- From: Simon St.Laurent <simonstl@simonstl.com>
- Date: Tue, 18 Jul 2000 21:28:43 -0400
- To: "Larry Masinter" <masinter@attlabs.att.com>, <mmurata@trl.ibm.co.jp>, <www-html-editor@w3.org>
- Cc: <mark.baker@canada.sun.com>, <dan@dankohn.com>
At 06:16 PM 7/18/00 -0700, Larry Masinter wrote: >> Consider an XHML docment labeled as text/xml or application/xml. If >> this XML document has a linking PI to an XSLT stylesheet, what will >> happen? > >I think for the question to be useful, it is necessary to be explicit >about the context -- what will happen when? When an email is received >by a regular email client, labelled text/xml? By a web browser? Hey! I agree with Larry entirely here. I think the context makes the difference. Older HTML browsers, for instance, might ignore the processing instruction, but would probably be willing to process an entity body with an <html> in it some place. >> Will the document be handled by the XSLT processor or XHTML >> user agent? In my understanding, the result is undefined. > >Is it bad that the result is undefined? I'm not sure it is definable, at least given 'Associating XML documents with Stylesheets'. Given what seems to be a real transition from HTML to XHTML at the W3C, that may not be a bad thing. >> IE 5.0(1) >> invokes the XSLT processors. In the case of CSS, I believe that >> the XHTML user agent is invoked and it uses the CSS stylesheet >> for rendering. > >Is this OK? It seems fine - perfectly coherent, anyway - to me. Why not just add a warning in XHTML 1.1 that putting that PI at the front of a document may generate unpredictable results depending on whether the PI and XSLT are supported by any given environment? Simon St.Laurent XML Elements of Style / XML: A Primer, 2nd Ed. http://www.simonstl.com - XML essays and books
Received on Tuesday, 18 July 2000 21:26:16 UTC