[uf-discuss] ISBN mark-up

Tantek Ç elik tantek at cs.stanford.edu
Tue Apr 25 07:31:29 PDT 2006


On 4/24/06 8:09 PM, "Benjamin Carlyle" <benjamincarlyle at optusnet.com.au>
wrote:

> On Mon, 2006-04-24 at 23:39 +0100, Andy Mabbett wrote:
>> In message <D53CA839-59AC-4E9E-AEDB-BE3ABB2B0C63 at hubmed.org>, Alf Eaton
>> <lists at hubmed.org> writes
>>>> How would you see that sitting inside HTML mark-up?
>>> <abbr class="uri" title="urn:isbn:0950788120">0 9507881-2-0</abbr>
>> I'm very concerned that that may be an abuse of "abbr"; and wonder how
>> it will sound when read to someone using assistive software?
> 
> It would be nice to have a more specific html element to work with. What
> we use abbr for most of the time in microformats is not to provide an
> expansion of an abbreviation, but to provide a canonicalisation.

Not quite.  We absolutely do use abbr to provide a precise expansion of an
abbreviation, most often, a date or time.  The format we choose for that
expansion happens to also be regular enough to be consistently machine
readable in a culture-independent manner (as opposed to dates like 5/11
which mean different things in different cultures).

> I think it becomes odder as you cross language and culture boundaries.

Actually, since the expansion is culture independent, our use of abbr is
actually more useful as you language and culture boundaries.


> In these cases canoncalisation to a machine-readable format may be
> increasingling important,

Right.

> but the mapping to "abbr" increasingly skewed.

It's not skewed at all.  We are simply using abbr in one particular way it
can be used.  Just because the one particular way we are using is something
the authoring community is not used to doesn't mean it is skewed, merely
novel.


> Hixie: Help! :)

Search for datetime element and you'll find that this has been
proposed/discussed for a few years now.  I personally prefer <t>, based on
the universal usage of t in physics to indicate datetimes (and that it is
shorter), but I've left it to Hixie to decide what he thinks would be best
for HTML5.

Thanks,

Tantek



More information about the microformats-discuss mailing list