PCN met Friday afternoon, 1 April, from 2:15 to 3:15 pm. Scott Bradner
presided.

minutes by Tom Taylor

NOTE WELL was presented.

Agenda accepted.

Status

Edge Behavior Drafts
=====================
Tom Taylor

For O & M sections, use our judgment on how much has to be said. Do not
have to go into details of MIBs. Editor will try to turn around within
two weeks.

Encoding Comparison
===================
Georgios Karagiannis

Regarding appendix: not a good idea to replace with a reference to an
Internet Draft.

Gorry Fairhurst. Suggests changing "comparison" in title to "overview",
which is closer to truth. Useful document.

Signaling Requirements
=======================
Georgios Karagiannis

Agreed to use "IP addresses" on issue 1
Issue 2: agreed to keep addresses
Issue 3: agreed to leave 2.3.2 based on Georgios's suggested
interpretation
Issue 4, 5: clarify
Issue 6: OK
Issue 7: use proposed text with RFC 2119 language
Issue 8: will rephrase
Issue 9: OK
Issue 10: question is whether some text on security is needed. Agreed
that this would be redundant. Philip Eardley suggested that the whole
of 3.3 is redundant to the rest of the document. Discussion of
structure.  Georgios will review.

Will WGLC when update submitted.
Philip Eardley says he has more comments.
Georgios: aim for WGLC in two weeks.
Discrepancy between slides on web and slides as presented. Problem is
slow update -- new version of slides submitted this morning.

3-in-1
======
Bob Briscoe

Draft already last-called, but some additional text needs
consideration. 
Need also to consider "updates 5696" implications.

Need to have a little text saying that 3-in-1 is now the standard. 
Experimental use of the codepoints is possible but out of scope. Agreed
on the additional text for e2e ECN.

Ready by the end of the month. Toby takes the token.

Alternative Edge Behavior
==========================
Michael Menth

A number of questions for clarification.
Tom Taylor recalled the piggybacking proposal -- same signaling, but 
decisions based on measurements. May be difference in variance.
Question of whether policy can be applied in termination -- yes, can 
exercise some discretion.

Looking forward: Georgios Karagiannis sees as potentially interesting. 
Philip Eardley noted difficulty in completing current work.
Bob Briscoe (also noted earlier): would have to PCN-mark signaling
class -- has implications.

Scott: agrees with Philip that we should concentrate on current work,
but discuss this one on mailing list. Can confer with AD depending on
outcome.