Re: Initial feedback on Auto Updates document

Hi Michael,
Thanks for your comments regarding the auto updates document. I'll be
rewriting it over the next few weeks and will take your feedback into
account. It would be good to get your comments once the document is a
bit more mature.

Kind regards,

On Wed, May 7, 2008 at 12:10 AM, Michael Davey <> wrote:
> I have some comments on the "Auto Updates 1.0" W3 Editor's draft,
> I understand that this is not in a public draft state yet but would
> like to provide some early feedback on the document.
> 1.  Overall the doucment shows a lot of promise. I would very much
> like to see a document along these lines but some thought needs to be
> given as to how one can provide multiple URLs for a resource,
> particularly when the resource is available via multiple, different,
> transport protocols.
> 2.  I'd like to suggest that section Introduction.6 is replaced with
> the following text:
> If the update details contained a hash element with a valid hash
> value, then the UA can derive the hash value for the updated resource
> using the method specified by the hash element's type attribute
> (possible types are "SHA-1" or "MD5"). If the attribute is missing,
> the following table is used to determine the method:
> value is 160 bits long: method is "SHA-1"
> value is 128 bits long: method is "MD5"
> any other length: the element is in error.  Implicitly this means the
> entire document is in error (as the specification is currently
> defined)
> If the method has been determined, the UA then performs a hash check.
> If the values don't match, the UA may ask the the end-user how to
> proceed (ie. discard the update resource or try again?)
> 3.  Capitalisation of element names should be consistent with other
> Widgets documents.  Some documents currently seem to use lowercase
> while others use UpperCamelCase or Initialcase.  As mentioned in a
> previous mail, the name and content of the method/type attribute of
> the hash element should be consistent with the Algorithm attribute of
> the DigestMethod and SignatureMethod elements from the Widgits-DigSig
> specification.
> Regards,
> --
> Michael

Marcos Caceres

Received on Sunday, 18 May 2008 21:33:26 UTC