During the meeting yesterday we made plan how to go forward. The plan
is summarized below:
----------------------------------------------------------------------
- Confirm during the list that we do not have consensus now for week,
after that continue this process if so
- Timeline
- Aug: Working group to review current drafts and send comments
- Sep: A full but not final specification
- For tcpcrypt: means including all the changes resulting from WG input
- For TLS means a complete profile for TCPINC
- Oct: Provide comparison of the proposals
- Nov: An implementation that matches the specification that the
implementers claims that matches the above specification
- If both fail to deliver: we close the WG
- If one fails, we choose the other one as starting point
- If both deliver,
- We do a consensus call, for 2 weeks
- If no consensus : we do a coin toss (something else?)
- If the room consensus on this way forward is not confirmed in the
list, we close the WG.
----------------------------------------------------------------------
So that will mean that during this week we (chairs) will make straw
poll on the list to confirm that we do not have consensus about which
draft to pick as starting point for the tcpinc.
Then during the August work group members should read the current
drafts, and send comments, so draft authors can update their draft as
needed. For TLS based draft it would actually be more useful to get
one update as soon as possible to include proper profile to be used in
the tcpinc, so WG have some time to comment on that.
Before the end of September, the draft authors will submit the "full
but not final specifications". For the tcpcrypt, that will mostly mean
addressing comments raised by the WG, for the TLS it would mean proper
profile (hopefully we will get that earlier, but this is the latest
time for it).
Then during October, we and draft authors can compare the proposals,
and try to summarize what is the main differences between them and
what are pros and cons for each version compared to each other.
Finally before the November meeting, we would like to see some kind of
implementation that matches the specification. This does not mean
production quality code, but more or less proof-of-concept type of
implementation. Mostly just to show that the specifications are in
such shape that they can be implemented based on the text provided.
These will most likely come from the draft author teams, but if
someone is willing to make independent implementations that would be
best...
Then in the November meeting we will try to see if we have consensus
on which draft to pick as starting point, and we will not have
consensus, then we will pick one by coin toss (or some other suitable
method).
This email is just to confirm in the working group that we do have
consensus about this way forward....
--
[email protected]
_______________________________________________
Tcpinc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/tcpinc
