xml-dev - RE: Another errata?

[Andrew Layman <andrewl@microsoft.com>]

The example 

 <html:a href="foo"> is identical to
 <html:a html:href="foo">

used in earlier mails to this list is less than ideal for the purpose of
explaining namespaces on attributes.  

Let's back up.  

The A element in HTML is defined in http://www.w3.org/TR/REC-html40/ as

<!ELEMENT A - - (%inline;)* -(A)       -- anchor -->
<!ATTLIST A
  %attrs;                              -- %coreattrs, %i18n, %events --
  charset     %Charset;      #IMPLIED  -- char encoding of linked resource
--
  type        %ContentType;  #IMPLIED  -- advisory content type --
  name        CDATA          #IMPLIED  -- named link end --
  href        %URI;          #IMPLIED  -- URI for linked resource --
  hreflang    %LanguageCode; #IMPLIED  -- language code --
  rel         %LinkTypes;    #IMPLIED  -- forward link types --
  rev         %LinkTypes;    #IMPLIED  -- reverse link types --
  accesskey   %Character;    #IMPLIED  -- accessibility key character --
  shape       %Shape;        rect      -- for use with client-side image
maps --
  coords      %Coords;       #IMPLIED  -- for use with client-side image
maps --
  tabindex    NUMBER         #IMPLIED  -- position in tabbing order --
  onfocus     %Script;       #IMPLIED  -- the element got the focus --
  onblur      %Script;       #IMPLIED  -- the element lost the focus --
  >

<!ENTITY % URI "CDATA"
    -- a Uniform Resource Identifier,
       see [URI]
    -->

It happens that the href attribute in HTML is also defined for a few other
element types, for example LINK and BASE, and for these elements it is also
CDATA with essentially the same rules and meaning.  This might lead one to
think that the HREF attribute of an A tag is the same attribute type as the
HREF attribute of the LINK tag.  One might even go so far as to imagine that
both HREF attributes are defined in some global scope, at the same level as
elements, in other words, in the "HTML" namespace.

The English text of the specification clearly implies that these attributes
are to be seen as essentially the same.

But the DTD does not say this.  The English text and our experience using
HTML gives us information not written in the DTD. DTDs today do not have any
facility to express that two attributes with the same name, used within
different element types, are the same in type or meaning.  Within the
expressive power of DTD, all we can observe is that the A tag has an
attribute named HREF with certain properties.  That is, there is an "href"
attribute used within the "A" tag.  There is also an "href" attribute used
within the "LINK" tag.  There is also an "href" attribute used within the
"BASE" tag.  And so forth.

They happen to have the same properties.  We know that they have the same
meaning.  But a DTD could equally well have defined an "href" attribute for
a "TABLE" element that was a NUMBER and had no relation to the "href" of an
"A" element.  The DTD can only go so far as to say that an "href" used on an
"A" element has certain properties, and that there is also an "href" on a
"TABLE" element that has certain other properties.  

In other words, the definition of an attribute, within the expressive powers
of DTD, is relative to its containing element. An unqualified attribute is
identified by the triple consisting of the namespace of the element, the
name of the element, and the name of the attribute. (See appendix A.3 of the
namespaces spec.) This is a long build-up to the conclusion that an
unqualified attribute cannot be presumed to come from the same namespace as
its element, without qualification. Rather, the element provides a local
namespace for its attributes.  

So is it true that

 <html:a href="foo"> is identical to
 <html:a html:href="foo">

?

There is nothing in a DTD that would allow us to decide that they are
identical.  They are no more the same than if the HTML DTD defined another
element called "ANCHOR" and the English spec said that it has exactly the
same meaning as the "A" element.  We might, reading the English spec, know
that they are the same.  But qua DTD, they are different.

The namespaces specification is designed to support the current practice of
DTDs, and so defines unqualified attributes in this way, while leaving the
door open to future forms of schema that express richer and more-global
attribute scopes.

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo@ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo@ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa@ic.ac.uk)