W3C home > Mailing lists > Public > www-html@w3.org > February 2007

Re: [html] Elements within "title"?

From: Lachlan Hunt <lachlan.hunt@lachy.id.au>
Date: Thu, 01 Mar 2007 02:20:44 +1100
Message-ID: <45E59DCC.4080304@lachy.id.au>
To: "Jukka K. Korpela" <jkorpela@cs.tut.fi>
CC: www-html@w3.org, help-whatwg.org@lists.whatwg.org

Jukka K. Korpela wrote:
> On Tue, 27 Feb 2007, Henri Sivonen wrote:
>>> Compatibility with what?
>> With the existing practice.
> Well, any change means _some_ deviation from the existing practice. You 
> seem to imply that any future version of HTML has to be suitable for 
> existing browsers.

In practice, browsers will parse HTML5 exactly the same as they will 
parse legacy HTML pages (excluding some minor parsing differences in 
quirks mode).  The HTML5 parsing algorithm has been written with that in 

> While that has roughly been the idea when HTML has 
> been extended, it has never been very systematic. For example, when 
> tables were added a long ago, they weren't compatible in the sense that 
> non-table browsers would render them in any reasonably way (though there 
> were some tricks to get a somewhat reasonable rendering).

HTML5 also introduces new elements that aren't fully compatible, like 
table when it was introduced, but adding new features is different from 
changing and breaking existing ones.

> My question was whether that would be the case. I still don't see how, 
> though I'm not saying that allowing normal inline content in <title> 
> cannot seriously break anything. It's still an open question.

There are legacy documents on the web that include markup significant 
characters like < and >.  Changing the parsing to handle them as markup 
would break such pages simply because it is different from what browsers 
do today, and what authors rely on.

>> I think he meant the serializability of conforming XHTML document 
>> trees as HTML.
> Perhaps, but that really amounts to feeding XHTML, written with some 
> special restrictions, to tag soup processors. XHTML is as serial as HTML.

No, that's not right at all.  In HTML5, the HTML serialisation must be 
served as text/html and the XHTML serialisation must be served as XML. 
But when either is parsed, the DOM (or other model) is still created 
with the elements in the XHTML namespace (though that is a change from 
existing browser behaviour).

Lachlan Hunt
Received on Wednesday, 28 February 2007 15:22:08 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 30 April 2020 16:21:01 UTC