I am the ART area reviewer for this document. This review is intended primarily
for the ART AD's, and should be treated as any other last-call feedback by
everyone else. There is nothing actionable here.
This specification uses CDDL, CBOR, CWT, etc. Assuming it uses those properly
-- and according to the shepherd writeup[1] it has been reviewed by those
groups -- then most of the ART issues are resolved "automatically."
It might be that I am not familiar enough with the field to understand all the
aspects, but I think this document requires a very careful copy-edit and
proofread. For example, Sec 2 talks says it is defining "one possible internal
representation"; are others expected? Would the document be more clear if it
instead talked solely about the items/factors/inputs to the evaluation? Why
is an internal model described at all?
Another thing is that this could have been split into two documents, with
sections 8 and following in a separate document: one describes the data, the
second describes the evaluation of the data.
Were I on the IESG, I would vote "no objection" if I trusted the shepherd doc,
WG chairs, document authors, and the Linux Foundation are right in their
conclusion: this is needed. I'm skeptical.
I am the ART area reviewer for this document. This review is intended primarily
for the ART AD's, and should be treated as any other last-call feedback by
everyone else. There is nothing actionable here.
This specification uses CDDL, CBOR, CWT, etc. Assuming it uses those properly
-- and according to the shepherd writeup[1] it has been reviewed by those
groups -- then most of the ART issues are resolved "automatically."
It might be that I am not familiar enough with the field to understand all the
aspects, but I think this document requires a very careful copy-edit and
proofread. For example, Sec 2 talks says it is defining "one possible internal
representation"; are others expected? Would the document be more clear if it
instead talked solely about the items/factors/inputs to the evaluation? Why
is an internal model described at all?
Another thing is that this could have been split into two documents, with
sections 8 and following in a separate document: one describes the data, the
second describes the evaluation of the data.
Were I on the IESG, I would vote "no objection" if I trusted the shepherd doc,
WG chairs, document authors, and the Linux Foundation are right in their
conclusion: this is needed. I'm skeptical.