RE: Communication means in draft-ietf-impp-cpim-pidf-05

> > But if those systems are interoperating, it would be bad if they applied
> > those same display techniques to the information received from the other
> > system.
>
> How could they do otherwise? The act of inserting a gateway
> between the two systems doesn't change the behavior of the other
> components of those systems. Or are you saying that as part of
> establishing interoperation, each system must enhance its
> intrinsic model to embrace the concepts of the other? That seems
> much less likely to happen than the installation of gateways.

I'm not expecting them to "embrace the concepts of the other." I'm saying
that one of the consequences of our inability to resolve a "single right
way" of partitioning a presence system is that a presence UA has to take a
"fuzzier" view of how the foreign presence system represents the world.

I don't think this is a huge problem. Even if we worked out a "single right
way", I would still expect presence systems to have various fancy
non-interoperating features for "native" users of the system.  So there
would already be some distinction in the UA between "native" and "foreign"
presence information. I was just noting that this is an area where it's
quite easy for people to believe that they are using the "obvious, natural"
approach to partitioning between presentities and tuples, and to assume that
foreign systems will partition likewise -- which will lead to problems.

I'm just noting that all you really know about presence information from a
foreign system is what's there in the message, and that user agents should
avoid ascribing additional semantics that are true for their local
implementation but might not be true elsewhere.

--Mark




  [reminder: [email protected] for non-technical discussions, please]