Clarification sought on a few semi-obscure DOM points

Looking through the latest set of DOM-TS tests has turned up a couple of 
issues where I'm not sure what the DOM specs say (if anything). Would 
appreciate any comments/resolution on:

1. Can Text nodes exist inside the Document node to represent white space?

I don't see anywhere where Core specifies there aren't, and it would 
seem useful to be able to preserve the formatting of the document (when 
not outputting in canonical-form, which specifies exactly one newline 
character between each document-level node).

On the other hand, document-level white space is not part of the infoset.

If there *can be* white space Text nodes inside the Document, should 
they be considered 'element content whitespace'? They're not inside an 
element declared to contain no textual content (because they're not in 
an element at all), but they *are* in a place which can't contain 
textual content.

2. Does the DOMConfiguration parameter 'canonical-form' have any effect 
on LSParser operations?

I had seen canonical-form as purely an output-stage feature, but a few 
DOM-TS tests are requiring some of the effects of canonical-form (eg. no 
doctype, default attributes becoming specified) to occur when it is set 
on LSParser.domConfig.

(It is possible this is by accident, as one of the test descriptions 
mentions it is supposed to be checking normalisation. However the 
question still stands.)

Any thoughts?

-- 
Andrew Clover
mailto:and@doxdesk.com
http://www.doxdesk.com/

Received on Sunday, 23 May 2004 17:09:25 UTC