Le Mon, 04 Dec 2006 18:10:18 +0200, Michel Fortin
<[EMAIL PROTECTED]> a écrit:
Le 4 déc. 2006 à 6:10, Mihai Sucan a écrit :
However, in the same "spirit", a middle way for those who want XMLiness
in HTML, would be to allow the xmlns:?.* attribute, xml:base, xml:id,
and xml:lang. Yet, define them as meaningless. Just for validation
purposes, just for helping people who do such things on the server-side.
I disagree. I think xmlns="http://www.w3.org/1999/html" is fine in HTML
because it just states what is already implied. But I think allowing
attributes with no meaning to validate will just reenforce the idea that
they are meaningful. If you use them, fine, but you'll have a validation
error there that will warn you that this is no HTML.
I know I suggested xml:lang before, but that was when I thought it was
parsed in HTML. Now I think a more clever approach would be to allow
html:lang to validate in XHTML, because XHTML already mandates that
html:lang be taken into account when determining the language.
Doesn't html:lang validate in XHTML? If not, this is news to me.
IMHO, it should validate. Why so? It's said that the HTML5 and XHTML5 are
one and the same language, with different serialization (HTML and XML,
hence different parsers must be used). The following code should be valid
in XHTML5 and HTML5:
<html lang="fr" xmlns="http://www.w3.org/1999/xhtml">
<head>
<title>Sans titre</title>
</head>
<body>
<p>Bonjour à tous!</p>
<p lang="ro">Bună ziua tuturor!</p>
<p><img src="merci.png" alt="Merci!" id="mon-image" /></p>
</body>
</html>
(apologies if mistakes in code, I am in a hurry)
As for xml:id, why not use "id" purely and simply? It works with both
HTML and XHTML anyway.
As I have said, I am not a supporter of allowing attributes just for the
sake of validation. I was just providing a "middle" suggestion, which
could satisfy some who want XMLiness.
Hence, I wish good luck to Hixie with saying "no" to some requests.
--
http://www.robodesign.ro
ROBO Design - We bring you the future