RE: Issue: PROP_ATTR

> From: w3c-dist-auth-request@w3.org
> [mailto:w3c-dist-auth-request@w3.org]On Behalf Of Greg Stein
> Sent: Sunday, April 15, 2001 10:54 PM
> To: WebDAV WG
> Subject: Re: Issue: PROP_ATTR
>
>
> On Sun, Apr 15, 2001 at 11:14:33AM +0200, Julian F. Reschke wrote:
> >...
> > As I said, both point of views are possible. However, your
> position seems to
> > be:
> >
> > value = Infoset of all child elements *plus* the optional
> xml:lang attribute
>
> Yes, because the xml:lang value is scoped, and the property value occurs
> within that scope.
>
> And this isn't just because xml:lang occurs on the property name
> element; it
> could be three levels higher in scope.

I've got a concern here... RFC2518 makes a special case for xml:lang, which
I think is a bad thing. "Canonical XML" [1][2] specifies rules for
generating canonical representations of document subsets (which seems to be
relevant to our problem), and it states that *all* attributes in the xml
namespace are inherited (as of today, this would include xml:space as well).

Also note that other working groups might be adding more attributes to the
xml namespace, such as xml:base [3]. I think it should be considered to
either drop the requirement to persist xml:lang with the property value, or
to extend the requirement so that all attributes from the xml namespace [4]
are treated equally.

Julian


[1] <http://www.w3.org/TR/xml-c14n#DocSubsets>,
[2] <ftp://ftp.rfc-editor.org/in-notes/rfc3076.txt>,
[3] <http://www.w3.org/TR/2000/PR-xmlbase-20001220/>,
[4] <http://www.w3.org/XML/1998/namespace>

Received on Monday, 16 April 2001 08:25:42 UTC