Le Sat, 10 Mar 2007 12:39:46 +0200, Jorgen Horstink
<[EMAIL PROTECTED]> a écrit:
On Mar 10, 2007, at 11:16 AM, Mihai Sucan wrote:
Adding a new DOCTYPE switch is not a solution to Microsoft's problem.
As far as I understand it, the new DOCTYPE switch is meant to 'tell' to
browser the document follows the HTML5 specification. HTML5 is set up to
be backwards compatible with HTML4 documents. The opposite does not
hold. There must be at least one new DOCTYPE to 'tell' the browser HTML5
is being served.
<!DOCTYPE html> seems to be a suitable candidate. This doctype can be
used by vendors to proxy the content to the right rendering engine.
Vendors can either rebuild a new engine from scratch, or improve
specific parts of their rendering engine.
I believe this is wrong.
This DOCTYPE is *not* meant to 'tell'/inform/advertise the document as
HTML5. It's definitely not the case, see the FAQ [1].
For one, HTML5 is a specification defining new features, and redefining
parsing, breaking the SGML and XML rules. All the error recovery, and all
the parsing rules are picked so that an implementation of HTML 5 will
properly support HTML 4 as used today on the majority of web sites.
Backwards compatibility is the key. Of course, it's utopic to believe that
a specification can be written to accomodate 100% of all the web pages
found. Yet, HTML 5 does provide parsing algorithms that work on the
majority of code found on the web.
There's no way to advertise the document as HTML 5, and it's certainly not
the purpose of the specification to do so.
(If I am wrong, an expert should correct me.)
[1] http://blog.whatwg.org/faq/#doctype
--
http://www.robodesign.ro
ROBO Design - We bring you the future
