[uf-dev] hCard locations in hCalendar

[Ryan King]

I've been looking at ways to better markup up (and parse) locations  
in hCalendar. I had been under the assumption that iCalendar's  
LOCATION field was just plain text. [1]

However, it turns out that, per RFC 2445, section 4.8.1.7, LOCATION  
can take an ALTREP parameter, which is "a URI that points to  
directory information with more structured specification of the  
location". So, in addition to the text, we can give a link to a  
structured representation.

This brings up several issues.

Assumption: if an agent is consuming iCalendar, they'd prefer to  
consume vCard, rather than hCard.

1. How could a service like X2V handle this?

   * It could supply links (piped through X2V's vCard service) to a  
vCard for the location. Ex (from: http://austin.adactio.com/, @id  
added):

   <span class="location vcard" id="six">
     <span class="fn org">Six</span>,
     <abbr class="geo adr" title="30.266128;-97.743816">
       <span class="street-address">117 West 4th St</span>
     </abbr>
   </span>

   could be transformed to:

   LOCATION;ALTREP="http://feeds.technorati.com/contacts/http:// 
austin.adactio.com/%23six": Six 117 West 4th St

   Note that for cases where there is more than one hCard on a page,  
the fragment identifier is extremely useful (probably necessary).

   * it *could*, in theory, use multipart MIME. Has this ever been  
done for HTTP?

2. How can we specify this in our test suite?

  * MIME CIDs, though ugly, would keep things in one file
  * putting the vCards in separate vcf files. (not sure if ALTREP can  
take a relative URI, though. it should right?)


3. Should we let authors specify the ALTREP URI? Could it just be the  
UID?


-ryan

1. I blame tantek: http://microformats.org/wiki/hcalendar- 
brainstorming#hCard_locations
--
Ryan King
ryan at technorati.com