OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   RE: A philosophical question?

[ Lists Home | Date Index | Thread Index ]
  • From: "Aaron Skonnard" <aarons@develop.com>
  • To: <xml-dev@xml.org>
  • Date: Thu, 11 May 2000 12:48:18 -0600


> > > In the XPath spec(and, I suspect, in the XML InfoSet spec, although I
> > > haven't read that one yet- it is on the list), the following statement
> > > made in the Data Model (section 5.3, for example, also 5.4 re:
> > > nodes) "... the element is the parent of each of these attribute
> > > however, an attribute node is not a child of its parent element."
> > >
> > > I try not to be too obsessed by symmetry, but I find this statement
> > > perplexing.  Any insights?  And what are the implications, if any, for
> > > implementation?
> >

This once again illustrates the problems that have surfaced from not
standardizing on an abstract data model before defining the layered
specifications (see the archives for similar discussions). This is exactly
why we need the Infoset to explicitly define the abstract data model once
and for all.

> > It's simple... You have an element node, say the method to get the
> > children is getChildNodes(), this shouldn't return the attributes or the
> > namespace nodes.
> >
> > However if you've got an attribute node, calling getParent() should
> > the element node.
> Note that the W3C DOM Level 1 specification also says that elements
> are not parents of attributes, to get rid of these sorts of
> misunderstandings. Thus using the DOM getParent method will not get
> you the element which the attribute is associated with, you have to
> use the Level 2 method getOwnerElement (see
> http://www.w3.org/TR/DOM-Level-2/core.html#ID-637646024 for details)

If you read the Infoset, you'll see that elements have an [attributes]
property and that attributes do not belong to an element's [children]
property. Furthermore, you'll see that attributes have an [owner element]
property as opposed to a [parent] property. Once the Infoset becomes a
Recommendation, it should be considered authoritative on issues like this.
If existing specs conflict with the Infoset model (as with this XPath
example), hopefully future versions will re-align themselves.


This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS