- From: David Carlisle <davidc@nag.co.uk>
- Date: Thu, 31 Aug 2006 12:35:15 +0100
- To: www-math@w3.org
> 1) The html extension continues being the same since JS cannot change > 1) that. The extension is irrelevant (according to the specs) and even in IE where it is used as a hint as to the mime type, it's the mime type that's more important. > 2) The original HTML 4 DTD is removed from the tree, but there i no xml PI > declaration. The XML declaration isn't a PI and it wouldn't mean anything in a DOM anyway as its main purpose is to specify the encoding used in a linearisation of an XML document so has no meaning in a tree model where all characters are unicode. > 3) The final DOM continues being typed as _text/html_ again mime types don't really mean much by the time you have an model in memory. > Therefore, all i can say is that MathML is being encoded as text/html, > without namespaces and html extension. it's true you can say that, yes. > On any case the page i cited proves that you can display math in html > pages even if is an indirect way and that this apparently works for > MSIE+MathPLayer. By the time that it gets to the renderer, IE needs html and mozilla needs xhtml, the pmathml.xsl or auto-detect doctype solutions work by the server serving xml and IE internally transforming that to html. The js mechanism works the other way, with pages being served as html and mozilla transforming to xml. neither mechanism changes the basic rendering behaviour. It's like saying "<p>hello</p> produces a paragraph that says hello" That statement is basically true even though it's easy to generate documents served as html or xhtml together with javascript or xsl or even css that makes the actual rendered result say something else. David
Received on Thursday, 31 August 2006 11:35:26 UTC