Re: issue-base-param

I'll make a test case for xml:base, not dissimilar to the one in the 
e-mail. We seem to all agree as to the correct reading.

The html base issue probably does need some clarification, since, as in 
earlier comment, the scope of the rel="transformation" rule can be 
(mis)read as limited to valid XHTML. Since xml:base does not appear in 
the DTD, it is not permitted. Also xml:base and the <base> element 
duplicate functionality.

Test html doc something like

<!-- NO DTD -->
<html xmlns="http://www.w3.org/1999/xhtml"
       xml:base="http://www.example.com/xmlbase/" >
   <head profile="http://www.w3.org/2003/g/dataview">
     <base>http://www.example.org/htmlbase/</base>
     <link rel="transformation" href="xform" />
   </head>
   <body/>
</html>

Is xform http://www.example.com/xmlbase/xform
or http://www.example.org/htmlbase/xform.
I think HTML is clear that it is the latter.

I'll read the specs (again!).

More later, with suggested text on both of these.

Jeremy



Dan Connolly wrote:
> On Thu, 2007-03-29 at 17:14 -0400, Chimezie Ogbuji wrote:
>> On Thu, 29 Mar 2007, Dan Connolly wrote:
> [...]
>> They are answered by the chain of normative dependencies.
>>
>>> The base IRI of E, in this case, is http://www.example.org/ .
>>> Or is it... XPath doesn't cite xml base... crud... don't
>>> tell me we need a normative dependency on the XQuery data model...
>> No need at all (XML dependency chain):
>>
>> * http://www.w3.org/TR/xpath
>> * http://www.w3.org/TR/xml-infoset
>> * http://www.w3.org/TR/xmlbase/
> 
> Huh? xpath doesn't cite infoset. Oops... yes, it does,
> but only in "B XML Information Set Mapping (Non-Normative)".
> 
>>> I added a todo... the rule box to think about this.
>>> http://www.w3.org/2004/01/rdxh/spec
>>> Revision 1.241  2007/03/29 19:51:01  connolly
>>> considering citing Infoset spec re xml:base
>> You don't need to, we already have a normative dependency on XPath 1.0
> 
> I'm not sure that's enough. I'm standing by for further advice.
> 
>>> I mostly think this is editorial, i.e. it doesn't merit re-opening
>>> issue-base-param. Harry, you might want to give that a think.
>> It certainly does not merit re-opening the issue, IMHO.
> 
> 

-- 
Hewlett-Packard Limited
registered Office: Cain Road, Bracknell, Berks RG12 1HN
Registered No: 690597 England

Received on Friday, 30 March 2007 09:32:09 UTC