# Routing area open meeting minutes ## Agenda: - Agenda Bashing - Updates from the Area Directors - Updates from the Routing Area Directorate chairs - Routing Area Working Group agenda items + Discussion - Presentation from Sue Hares ## Agenda Bashing Adrian Farrel: Routing Area open meeting conflicts with the security area open meeting. Please try not to conflict with other area open meetings in the future. ADs: Good point, noted! ## Updates from the Area Directors John: welcome Yingzhen Qu as new cochair of LSR. Thank you! Andrew: Adrian Farrel has agreed to serve as interim chair of MPLS. Thank you! John: we keep experimenting with the format of this meeting. In the past, we used to have every working group report status every time. These reports didn't seem to be that helpful, it was just a version of reading their document list. Later, we shifted to having a few working groups report in detail, rotating between working groups at different meetings. That was more satisfying to me, anyway. Prior to this meeting, we considered whether to do that, and decided not to, since we've had detailed reports from all of our working groups fairly recently. All of this is to say, if you have any suggestions for how we can make your area open meeting work better for you, please let us know! ## Updates from the Routing Area Directorate chairs Andrew: introducing Daniam as area directorate co-chair, and thanking Luc Andre for his past work. (see slides, Daniam presenting) John: if you're going to reject an assignment, that's OK, please reject as soon as you know you won't be able to do it. ## Routing Area Working Group agenda items + Discussion (Clarification point, the title refers to agenda items in any working group in the routing area. The potential confusion with the full name of RTGWG is unfortunate.) (see slides, Andrew presenting) Jim: discussion during the working group session is so much more meaningful if someone other than the presenter knows what the presentation is about! Anthony Somerset: this seems like common sense. Shouldn't this apply to all areas, not just routing? John: we're the routing area directors, we don't get to set policy for other areas. Valid point, though. Jeff Haas: sometimes work from Group A is presented in Group B. Example, Segment Routing thingy needed to be encoded in BGP-LS, "we're presenting this thing, it's a trivial encoding, who from IDR is following this work in SPRING?" John: I think the meta-point is... I at least don't want to make this such a stern policy that chairs can't exercise discretion. But even in the case you just raised, that seems like something that's foreseeable. Sometimes you have a diving catch -- it's Tuesday, you're at Group A, and you suddenly realize the work should be advertised at Group B on Thursday. That's fine, the ask is even if it's a liaison from another WG, please flag it to your mailing list. Jeff: The thing that makes it challenging, is sometimes when we're flagging items, we don't know that the thing being flagged is work being done (or NOT done!) in another WG. Example, "we're being asked to stick metric foo in the Segment Routing TLV, suddenly someone on the SPRING side says 'what are you *doing*, we don't do it this way, it doesn't belong here!" This is obviously a working group chairs coordination problem, but it overlaps this subject. John: let's discuss more off-line. Adrian Farrel: +1 to what Jeff said, and +1 to SHOULD not MUST, and maybe on "*a* mailing list" and not "*the* mailing list"? John: That covers a lot but I'd still like to see it flagged to the list of the WG in question. Adrian: Yeah. Possibly what underlies this, is when constructing the WG agenda, work out what needs to be talked about and why and how, and then all this goes away. John: Yes. This is a constant tension. If we say "chairs, try to be good chairs, as long as you do that we don't need rules". OTOH having chaired some things, that wasn't always helpful to me. Sometimes it was desirable to have, if not rules, then very firm guidelines to help me do my job. That's what we're attempting here. That's also why we are having this as a conversation here, and not just email, in case we've gotten it wrong in some way. Adrian: Thanks, dad. Tony Li, TVR co-chair: I have the opposite problem. People in my WG don't do much on the list, they're much happier commenting in person. It's a chicken and egg problem. How do I get around this? John: ... chair harder? But seriously, yes it's a problem, different communities have different ways of working, and if your community works in person, you're the chair, use your discretion, kind of what Adrian said. Andrew: These are guidelines, the real goal is effective use of agenda time. Roland Bless: I understand the idea behind it. RTGWG should be open to new work, it could become harder if you're bounced because no discussion. There should be at least some slots even if they're only 5 minutes, to make people aware. It works much better in person than in email. Email is too easy to ignore. Raising interest is not that easy. I'm in favor of prioritizing the WG's "real work", go ahead and prioritize chartered items, but don't exclude the possibility of new work. Andrew: I hope that by doing this we might actually create space for the kind of work you're talking about, by pushing back on work that isn't getting discussion. We'll confer more about this and keep your point in mind. [AI] Lou Berger: Roland just said one of my points, I agree, we don't want to exclude new people/thoughts/ideas. I'm a little worried about phrasing of second bullet (Second bullet is: "Where working groups want external presentations from third parties that aren't drafts, the slides for these presentations should be sent to the working group within the same time frame.") I'd say modify it to say "new drafts" or "new drafts and external presentations". Andrew: I think they're two separate points. Should also address new drafts, you're right. Lou: Separate is probably better. I'm not in a lot of WGs that have time for external talks. Allowing for new drafts/contributors is critical. John: The spirit behind this (and we may have got it wrong) is two things. One, with 00 drafts (not brand-new ideas necessarily) don't submarine your idea til the draft cutoff date, then publish and expect to get useful work done a few weeks later at the in-person. Everyone is deadline driven, and this is a way of saying "let us help you be deadline driven in a way that's more productive for everyone". Lou: I love the idea that new work that hasn't been discussed needs to be submitted X time ("a week") in advance, so there's time for adequate review before the meeting. John: We'll have a follow-up at an upcoming chairs