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

RE: Screen-reader behaviour

From: DESCHAMPS Stephane ROSI/SI CLIENT <stephane.deschamps@orange-ftgroup.com>
Date: Fri, 31 Aug 2007 11:38:00 +0200
To: "'Philip Taylor \(Webmaster\)'" <P.Taylor@Rhul.Ac.Uk>
Cc: "'HTML Working Group'" <public-html@w3.org>
Message-ID: <002a01c7ebb2$9f229be0$8911b30a@stquentiny.francetelecom.fr>


> -----Message d'origine-----
> De : public-html-request@w3.org 
> [mailto:public-html-request@w3.org] De la part de Philip 
> Taylor (Webmaster)
> Envoyé : jeudi 30 août 2007 19:28
> À : Gregory J. Rosmaita
> Cc : HTML Working Group; wai-xtech@w3.org
> Objet : Re: Screen-reader behaviour


> 	<span class="Book-title">The Quare Fellow</span>
> 	<em class="Linnaean-binomial">Lagopus hyperboreus</em>
> 	<i class="Loan-word">Zeitgeist</i>
> 
> I have always assumed (somewhat naively !) that such markup 
> might be of help to those relying on screen readers, but if I 
> understand your answer correctly, this is not the case (at 
> least for JAWS).

Philip,

You *do* raise an interesting question, but the extensibility (it seems to
me) cannot possibly be implemented in a screen reader: the way it works, a
screen reader relies on a set of known HTML components (h*, lists, form
elements, paragraphs, tables) or to a lesser (but growing) extent on ARIA
roles and states (I've just tested Becky Gibson's treeview [1] in Firefox 2
with Jaws 7.1 and was baffled by the excellence of the information it gives
me).

How, even if a screen reader is going to be able to learn a new DTD
dynamically, is it going to know what to do with it: stress it? explain what
this or that is relying on descriptive mechanisms in your DTD?

Another question: Should it even telle me precisely what an element is? As
far as it's concerned, the notion of nonce-elements you're proposing doesn't
have to do with structure (whereas header, table, list do) or with different
content type (whereas image, plugin do). It has to do with text content,
just a bit more precisely so. From what I know, screen reader users set the
verbosity to a minimum anyway, because otherwise reading a whole page is
made very tedious.

On the other hand, what is really differentiating your three examples is
their language: and on this point screen readers do a very good job. So, to
be a bit more 'semantic' (sorry if I'm using the word too loosely), your
code could be:

<cite class="Book-title" lang="en">The Quare Fellow</cite> (especially
considering I'm on a french Os, and <cite> applies to book titles)

<em class="Linnaean-binomial" lang="la">Lagopus hyperboreus</em>

<i class="Loan-word" lang="de">Zeitgeist</i>

JAWS either reads using the appropriate phonologic engine if it is installed
(I can save sounds if anyone needs examples, I've got french, english,
italian on my machine), or announces the language if it's not installed
("la" for latin, for instance).

Considering, though, that accessibility is not for screen readers only,
marking up your content with microformats, RDFa, your own namespace or your
own classes would help in future-oriented ways, such as cataloging for
instance.

(I hope I was not too long-winded on my answer)


-- 
Best regards,
Stéphane Deschamps
  Web HCI expert
  orange / france telecom group

[1] http://www.weba11y.com/AjaxDemo/sample.html - Dojo has put this approach
to work in its tree component in Dojo 0.9, and it's impressive work.


*********************************
This message and any attachments (the "message") are confidential and intended solely for the addressees. 
Any unauthorised use or dissemination is prohibited.
Messages are susceptible to alteration. 
France Telecom Group shall not be liable for the message if altered, changed or falsified.
If you are not the intended addressee of this message, please cancel it immediately and inform the sender.
********************************
Received on Friday, 31 August 2007 09:38:11 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:05 GMT