W3C home > Mailing lists > Public > public-html@w3.org > November 2007

Re: HTML syntax

From: Thomas Broyer <t.broyer@gmail.com>
Date: Wed, 21 Nov 2007 18:01:43 +0100
Message-ID: <a9699fd20711210901h61f9c694o33bfc3dfb38a0e90@mail.gmail.com>
To: public-html@w3.org

2007/11/21, Dean Edridge:
> Anne van Kesteren wrote:
> > and believing we might move the whole Web to use XML syntax has been
> > proven wrong before.
> No it hasn't been proven wrong. I think that's just your opinion. What
> did you mean exactly? Did you mean that changing the world to XHTML
> didn't work.
> Millions of people use wordpress and that uses the XML syntax.

Wordpress uses XML-*like* syntax, but uses string concatenation and
serves the whole thing as text/html (so no browser even try to use its
XML parser; all browsers out there use their HTML/tag-soup parser).
Moreover, it's really easy to change Wordpress's template to output
HTML5 (though Wordpress has some hard-coded "/>" in it; if Wordpress
wordked with a DOM, it'd only be a matter of serialization).

> > As Maciej already explained and you know doing both HTML and XHTML at
> > the same time is hard.
> I'm not talking about that, just the quoted attributes & solidus in the
> void elements.
> It's not just what I wish to do in the future. It's about the whole
> world, it's about all the documents that will be circulating out there.
> They will be incompatible with each other.
> Because a few people are too proud and stuborn to change. Therefore
> millions of people are going to have to deal with the unnecessary
> discrepancies.

There are really few people really doing XHTML (i.e. serving it as
application/xhtml+xml). Most people pretending they do XHTML (i.e.
serve XML-like documents as text/html) are failing to do it right
(someone sent some figures to this list or the WHATWG a few
weeks/months ago).

> How on earth do you think that HTML5 and XHTML5 will live on the web at
> the same time without some type of increased normalisation between the two?

I don't expect XHTML5 to live on the Web.
XHTML is a wonderful thing when mixing HTML with other XML
vocabularies and/or using an XML toolchain (and even then, you could
just use an HTML parser and then work with the DOM); but I do think
that XHTML on the Web has failed.

> > It is, actually. Just because you want "the one syntax"
> It's not just about me.
> > doesn't mean I should suddenly change all my HTML pages and create new
> > resources for the scripts they contain.
> Why not? Who isn't going to have to make changes to their sites, CMS,
> WYSIWYG editors and so on to suit (X)HTML5.

First, HTML5 has been designed to be backwards compatible (at least
wrt parsing, if not at the validation stage) so that most valid HTML 4
(or previous versions) documents will still be valid HTML5 documents.
Second, if you're doing HTML 4 correctly, why would you bother moving
to HTML5? (provided you have no need of HTML5 additions)

Many pages on the wild are still HTML 3.2, most pages still use
<!--comments--> around their inlined <style> and <script>, etc. Most
people have made no change to their pages for ages, so why would they
suddenly switch to HTML5?

New products (or new versions of existing products) will probably
align on HTML5, but so what? Many people are still using IE 6, some
are still using Firefox 1.5, why would they switch to an
HTML5-compliant browser? and such a browser would still have to be
"compatible" with non-HTML5 pages, so why switch to produce HTML5
pages? (provided you have no need of HTML5 additions)

> >> I don't know why I would need to use some "weird workarounds". Unless
> >> you are referring to document.write and innerHTML.
> >
> > Weird workaround are needed if you want inline scripts and style
> > sheets, which people definitely want.
> I said that the quoted attributes could be normallised between
> serialisations. I never said anything about converting the world to XHTML.

You said that adding the xmlns declaration would be done automatically
when converting to XHTML; Anne replied that html5lib (among others) is
capable of parsing an HTML5 document to DOM which can easily be
serialized back to XHTML, so other differences between the two
serializations can be worked out that way.

Thomas Broyer
Received on Wednesday, 21 November 2007 17:02:01 UTC

This archive was generated by hypermail 2.4.0 : Saturday, 9 October 2021 18:44:24 UTC