Russ (and friends),
I have two comments to your proposal:
1) I believe that Stephen Hanson's comments should be addressed. In an
earlier message exchange regarding countersignatures, the recommendation was
to allow a countersigner to calculate the contersignature without requiring
validation of the original signature. Paul Hoffman recommended that the
following text be added to CMS (except that I corrected his last sentence
based on Bill Ottaway's comment):
A countersignature can be created without the countersigner knowing the
original content. The recipient who is validating the countersignature has
no way of knowing if the signature that was countersigned is valid without
checking that signature as well. Thus, a recipient who can validate a
countersignature but cannot validate the original signature must not infer
that the content that was signed has not been modified, and must not infer
that the countersigner actually had access to the content.
2) Your proposal states that there can be multiple countersignature
attributes in a signerInfo UnsignedAttributes SET OF Attributes. I am not
strongly opposed to your proposal, but the countersignature would be the
only attribute type for which multiple instances would be permitted. This
may be confusing to the implementors since all of the other attribute types
are limited to 0 or 1 instances. I believe that it would be beneficial to
state a rule limiting the number of countersignature attributes to 0 or 1.
This would be consistent with the processing rules for all of the other
attributes.
- John Pawling