RE: [Simple] About tuple ids and namespaces in presence data model

["Dag Ekengren" <dag.ekengren at hotsip.com>]

Jonathan Rosenberg wrote:

>>
>> If service tuple were required to have contact uri:s, I guess the
>> tuple-id would be unnecessary. But that would be incompatible with
RFC
>> 3863.
>
>No, even if you have contact URIs, the server still needs to see them
>differently. I could be attempting an override, in which case two
>different PUA would publish tuples with the same service URI. Even if
>you ignore that, if you wanted to allow a publisher to change the URI
>and consider this a change in tuple, as opposed to a new one, you'd
>still need ID. Though I suppose there is hardly a reason to do that.
>
>We don't have an id for the <device> element, which would seeem to
imply
>that the device-id+etag form the handle. For services, its id plus the
>etag. I'll make this explicit.
>

Well, I think that the contact uri would be sufficient to identify the
service tuple within the presence document, just as the device-id is. Of
course one would have to keep in mind that there may be service tuples
in other documents with the same uri (that may attempt to override this
tuple). This should be no different from the existence of several device
elements with the same device-ids, but in different presence documents.
Changing the contact uri would then create a new tuple and delete the
old one as you say.

I realize that it is valuable to keep the id to keep things compatible
with RFC 3863.

>
>
>> 2.	The new elements "person" and "device". I would like to put a
>> note on the device, would the note element be from the pidf namespace
or
>> do you imagine that this element would be added to the
>> urn:ietf:params:xml:ns:pidf:person namespace? RFC 3863 says that a
tuple
>> element can have optional note elements, but the person and device
>> elements aren't tuples, at least they aren't called <tuple>.
>
>THis is a good point. I would argue that the note in the presence
>document that resides outside of the tuples probably "belongs" to the
><person>, even though its not within it. Thus, there would not be a
need
>for another note element. I think thats consistent with current usage.
I
>will add text describing this.

Ok. I would also like to be able to use a <note> on the device. I think
it is valuable to be able to provide a textual description such as "my
mobile phone", "IBM R40" or "Home PC".

Thanks,
Dag



_______________________________________________
Simple mailing list
Simple at ietf.org
https://www1.ietf.org/mailman/listinfo/simple