Came up when @bob-beck and I were talking about things. Having the log hash as a separate AlgorithmIdentifier means every possible hash might be reachable from MTCs, including say MD5. It also makes it challenging should someone want to do, I dunno, specific truncation of some hash.
If we instead did id-pe-mtcCertificationAuthority-SHA256 and said that other MTC log hashes get their own OIDs, it's the same extensibility, but keeps the parameter space more controlled and makes it easier to encode funny things if that's what people want.
Not sure if we'd also id-alg-mtcProof-SHA256 to match or if that's just id-alg-mtcProof. Strictly speaking, we don't need to bifurcate that one (and currently don't).
Came up when @bob-beck and I were talking about things. Having the log hash as a separate AlgorithmIdentifier means every possible hash might be reachable from MTCs, including say MD5. It also makes it challenging should someone want to do, I dunno, specific truncation of some hash.
If we instead did
id-pe-mtcCertificationAuthority-SHA256and said that other MTC log hashes get their own OIDs, it's the same extensibility, but keeps the parameter space more controlled and makes it easier to encode funny things if that's what people want.Not sure if we'd also
id-alg-mtcProof-SHA256to match or if that's justid-alg-mtcProof. Strictly speaking, we don't need to bifurcate that one (and currently don't).