--On Monday, December 28, 2020 20:53 -0800 Benjamin Kaduk <kaduk@xxxxxxx> wrote: >> Kind of agree, but really why do we need to consider the >> local time, at least for some of the cities we are >> (virtually) going to have very few local participants. > > I don't have the answer for you. My understanding is that > SHMOO is chartered to consider topics such as this, but > regrettably I personally don't have time to spend thinking > about it, given my other commitments. > >> I will not do any meetings that is before 6am or after >> midnight. I think it would be better to try the 6 hours that >> is reasonably hard for the participants that get the hardest >> deal, maybe even find two alternate 6 hour slots. But they >> need not have anything to do with the "local time". > > It's your prerogative to place those hard limits, sure. (109 > was 9pm-3am for me and I managed, but I am very fortunate in > my home environment.) But my understanding is that for any > 6-hour block you pick, "the participants that get the hardest > slot" will have that slow overlap with their midnight-6am > window, and it's far from obvious that we want to always give > the hardest slot to the same limited set of people. Hence, > having a WG to consider it... Ben, I agree with most of this. Despite inconvenient (one might say pathologically so) times, I managed to get to most of the sessions at IETF 109 that I considered really important. WHen I look at this thread, I think I see three groups of people: (1) Those who are going to attend all of the sessions relevant to them, independent of time zone. (2) Those who will get to most or all of the sessions that are really important to them, but will skip anything that, in a f2f meeting with all of their week dedicated to the IETF, they might have dropped in on based on anything from interest (but not high-priority interest) to vague curiosity. That, along with WG sessions that are not scheduled during IETF, is almost certain to have a negative impact on cross-area review. I am guessing, but just guessing, that few people attend online interim meetings of WGs out of vague curiosity either. (3) Those who simply will not attend sessions that are sufficiently inconvenient or problematic vis-a-vis their "normal" timezones. That, of course, has a negative impact on everything unless either IETF meetings and/or cross-area review are not important or there is no different in expertise and perspective between those who do not attend a give session or set of sessions and those who do. What concerns me about your suggestion is the "have a WG consider it" part. In my experience, WGs are successful when they start with clear criteria for success and, ideally, with some drafts about which there is general agreement about principles even if considerable work is required about specifics and details. Having a significant number of people who are determined to get the work done, get to a conclusion, and get one or more documents out is often critical too. When we have a vaguely defined problem and no agreement about principles, pushing things off into a WG seems to very nearly guarantee going around in circles, spurts of activity and long breaks in between, and discussions among those who feel most strongly about the topic (or who have too much time on their hands). If there is an outcome, it is usually one that ends up being attacked either during whatever passes for Last Call (including "Consultations") or, sometime after a decision was supposedly made, on the IETF list. This situation, the previous evaluations of meeting times, and the experience in SCMOO seem to be rather good illustrations of this. As an example of the principles in this area on which there is no agreement, consider whether it is more important to optimize for maximum attendance, maximum attendance by important people, or maximum geographical or other diversity. From observing the discussions of meeting location principles over the last several years and particularly the last one, there appears to be a clear conclusion: we don't have anything resembling consensus. And no amount of "lets charter a WG to look at that" is going to change that lack of consensus not matter how effective it might be at postponing the decisions and/or shifting responsibility. best, john That analysis may not tell us much
