On Dec 14, 2010, at 05:17, David Carlisle wrote:
> I've no complaint with html5 having defined fixup rules to give consistent error recovery from overlapping markup and other horrors, but I think the fact that it parses well formed XML and produces different trees is just wrong.
The easiest proof why it has to be this way is:
There are Web pages that rely on the <html> tag getting implied per HTML 4 when not present in the source text. Therefore, the HTML5 parsing algorithm always outputs a tree whose root element is html. There are XML documents whose root element is not html. Therefore, it has to be that there are well-formed XML documents that parse into different trees using an XML parser and an HTML parser.
--
Henri Sivonen
http://hsivonen.iki.fi/
