W3C home > Mailing lists > Public > public-webcrypto@w3.org > August 2012

New Editor's Draft Published

From: Ryan Sleevi <sleevi@google.com>
Date: Fri, 31 Aug 2012 13:02:53 -0700
Message-ID: <CACvaWvZ8w_RSbsMrYJs61sh_B4k6g7s6QoLFhg_3JWJCvQAQ4w@mail.gmail.com>
To: public-webcrypto@w3.org
Version 1.44 has been pushed out now, a substantial update from the
previous 1.21

Changes include:

- Update acknowledgements and mark the key format as a readonly attribute
- Define NamedCurve as an enumeration type. Rename the NIST curves to
the more commonly used "P-256", "P-384", and "P-521".
- Add the ECPoint type for EC public keys.
- Rename EcdsaKeyGenParams to EcKeyGenParams so that it can also be used
for generating ECDH key pairs.
- Fix an error in the length in bytes of an ECDSA signature.
- Specify the ECDH algorithm (incomplete).
- Add X9.63 as a normative reference and NIST SP 800-56A as an
informative reference for ECDH.
- Clarify that the output of RSA/ECDSA/DH key generation is a key pair,
while the output of AES key generation, DH key derivation, and PBKDF2 key
derivation are Key objects.
- Add a type enum to the Key object, to disambiguate key pairs and
asymmetric keys
- Add further clarification/linkification to "Terminate the/this algorithm"
- Fix typos in the BigInteger and RSA-PSS sections.
- Add the NamedCurve typedef.
- Add missing periods at the end of sentences.
- Specify the ECDSA algorithm.
- Add X9.62 as a normative reference for ECDSA.
- Add Virginie's proposed modifications to the Scope section
- Add terminology to clarify "Terminate the algorithm"
- Add text from Mark from ACTION-38, including proposed clarifications
- Add references to ISSUE-25, ISSUE-35, and ISSUE-36
- Move key tainting from the security considerations text to the editor's
note, due to concerns about how it might be implemented inconsistently by
user agents.
- Rename "userAttributes" to "extra", to be clear that some attributes may
be predefined, as part of the KeyAttributes interface (eg: pre-provisioned
keys)
- Clarify that window.crypto.keys does not require the user agent to block
nor to guarantee that the underlying keying material is available.
- Resolve ISSUE-13 by adding a use case for JOSE
- Add more normative text regarding KeyAttributes and the possible values
for the indexed and named attributes.
- Moved BigInteger out of Algorithms definition
- Move the Cloud Storage use case right below the Protected Document
Exchange use case. Point out the similarity between them.
- Add Data Integrity Protection as a use case for signature verification.
The text is contributed by Vijay Bharadwaj.
- Define the BigInteger typedef.
- Add PKCS #3 as a normative reference.
- Added text referencing ISSUE-37, related to the create* naming scheme
- Move the CryptoOperation internal state documentation to immediately
follow the WebIDL, and remove a reference to "readyState" in favour of
linking to the internal state.
- Rename key methods to follow the create* pattern
- Add notes for ISSUE-28, regarding the (Algorithm or DOMString) typedef
for shortnames
- Add a note to ISSUE-26, noting the outstanding question regarding
multi-origin key access.
- Remove Arun from editors, per ddahl
- In KeyAttributes, change "key/value pairs" to "name/value pairs".
- Add an editorial note on the current browser support of
getRandomValues and clarify the length of 'array' means the 'byteLength' of
'array'.
- ACTION-37: Rename SHA-2-* to just SHA-*. Update reference to FIPS 180-4.

If I've missed any feedback, apologies - there was a lot of spirited and
helpful discussion following our last conference call. Please be sure to
file ISSUEs or ACTIONs (since we're not currently using Bugzilla) if there
has been anything omitted.
Received on Friday, 31 August 2012 20:03:22 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 31 August 2012 20:03:23 GMT