Illustration of Different Node Trees for XML (W3C(XT)/IE5)

[G. Ken Holman]

At 99/04/09 12:11 -0700, Chris Lovett wrote:
>(1) IE5 gives you the option on text nodes between elements.  These are
>called "ignorableWhiteSpace" and we have a switch on the IXMLDOMDocument for
>this.

>From the perpective of a stylesheet writer (me!), WD-XSL-19991218 states
the default for the source tree is all whitespace element content is
significant and the default for the stylesheet tree is XSL-type-ignoreable
element content whitespace is ignored (neither of which appears to be true
for IE5-XSL) ... if I counsel someone regarding portable stylesheets, the
WD describes the behaviour I expect, so the fact that I may be able to
manipulate the DOM (which I can't with Working Draft 2) doesn't help me here.

>(2) IE5 gives you all PI's.  I don't see why the <?xml delcaration is not a
>valid PI for representation in your DOM.  What if you want to change the
>encoding and save the document back out ?

According to REC-XML production [23], the XML declaration is just that, and
*not* a processing instruction.  The stylesheet can expose processing
instructions, not the XML declaration.

Again you mention the DOM ... I'm talking about portable stylesheets and
the document structure presented to me and my customers with our stylesheets.

>(3) Again, namespace attribute are attributes.  What if you want to promote
>a namespace declaration to a higher level in the tree ?

Again you are speaking of someone manipulating a DOM ... according to
WD-XSL-19991218 section 2.4.4, attributes whose name starts with "xmlns:"
create a namespace node, not at attribute node, and my stylesheet was
written to expose attribute nodes, not namespace nodes ... hence I feel the
behaviour witnessed is not correct.
>(4) Well this all points out the fact that the DOM group refused to consider
>namespaces in level 1.  IE5 gives you the namespace info as separate
>properties called "namespaceURI" and "prefix" rather than inventing a new
>node name format.  This way someone can add a namespace to a document
>without breaking an app that is already written to the nodeName as a simple
>GI.  This give a better migration story.

Indeed for one manipulating a DOM for later emission as a new document ...
but that wasn't my perspective.

I was analyzing the state of a node tree as presented to an XSL stylesheet
writer ... not analyzing what I could and could not do with a node tree in
general.

Thank you for taking the time to share your observations ... I hope I've
clarified my own goals in my analysis ... I felt it important to share with
others who may be writing stylesheets who should be considering the
portability of stylesheets and their behaviours (some particularly strongly
held feelings on my own part).

........... Ken


--
G. Ken Holman                  mailto:gkholman at CraneSoftwrights.com
Crane Softwrights Ltd.           http://www.CraneSoftwrights.com/x/
Box 266, Kars, Ontario CANADA K0A-2E0  +1(613)489-0999  (Fax:-0995)
Website: XSL/XML/DSSSL/SGML services outline,  XSL/DSSSL shareware,
         stylesheet resource library, conference training schedule,
         commercial stylesheet training materials, on-line XSL CBT.
Next instructor-led XSL Training:                   WWW8:1999-05-11


xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)