[uf-discuss] An Inconvenient hCard

Paul Wilkins pmw57 at xtra.co.nz
Tue Mar 13 10:58:35 PST 2007


Tim White wrote:

>----- Original Message ----
>>From Ara
>
>On 3/12/07, James Craig <jcraig at apple.com> wrote:
>  
>
>>Paul Wilkins wrote:
>>
>>    
>>
>>>With the abbr design pattern, you encode the machine-readable info
>>>around the human-readable words.
>>><p class="tel"><abbr class="type" title="fax">Téléc</abbr>: <span
>>>class="value">(514) 123-4568</span></p>
>>>      
>>>
>
>  
>
>>This is a misuse of abbr at best.
>>
>>See: open issue! 2007-01-26
>>http://microformats.org/wiki/hcard-issues
>>    
>>
>
>  
>
>>Okay, so what does a guy do in a case like this then? 
>>    
>>
>
>What about something along the lines of: 
>
><p class="tel"><span class="type" title="fax">Telec</span>: <span class="value">(514) 123-4568</span></p>
>
>Title is a valid HTML attribute and this avoids abusing the <abbr>. 
>  
>

The possibility of using the title attribute to hold generic machine 
readable information was considered and rejected.

The title attribute is already commonly used on many elements today. If 
people were to add a microformat to their page, they would have to 
remove all title attributed content except for that which they 
specifically intended to be the machine readable information. This then 
becomes an abuse of the title attribute and as such is not allowed.

Another reason for discarding the idea was that it makes it too easy for 
people then to hide information. If someone is not willing to make the 
information available in plain text, then that information shouldn't be 
used.This is why we people shouldn't be encouraged to store information 
invisibly because that just leads to bad things.

Restricting the title attribute for microformatted information to just 
the abbr element itself was considered and allowed, because that element 
at least, is purpose built to show a different representation of the 
information within it, and that's why it's become so useful with dates 
and other content, to allow human readable content with a machine 
understandable representation.

When it comes to languages other than English though, this is where 
things get tricky, and is covered in the FAQ at 
http://microformats.org/wiki/faq#Q:_How_do_microformats_breach_language_barriers.3F
Effectively, the language barrier is a pre-existing unsolved problem.

While it's fine for class names to be represented in English, what 
happens when the readable content is used as keyword data. We are left 
with one of two options. Either the readable data needs to remain as 
English, or some other way needs to be found to provide a representation 
of that non-english data in an english form.

Currently there hasn't been a more appropriate answer to this than the 
abbr design pattern, and it is possible to justify. The computer is able 
to understand only a very small number of terms, and each of those terms 
can be expressed in a wide number of human readable formats. So what 
we're doing with the abbr design pattern is to take one of these vast 
number of human readable formats and abbreviate all of those 
possibilities down to one common computer understandable term.

<abbr class="type" title="fax">Téléc</span>:

The only viable alternatives are either to spell it out in full,

Téléc (<span class="type">Fax</span>):

Or to hide the english form of the word

Téléc <span lang="en" style="display: none;" class="type">Fax</span>:

This is one of those situations where whateveryou do, you're guilty of 
something.

-- 
Paul Mark Wilkins
New Zealand Tourism Online
pmw57 at xtra.co.nz <mailto:pmw57 at xtra.co.nz>
109 Tuam Street
Level 1
Christchurch 8011
New Zealand
+64 3 963 5039


More information about the microformats-discuss mailing list