[uf-discuss] A couple of real-world usage problems

[Tantek Ç elik]

On 7/11/06 2:07 PM, "Ryan King" <ryan at technorati.com> wrote:

> On Jul 11, 2006, at 6:18 AM, Ben O'Neill wrote:
>> On 11/07/06, Mike Rumble <mike.rumble at lawton.co.uk> wrote:
>> 
>>> What I'm getting at is does the user-agent or search engine see
>>> these a
>>> two separate events? If not, is there value in using hCalendar in
>>> both
>>> places?
>> 
>> The spider/browser should see they have the same uid and therefore
>> know that they are the same event. It should also be clever enough to
>> add any info it learns from the more detailed version to the knowledge
>> it learnt from the less detailed.
>> 
>> That's how I hope it would work, but I'm not familiar with the
>> hCalendar spec.
> 
> Your instincts are right. UID is a universal identifier for an event.
> Two events with the same UID should be considered the same event
> (assuming there are no trust issues).

Agreed.  Though we will have to figure out how to treat two events with the
same UID but found on different URLs.

Which content is authoritative if the two conflict?

I believe the short answer is to use class="url uid", that is, use an actual
URL to a definitive version of the event as its UID, and then you can always
resolve conflicts between two sites posting the same UID by actually
*loading* that UID (since it is also a URL) and getting the definitive
details there.

Thanks,

Tantek