[uf-discuss] definitive hCard on a page

[Scott Reynen]

On Jul 4, 2007, at 4:06 PM, Andy Mabbett wrote:

>> Interesting question, if not what Thom needed.  I try to use <h1> to
>> identify the subject of a page, so perhaps something like <h1
>> class="fn">Paul McCartney</h1> could cover that scenario.
>
> Not if the existing page has:
>
>         <h1>Acme Rock Biographies Inc</h1>
>         <h2>Paul McCartney</h2>

I think the subject is still clear here, if not what was intended.   
Headers describe the topic of a section of content.  In this case,  
this headers say the content is primarily about Acme Rock  
Biographies, and secondarily about Paul McCartney.  Of course, many  
people are still using headers primarily for their default visual  
rendering without regard to their meaning, so there's still a lot of  
work to do convincing publishers to use HTML markup meaningfully.

>>  But it's  hard to know without an actual web page to look at, and  
>> I'm
>> not sure  what we'd do with the subject after we'd identified it.
>
> One application could be for a search engine - "find me pages *about*
> Paul McCartney, but don't bother with those that just /mention/ him".

Search engines already analyze the structure of HTML documents, e.g.  
headers, to determine subject matter, but they don't treat the markup  
as definitive.  I think the only thing required to accomplish this  
scenario would be for a search engine to assume the structure of  
markup actually reflects the structure of the content in HTML  
documents, i.e. assume the web is semantic.  I doubt they'll see much  
value in that until much more of the web actually is semantic, but  
I'd certainly be interested in seeing the results of such a search  
engine.

Peace,
Scott