W3C home > Mailing lists > Public > public-html@w3.org > January 2011

Re: rationale for inclusion of hgroup in html5?

From: James Graham <jgraham@opera.com>
Date: Thu, 27 Jan 2011 18:08:28 +0100
Message-ID: <4D41A68C.8080202@opera.com>
To: Anne van Kesteren <annevk@opera.com>
CC: Jirka Kosek <jirka@kosek.cz>, Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>, Steve Faulkner <faulkner.steve@gmail.com>, HTML WG <public-html@w3.org>
On 01/27/2011 05:14 PM, Anne van Kesteren wrote:
> On Thu, 27 Jan 2011 17:11:54 +0100, James Graham <jgraham@opera.com> wrote:
>> On 01/27/2011 04:15 PM, Jirka Kosek wrote:
>>> - all<hx> elements are considered
>>> - if there are multiple<hx> elements inside<header> only first of them
>>> is considered
>>
>> That was, in fact, the very first solution that we developed. However
>> people seemed to find it confusing (iirc they didn't want <header> to
>> have anything to do with the outline algorithm). <hgroup> was an
>> attempt to fix that. It seems that the design is still considered
>> suboptimal, hence ideas like <subhead> and so on.
>
> It does seem to be the one that most naturally maps to existing markup.
> Maybe we got scared off to quickly by a few confused people?

I think the <header> model is prone to accidental misuse in a way that 
<hgroup> or <subhead> are not. For example I would expect to see lots of 
instances of:

<header>
<h1>My blog!</h1>
<h2>What I had for breakfast</h2>
</header>
<ul><li>Bacon

with the author being unaware that they had removed the article title 
from the outline structure.
Received on Thursday, 27 January 2011 17:09:05 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:17:21 GMT