W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > April to June 2003

Re: accessify.com's review of RNIB relaunch

From: Matt May <mcmay@w3.org>
Date: Wed, 25 Jun 2003 13:49:42 -0700
Cc: tina@greytower.net, w3c-wai-ig@w3.org
To: Kynn Bartlett <kynn@idyllmtn.com>
Message-Id: <8BA02FA3-A74E-11D7-9CD0-000393B628BC@w3.org>

On Wednesday, June 25, 2003, at 10:40  AM, Kynn Bartlett wrote:
> It's my contention that as long as the _relationship_ between
> the headers is reasonable, the exact _numbers_ employed need
> not matter -- you will still be able to construct an appropriate
> hierarchical structure of the site even if the document
> only uses the even-numbered elements (<h2>, <h4>, <h6>).

I want to make sure that when someone collapses headers into a tree (as 
some authoring tools do) or has them read out in header view, the 
relationships they draw are logical. If an h4 that follows an h3 is not 
in fact a child of that h3, I consider that bad design.

I don't intend to make people swear that they will forever avoid 
skipping headers (after all, Emerson had something to say about a 
foolish consistency [1]). But I do intend to make them aware that using 
markup consistently helps the overall design of their site.

>> XHTML 2 is introducing a <section> element, so that headers within a 
>> given section would "know" which level they are.
> This is actually support for my position -- it shows that the
> exact numbers used are unimportant, and that what matters is
> the relative hierarchy established by the heading tags.  If the
> exact numbers mattered -- if <h1> _had_ to be the first tag,
> for accessibility's sake -- then the XHTML 2.0 proposal of
> <h> tags would be an affront to accessibility.  However, the
> truth is that it's a boon to accessibility (as well as
> portability).

The numbers of the elements could be considered a hack. What (aside 
from sane, rational analysis of the content that I have) should stop me 
from exposing seven or more layers of hierarchy? But here, you only 
have six.

The <section> and <h> approach solves two problems: First, it allows 
headers to be semantically linked with their supporting content, which 
is not something that is hardwired into HTML. Second, it allows nested 
headers to inherit their depth from the number of sections above them, 
which forces the kind of hierarchy the designers of the language want.

> Ergo, the exact numbers do _not_ matter.  And an insistence
> on <h1> as the first header (instead of <h2>) is inappropriate.

I'd have to see "matter" defined. Does it rip a hole in space-time? No. 
Is thoughtful use of headers something I'd consider a best practice, 
from code quality and usability standpoints? Yes. (Are you hurting 
yourself in search engines by not giving your best searchable content 
[the title] the most important markup [<h1>]? Probably.)

[1] http://www.brainyquote.com/quotes/quotes/r/q101115.html

Received on Wednesday, 25 June 2003 16:50:56 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:36:16 UTC