Hi,
Thanks for the document. It reads clearly. I have some comments, questions
and nits I would like to get answers on.
A data structure which supports one or more Proof Types. This
property is conceptually similar to "alg" (1), it described
an algorithm used to maintain the verifiable data structure,
for example a binary merkle tree algorithm.
I assume "described" should be "describes", but even then there is
confusion about whether "it" describes "alg" (1) or this new VDS.
Verifiable Data Structure Parameters (VDP):
Should that be "VDSP" instead of "VDP" ? If not, lowercase the S
in Structure to indicate this?
During testing and development the experimental range SHOULD
be used, unless early assignment for a provisional entry has
been completed.
This seems to contradict the fact that Early Code Points are only
given out if there is a strong assurance that the code point won't
need wire/structure modifications. I think it is better just removed.
N/A 0 N/A N/A
Is there a reason it says N/A instead of Reserved ?
When designing new verifiable data structures, please request
the next available positive integer as your requested assignment,
for example:
Why not let the DE decide on the positive integer and have people just
request
a number without providing one? This avoids DE races too :) The DE
instructions
can include "assign the next available positive integer".
Section 4.2: is there a reason value 0 is not "N/A" here like above ?
In Section 5.2.1, I believe [I-D.ietf-cose-typ-header-parameter] and
[I-D.ietf-cose-cwt-claims-in-headers] are normative, not informative ?
Section 6 Privacy Considerations and Section 7 Security Considerations:
Why not provide direct links to the sections to not lose the reader
needing to navigate around?
I believe Acknowledgements are usually last, after IANA Considerations.
Section 9.2:
Provisional assignments to expired drafts MUST be removed from the
registry.
This sentence, along with the earlier "must pick the next number", could
lead to
issues where some implementations use a number for one thing, and other
implementations
use it for something else. "Provisional assignments", which I assume to be
Early Code
Point allocations, should be burned, not re-used. So perhaps the advise
here should be
to "MUST replace expired code point with 'reserved'" ?
Maybe 9.3.1 and 9.3.2 can be combined in a way to remove duplication ?
NITS:
about about -> about
this this -> this
Paul
_______________________________________________
COSE mailing list -- [email protected]
To unsubscribe send an email to [email protected]
