of the 13 strings as names
(“Adam’s second issue”).
The title* issue also requires some more thinking — do we even want to have
language tags on a constrained device. Right now, the document simply leaves
the whole subject out, but probably needs a position on that.
Finally, there is the issue that RFC 6690 percent-decodes and how that should
be reflected in the JSON and CBOR variants.
We added a whole subsection on converting back to RFC 6690, which discusses the
current pragmatic approach on solving the “to quote or not to quote” issue.
On the editorial side, we addressed the document title, added some language to
make it clear that the CDDL is informative (and actually added build tool rules
to check the examples against the CDDL, ouch), and quite a few nits. Thank you
for the text suggestions (and no, we didn’t add a reference to X.690 yet :-).
Based on Adam’s suggestions, we found a way to fix the bullet list in 2.2.
We still need to review the editorial comments on the detailed processing of
strings etc. (see may previous comments why these didn’t just drop in).
Please have a look at the new text at:
https://tools.ietf.org/rfcdiff?url2=draft-ietf-core-links-json-08.txt
Thank you!
Grüße, Carsten