Re: http content type authoritative for object data?

On 12/8/09 8:32 AM, Julian Reschke wrote:
> 4. Otherwise, if the resource type is unknown, and the resource has
> associated Content-Type metadata, then let the resource type be the type
> specified in the resource's Content-Type metadata."
....
> This suggests that when @type is present, the associated content type
> information (for instance HTTP Content-Type) is ignored

The "associated Content-Type metadata" bit means the HTTP Content-Type.

So it's only ignored if @type is a type that is supported via a plug-in.

> which seems to be both an incompatible change, and violate the "authoritative metadata
> principle".

Sadly required to not break existing content.  :(  I don't think 
anyone's particularly happy about it; if you have a better approach that 
still doesn't break existing content I'd love to know what it is.

> Was this discussed somewhere?

Yes; looks like on the whatwg mailing list.  See 
http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-August/022501.html 
as a starting point, I think, complete with the gory bug references, 
etc.  There are both earlier and later messages in that thread, but I 
think that one sums up the Gecko implementation experience with <object>.

> It it *implemented* this way?

In Gecko, yes.  See the discussion above.

-Boris

Received on Tuesday, 8 December 2009 20:28:28 UTC