- From: Ian Hickson <ian@hixie.ch>
- Date: Sun, 18 Aug 2002 21:33:20 +0000 (GMT)
- To: Elliotte Rusty Harold <elharo@metalab.unc.edu>
- Cc: "www-tag@w3.org" <www-tag@w3.org>, "www-style@w3.org" <www-style@w3.org>
On Sun, 18 Aug 2002, Elliotte Rusty Harold wrote: > > You have a very different definition of huge traction than I do. See my previous post. > Even more importantly I repeatedly notice that even the most > recent browsers fail to handle basic XML techniques like CDATA > sections and non-predefined entity and character references. With content sent as text/html or text/xml? If you send a document as text/html to a UA, it has to [1] handle it as HTML (tag soup), which doesn't have CDATA blocks and so on. If you send content as text/xml, then it should be handled correctly. [1] Not for spec reasons, but because otherwise it wouldn't be able to render the majority of the sites that I gave in my last post, and so would gain zero market acceptance (users don't like UAs that render fewer pages than the competition). It's a sad fact that XHTML is being authored as tag soup now, and thus UAs are forced to handle text/html as tag soup, even if it looks like XHTML. Thankfully, UAs are already correctly handling broken text/xml documents, so this problem won't be propagated to the post-text/html world. > By pure XML, I meant "make up your own vocabulary as you go" rather > than using something that's been defined for you. You should never, _ever_ send arbitrary markup in a language you made up over the network (unless you have full control over the target UA). Making up your own vocabulary is one of the worst possible things to do in terms of accessibility, semantic web content analysis, and user control. -- Ian Hickson )\._.,--....,'``. fL "meow" /, _.. \ _\ ;`._ ,. http://index.hixie.ch/ `._.-(,_..'--(,_..'`-.;.'
Received on Sunday, 18 August 2002 17:33:21 UTC