Paul Connolley wrote: > > Geoff Deering wrote: > > > I am talking about CSS applied to HTML and the rendering of the CSS as > > applied to the parsing of the document. But still, strictly speaking, > > an > > XML based document is bound to be more semantically correct because it > > is > > well formed. This means that the CSS can be applied without fear of > > the > > parser misunderstanding where a declaration could have finished. > > There is > > no possibility of any guess work in xhtml as it is well formed. > > You are talking about two distinctly different parsers. XHTML is a XML > subset, HTML is an SGML subset. > > For example: > > In XHTML: > > <ul> > <li>One item</li> > <li>Second item > <li>Third item</li> > </ul> > > This will throw an error because it is bad XML > > In HTML it will not. With the SGML parser it knows that when it arrives > at a new <li> it is the beginning of a new list item. The same would > apply, for example, to a <p> paragraph. Surely you don't believe that > it would render > > <p>This paragraph > <p>This other paragraph > > as a paragraph within a paragraph. You need to revise your > understandings. HTML is NOT XML. Admittedly, XHTML contains inheritance > to certain HTML objects but it is wholly a XML subset. You shouldn't > try to argue about parsing when they are parsed by two different types > of engines.
But that is entirely my point. > > You're missing the point. Closing tags is being completely accurate > > with > > punctuation, where markup is the punctuation. Not closing tags CAN > > lead to > > ambiguity. In XHTML there is no syntax ambiguity, in HTML4 there are > > possibilities. It may not happen when validating against the doctype. > > Once again you are arguing for ambiguity. It is wrong to assume this. > SGML allows for the omission of end tags because it follows a different > ruleset > That's exactly the point I am trying to make. ----- Geoff ****************************************************** The discussion list for http://webstandardsgroup.org/ See http://webstandardsgroup.org/mail/guidelines.cfm for some hints on posting to the list & getting help ******************************************************
