I have another "bad idea" regarding being overcommitted: Keep the commitment, but change the timeline. If we grouped work into 4 broad areas, and then work on 3 of them per IETF, (with some allowance for critical work in the other area) we can avoid both "DNS is always discussed in North America" and "we never have time to discuss DNS" because we will be doing 3/4 of the current work at each IETF, but take 1.25 times as long to get to the output. I really don't see how we can alter the burden without either: * introducing more parallelism (which nobody wants) * introducing more hours (which nobody wants) * introducing more meetings per year (which nobody wants) or changing the expectations of timelines, and making it easier per meeting, because we accept less can be done per meeting (which maybe nobody else wants, but I want) I think Christian Huitema's work on the timeline to produce an RFC is informative. Nothing happens inside a year, most things take between two and three years, some things take decades. If we accept this, then deciding to not talk about SNMP every single meeting, but talk about SNMP in 2 meetings out of 3 in any given year, would reduce our burdens by the cost of parallelism to talk about SNMP at a meeting. Again, I probably should stop inventing things. But this is what I think: we should accept the reality and work on less, per meeting. -G
