Attempt to summarize various threads

Here's my attempt at summarizing the current state of discussions and
apparent rough consensus:

1. Is cpim-srv necessary? Yes.

2. Does cpim-srv need to cover more than it does now?  Not as normative
text.[personal comment: incorporating some additional non-normative examples
and scenarios drawn from the recent discussion may be helpful for new
readers to understand how the SRV mechanism is used and/or why it's needed.]

3. Is it OK for there to be zero tuples in PIDF?  Yes, but we need to add
text to the spec discouraging any attempt to use e.g. <note> as a
non-interoperable substitute for status.

4. Should PIDF be split into "pure presence" vs. "IM-related" parts? No.

5. Is there a loop detection mechanism in CPIM?  Yes, Jon will add
"MaxForwards" (or some variant spelling) in the next draft.

6. Should there be a MaxForwards field on MSGFMT? No, MSGFMT has to be
immutable and so a decrementing hopcount won't work there.

If anyone thinks I got these issues badly wrong, please say something now.

I think we end up needing small changes to CPIM and PIDF, plus optional
non-normative explanations in CPIM-SRV.  In the absence of objections from
the WG to this summary, the editors should go ahead and make those changes
so we can move to WG last call on our docs.

Thanks,

--Mark








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