XHTML object "type" attribute description inaccurate

On Fri, 23 Jul 2004 00:21:04 +0900 (JST), Masayasu Ishikawa
<mimasa@w3.org> wrote:
> 
> Hello,
> 
> After long editorial work, finally the sixth public Working Draft of
> XHTML 2.0 is now available at:
> 
>     http://www.w3.org/TR/2004/WD-xhtml2-20040722
> 
> The HTML Working Group thanks the great many people on this list for
> lots of thoughtful feedback.  We are still actively working on
> resolving remaining issues, and appreciate your feedback.
> Public discussion may take place here, but if you have an issue,
> please make sure to send your issue to www-html-editor@w3.org so
> that your issue can be properly regisitered as an issue.

There seem to be some lingering inaccuracies in the description of the
"object" element's "type" attribute. A table in 21.1.3 includes the
description:

  Defining the mime type of the object data will assist the object handler in
  determining whether the object data can be processed by the user agent
  or if an external application needs to be launched to process the object
  data.

There's really no way the UA could make such a determination until it
actually initiated the transfer, at which time it has access to the
Content-Type sent from the server, some equivalent protocol-dependent
metadata, or by some heuristic applied to identify the incoming data
(in the absence of such metadata). *All* the "type" attribute can be
used for is to indicate to the client whether it should bother
downloading the data at all. To suggest otherwise is inconsistent with
HTTP/1.1 (which XHTML 2.0 normatively references, making the spec
inconsistent with itself) as well as the principle given in 3.4 of the
July 5 working draft of "Architecture of the World Wide Web":

  Principle:  Data-metadata inconsistency:
  Agents MUST NOT ignore message metadata without the consent of the user.

  Thus, for example, if the parties responsible for "weather.example.com"
  mistakenly label the satellite photo of Oaxaca as "image/gif" instead of
  "image/jpeg", and if Nadia's browser detects a problem, Nadia's browser
  must not ignore the problem (e.g., by simply rendering the JPEG image)
  without Nadia's consent. Nadia's browser can notify Nadia of the problem
  or notify Nadia and take corrective action.

-- 
Braden McDaniel
<http://www.endoframe.com>

Received on Thursday, 22 July 2004 14:18:55 UTC