chat, but this is not just a chairs meeting, it's an area meeting and potentially touches all contributors which is why we wanted to float it. [AI] Lou: I'm sold on that part. One added point, Chris Hopps mentioned they're (in LSR) running an experiment where they'd stop the time early in order to leave a few minutes for discussion. I ran with that and have run three sessions with that approach. I really like it. "Steal" a few minutes of time so that the timer expires early and it leaves space for discussion. Jim: On the 00 drafts, with a 00 draft, the first time the WG hears of it should't be in the meeting. Even just an announcement that the draft exists and plan to present it, is to some extent "discussion". Lou: That makes sense. And I really like the idea of "send your *presentation* a week before", lower barrier to entry than reading a draft. Sue Hares: does this policy restrict interim meetings? Sometimes we've had a new drafts interim, it's announced as such. Andrew: thanks, that's a new point. We'll discuss amongst ourselves. [AI] Still seems helpful to announce. Acee Lindem: we used to have a separate 00 deadline. Another change I'd like to see is for any document for LSR (whatever group) to go to the list. John: So... the LSR list would get anything that matches on ".*lsr.*"? Acee: Yeah, something like that. You can come up with a regex that'll do the job. [...] We have a lot of people who are passionate about working on drafts, usually *their* drafts. What we need are people who are passionate about reviewing drafts. (To be fair, we do have some excellent reviewers.) John: Datatracker does have a regex-based watchlist feature, use it through your dashboard. Not sure if it will auto-email you too. My only concern with your suggestion is it's heuristic, maybe you end up with label switched routers hitting the LSR list. For all I know, datatracker already has this feature. Andrew: We'll check in on the tooling. Yingzhen Qu: For RTGWG, we get a lot of 00 drafts, because of our dispatch function. We try to prioritize WG document updates first. Brand new work may get "if time allows". For new authors, it may be hard for them to promote, communicate their idea on the list. Chairs can work harder with new contributors. I'd prefer this be a recommendation and not a hard requirement. John: It's a very firm recommendation. Maybe we should have an off-line conversation with you. I would think even a very green author ought to be able to say "we published this draft". Yingzhen: Yes, we'll do that. Lars Eggert: Datatracker power user. There is a "send me email" button on the Datatracker watch list feature. Tony Przygienda: The first bullet is a hare-brained idea. (first bullet is: "drafts should only be allowed on agendas if they have had discussion on mailing lists prior to the working group agenda cutoff.") I can't force a discussion. Second bullet, (Second bullet is: "Where working groups want external presentations from third parties that aren't drafts, the slides for these presentations should be sent to the working group within the same time frame.") I'm all for it. Material should be provided timely. If you want more participation, make the bloody authors list longer. It's the simplest way to get more engagement. John: That's interesting. What do others thing? Would removing backpressure on author lists get us more engagement? Tony: At least be more generous. Look at old ATM forum specs, look at modern science papers, people want recognition. Andrew: I also find that interesting. If you have more feedback like that, I'm interested to hear it, so are the rest of the IESG. Please let us know how you feel. Tony: Or you could make it one author per company. I'd say just make it a page. John: We don't participate as companies. Tony: That's another discussion. Lars: We've had situations where the front page was used to indicate corporate support. It was many more than five, on a document that had fewer than five pages. We could go back there. It's a continuous discussion. John: On that draft, was there a tremendous amount of engagement when discussing it? Lars: I don't know, wasn't in my area. Andrew: I've seen cases where people on the front page have clearly not read the draft. There's a tension, we're constantly discussing it. Tony: You can go both ways, it takes value away from author list padding. Or maybe it also makes author list padding even more competitive? We could look at git history of draft development... though that can be gamed too. And there should be consequences for gaming it. There's no perfect solution. Acee: for consensus building and combining drafts, the five author limit makes no sense. You have already gone through a lot of effort to merge two drafts, then push everyone off after that big effort? I'd rather fight with the IESG then reopen the consensus that was so hard to build. Andrew: Current document says if you have more than 5 you can appoint an editor or get an exception. If you go beyond 5 you need a justification, I think merger is a perfectly suitable justification. I think it's problematic when the justification is "because the authors wanted it that way". John: that's a quote Andrew: There are certainly cases where the number should be more than five, nobody denies it, but it needs to be justified. David Lamparter: If by making it easier to be on an author list of a draft/RFC, that could have the effect of working against people getting bonuses for authorship, which I think is a counterproductive practice. John: I think you just said, if we allow everyone to print their own money, then we'll have rampant inflation. Lars: The word "author" implies writing words. It means you've written text, it doesn't just mean you're on the front page. We do have cases where many more people than five have done the writing work, that's a good reason. But we have the limit for a reason and we need to think about whether to change it. Tony: Nothing wrong with that, but let me give you an example of another way to do standards. You didn't just rev out the drafts. You had a set of documents, people had to present the diffs they wanted, then you took a vote, and it was totally clear who had contributed. John: is your point that in principle, we could make it traceable, trackable, who sent text? Tony: Yes. Of course everything can be gamed, but we can make it more trackable. John: Thanks for this great conversation. We will need to digest and iterate. ## Presentation from Sue Hares John: and now for something completely different! Title: Do's and Don't's of evaluating your AD (Sue presenting, see slides/video) Andrew: Thank you, that was really interesting. I really appreciate the data-driven approach, will talk with you about it later.