List Info

Thread: The datetime screen reader problem is almost complete bollocks




The datetime screen reader problem is almost complete bollocks
country flaguser name
United Kingdom
2008-03-28 12:56:43
We had a very experienced screen reader user in today (http://www
.abilitynet.org.uk/webteam#robin 
) testing the new BBC iPlayer mark-up with abbr design
pattern.

He was adamant that the abbreviation title tag problem was
hardly  
going to affect any screen reader users. Why? Because by
default, the  
title attribute isn't used much. Users of JAWs 9, 7.1,
Windows eyes  
6.1, Hal 9.91 have to edit the preferences to start reading
the title  
tag in abbr. And users who don't change the default settings
are the  
most important to accommodate - they're in the majority.

 From these preliminary results it looks like the BBC can
push the  
date-time-pattern and screen readers can play catch-up if
needed to.  
We'll have more in-depth tests and reports done in a month
or so.

Adam Craven

PS. There was an instance where a JAWS preference,
'Abbreviation  
expansion/meta data' caused unexpected results. It reads the
title  
tag, even when the content is display:none.
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

Re: The datetime screen reader problem is almost complete bollocks
user name
2008-03-29 06:34:53
Adam Craven - Four Shapes wrote:
> He was adamant that the abbreviation title tag problem
was hardly going 
> to affect any screen reader users. Why? Because by
default, the title 
> attribute isn't used much. Users of JAWs 9, 7.1,
Windows eyes 6.1, Hal 
> 9.91 have to edit the preferences to start reading the
title tag in 
> abbr. 

Thanks for sharing this information with the list, though I
did note 
that reading TITLE was a non-default configuration
option during the 
extensive discussion of this problem last year:

http://microformats.org/discuss/
mail/microformats-discuss/2007-April/009416.html

> And users who don't change the default settings are the
most 
> important to accommodate - they're in the majority.

Sorry, but a problem being rare is not the same as it being
"bollocks".

Actually, from lurking on mailing lists for screenreader
users, changing 
settings seems commonplace. Of course, correspondents might
be 
disproportionately expert on those lists.

However, as I again noted last year, changing the behaviour
for TITLE 
seemed unlikely to be a popular configuration choice. I
don't recall the 
subject coming up on the lists I'm on since. Of course,
someone 
encountering an accessibility issue with microformats might
have no idea 
what was causing the problem.

Still, catering to the "majority" of screenreader
users is reasonable if 
we're using the underlying (X)HTML standards in the right
way.

I wasn't persuaded that the abbr-date-title pattern was a
correct use of 
ABBR or TITLE back in April 2007, and I'm still not persuaded
today:

* I just don't buy that "Monday" is an
"abbreviation" of 
"2007-08-08T12:12" in the same way as
"W3C" is an abbreviation of "World 
Wide Web Consortium". Consequently, I don't think it's
a correct use of 
ABBR. None of the alternatives suggested by the WASP
Accessibility Task 
Force had this particular problem:

http://www.webstandards.org/2007/04/27/haccessibility/


* More weakly: some versions of ISO date times are more
human-friendly 
(and more screenreader-friendly) than others, but I'm not
persuaded that 
publishers would normally choose such a representation if it
weren't 
intended it for machine rather than human consumption, so I
think it's a 
dodgy use of TITLE. Given part of the rationale for
using TITLE is to 
make this data "visible" to end-users, it seems a
tad inconsistent to 
bank on it not being rendered to screenreader users.

So I tend to regard abbr-title-date as a failure to live up
to 
microformats' mission to build "upon existing and
widely adopted standards":

http://microformats.or
g/about/

> We'll have more in-depth tests and reports done in a
month or so.

I look forward to these with interest.

--
Benjamin Hawkes-Lewis
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

Re: The datetime screen reader problem is almost complete bollocks
country flaguser name
United States
2008-03-29 12:51:05
Benjamin Hawkes-Lewis wrote:

> Actually, from lurking on mailing lists for
screenreader users, changing
> settings seems commonplace.

I was going to reply to Adam's message last night, but was
too tired so 
went to bed instead. Ben has beaten me to it, and my
thoughts on the 
matter are very much in line with his.

Yes, it is true, that for the most part, people do not
change their 
software's initial settings. However, there are a certain
class of 
settings that many people -- perhaps even *most* people
(yes, even non-
technical people) do change. Probably the best example is
desktop 
wallpaper. How many people do you really know who use their
operating 
system default desktop wallpaper?

The question is: is the behaviour of screen readers with
regards to the 
ABBR element one of this special class of frequently changed
settings, or 
is it not? Unless someone can cite actually research into
this matter, I'm 
unhappy to go along with the "it doesn't cause a
problem with real-world 
screen reader users" hypothesis.

Ben also raises another point that I wasn't going to mention
in my 
abandoned theoretical late night message. It's another point
with which I 
agree, and have probably mentioned before on this mailing
list: that of 
semantics. Even if all the accessibility problems could be
solved 
overnight, semantically ABBR is often still the wrong
element to use - 
certainly for the datetime design pattern, but also for many
other ABBR 
design pattern use cases. It's just icky -- especially for
hCalendar's 
dtend used with a date (rather than a full datetime).

What are the alternatives:

The Web Standards Project's suggestion to use OBJECT and
PARAM for 
datetimes <http://www.webstandards.org/2007/04/27/haccessibility/
> is 
nice, but can cause browser funniness, and it's too much to
type anyway.

There are issues with simply supporting "title" on
any old element. The 
title attribute is used too much in the wild for us to
easily use it. For 
example, how would this work?:

	<div title="My hCard"
class="vcard">
		<span class="fn">Toby
Inkster</span>
		...
	</div>

Personally my preferred solution is Andy Mabbett's
suggestion from 
February this year <ht
tp://microformats.org/discuss/mail/microformats-
discuss/2008-February/011583.html>. This has been
implemented in 
Cognition, my semantic web parser, of which version
0.1alpha6 is due out 
later today. I can vouch that it's no harder to implement
than ABBR, and 
it's very easy for a tool to support both.

-- 
Toby A Inkster BSc (Hons) ARCS
[Geek of HTML/SQL/Perl/PHP/Python/Apache/Linux]
[OS: Linux 2.6.17.14-mm-desktop-9mdvsmp, up 3 days, 4:45.]

                                 Earth Hour
            
http://tobyinkster.co.uk/blog/2008/03/29/earth-hour/

_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

[1-3]

about | contact  Other archives ( Real Estate discussion Medical topics )