[uf-discuss] Re: Precise Expansion Patterns

[Martin McEvoy]

Hello Paul

On Sun, 2007-12-16 at 10:27 +1300, Paul Wilkins wrote:
> On Dec 16, 2007 8:00 AM, Manu Sporny <msporny at digitalbazaar.com> wrote:
> > What was the problem with the SPAN approach, again?
> >
> > <span class="duration" title="PT3M23S">3:23</span>
> >
> > - You can set most, if not all, screen readers to not verbalize @title
> >   in SPAN.
> > - We're not abusing ABBR.
> 
> I've been looking carefully through the HTML 4.01 specs and realised
> that there is no actual mechanism provided for what we are wanting to
> achieve. This is why there has been so much debate about this and it's
> why compromises will have to be made if we're to achieve our end
> result.
> 
> One of the things I came across is that the INS element accepts a
> datetime attribute, but it has to be fully specified as a moment in
> time. That might be useful for other applications like the calendar.

No ins is good and is valid markup...

> 
> The A element is the other one that I came across. We appear to have
> been completely ignoring this one, even though it looks to be a very
> suitable candidate.
> The specs say that it doesn't have to link to anything, and can
> contain the class and title of what we require.
>     "Authors may also create an A element that specifies no anchors,
> i.e., that doesn't specify href, name, or id."
> 
> <a class="duration" title="PT3M23S">3:23</a>

a is good too...

> 
> And if you don't want the title to appear when you hover over the
> text, you can leave the A element empty.
> 
> <a class="duration" title="PT3M23S"></a>3:23

the problem is as you say the above two will still put machine data in
the users view! from an accessibility point of view Manu mentioned
earlier, you can just turn off the titles? the title area of a link or
span or anywhere on a web page is important to people who do use screen
readers, they usually give information about content or about the links
they are about to visit, so turning of titles really is out of the
question.

The empty <a></a> although valid, seems nasty to me and a little anti!,
microformats are supposed to represent the data they contain...are they?

Thanks

Martin

 
> 
> In both situations the A element is not included in the tabbed flow of
> the document, because there is no anchor involved.
> 
> With this technique we are able to embed machine-only information
> around or near-to the human readable text.
>