[uf-discuss] UID in iCalendar

[brian suda]

I do remember awhile ago we discussed using ID as UID, i can't find the
archived message at the moment. The one concern was that UID is to be
globally unique, whereas the ID attribute is only document unique, so if
we are going to construct an implied UID, then we could look to the RFC
recommendation and do something like:

<id>@<host-name>

where <id> is the fragment identifier for the page and <host> is the
exact page where the <id> was found.

For example,
http://events.example.com/#123
would become
123 at events.example.com

and

http://events.example.com/europe/#123
would become
123 at events.example.com/europe

we need to take the full URL so that there are not collisions in the UID
when there might be IDs called 123 on two different pages.

I need to now look to see if/what needs escaping in the URL string
(slash '/' needs to become '\/')
-brian

Tantek Çelik wrote:
> Agreed with Brian's interpretation.
>
> In addition, I think if we make a stronger suggestion (perhaps a SHOULD)
> that individual hCards and hCalendar events have a unique-to-the-document
> 'id' attribute on their root elements, then automatic construction of
> globally unique UIDs becomes fairly straightforward, and we can write an
> "implied UID" rule for hCard and hCalendar at a minimum.
>
> Thoughts?  
>
> I'd like to add "SHOULD use 'id'" and "implied UID" to hCard and hCalendar
> soon if no one objects.  And yes, we should update the creators so that they
> auto-specify uniqueish 'id' attributes on the root elements as well for ease
> of authoring.
>
> Thanks,
>
> Tantek
>
> On 7/3/06 12:17 PM, "brian suda" <brian.suda at gmail.com> wrote:
>
>   
>> According to the RFC 4.8.4.7 Unique Identifier, the only to requirements
>> are:
>>
>> 1) Value Type: TEXT
>> 2) Globally Unique
>>
>> Any URL fits the TEXT requirement. As for globally unique, all URL only
>> resolve to a single location, so that makes it globally unique.  The
>> only caveat is that if you have two events, those each need to use a
>> unique URL, not just http://events.example.org, but instead
>> http://events.example.org/event/1234
>>
>> The recommendation is just that, a recommendation, not a requirement.
>>
>> -brian
>>
>> Dimitri Glazkov wrote:
>>     
>>> Can UID in iCalendar be just a URL of the event? Do we have to adhere
>>> to the RFC2445's recommendation
>>> (http://rfc.net/rfc2445.html#s4.8.4.7), which is based on an
>>> email-style RFC822 addr-spec (http://rfc.net/rfc822.html#s6.1)?
>>>
>>> :DG<
>>> _______________________________________________
>>> microformats-discuss mailing list
>>> microformats-discuss at microformats.org
>>> http://microformats.org/mailman/listinfo/microformats-discuss
>>>
>>>       
>> _______________________________________________
>> microformats-discuss mailing list
>> microformats-discuss at microformats.org
>> http://microformats.org/mailman/listinfo/microformats-discuss
>>     
>
> _______________________________________________
> microformats-discuss mailing list
> microformats-discuss at microformats.org
> http://microformats.org/mailman/listinfo/microformats-discuss
>
>