Session Date/Time: 05 Aug 2026 14:00
[00:00:06] Joe Hildebrand: So are we having a meeting today?
[00:00:15] Mikolai Gütschow: I would have assumed so.
[00:00:18] Joe Hildebrand: According to Barry's latest note too, I don't know. I was just looking to see if there was anything on the mailing list from any chair. Yeah. I didn't see that either. It's strange that no chair, no Karsten. Yeah.
[00:00:38] Вадим Гончаров: But Karsten's on vacation at least.
[00:00:41] Joe Hildebrand: Yes. Well, it would've been nice to get a cancellation so I
[00:00:45] Вадим Гончаров: don't have
[00:00:45] Joe Hildebrand: to get up early.
[00:00:48] Laurence Lundblade: Yeah. I agree.
[00:00:51] Вадим Гончаров: Agreed. I was also quite in a hurry to get to this meeting. So still a notification for us.
[00:01:04] Laurence Lundblade: And probably give him a few more minutes Yeah. Before we give up.
[00:02:51] Joe Hildebrand: Laurence, I am planning on doing a review of your doc. I have not yet.
[00:02:58] Laurence Lundblade: Okay. Thank you.
[00:03:07] Joe Hildebrand: I'm interested to see how this if anybody puts in the work to to want to cut out anything out of, EDN.
[00:03:31] Вадим Гончаров: Barry is showing online at least in notes taken.
[00:03:39] Laurence Lundblade: Oh, here's Andy.
[00:04:04] Joe Hildebrand: I'm guessing we're waiting for chairs.
[00:04:06] Laurence Lundblade: Yes. We are.
[00:04:09] Joe Hildebrand: Yeah. I just sent a note to the list couple of minutes ago saying that. Yep. Christian, do we, are we having a meeting? It looks like he may not actually be here.
[00:07:35] Вадим Гончаров: I think he's still setting up. If I knew I'd better left in bed sleeping instead of too much coffee.
[00:09:06] Joe Hildebrand: Was there a published agenda for the meeting? It was terse.
[00:09:19] Laurence Lundblade: Yeah. No no specific, presentations were submitted or anything like that. It was just, I believe it was just open for discussion on the the usual topics.
[00:10:01] Joe Hildebrand: So I suggest a lack of quorum. Yeah. I'm gonna have to agree with that. The yeah. Sorry, everyone.
[00:10:21] Вадим Гончаров: Nothing concrete was focused. Probably, we want to hear from chairs about next procedures, about EDN.
[00:10:49] Laurence Lundblade: I believe working group last call on serialization ends in a week or two. So I may not be able to attend next meeting. I'll be at an airport at the time the meeting is happening, so I'm not sure I'm gonna be able to participate. So I this seems like the last opportunity to talk about serialization before the end of the last call. I don't know if there's anything anybody wants to say, but but I guess we can take that on the mailing list. That's fine.
[00:11:35] Joe Hildebrand: And I thought Christian's message about what to do on EDN was quite clear. So I don't know that we need to discuss that as long as everybody's read the mail on the list.
[00:11:59] Вадим Гончаров: I did not read the draft yet. That's final version. Does it include the section on ALDR or not?
[00:12:14] Laurence Lundblade: It does not call it ALDR. It discusses what to do for determine deterministic at the, data model level. So there's a section in there on that. It's been there, for probably six months and hasn't really changed. I I think it's appendix b or c.
[00:12:58] Вадим Гончаров: I mean, for example, I had an example in least several months ago, maybe even previous year, about specific kilogram session where the CBOR object maybe is only without specific map k called access hash is considered to be equivalent for purposes of hash. This goes beyond standard data model assumptions. So this is truly a application layer application layer determinism. So maybe so I guess it's not included. Probably, I need to contribute to paragraph or two about such complex cases.
[00:14:05] Laurence Lundblade: Yeah. I don't think that is included. I didn't understand everything you said there. But the the approach is, you know, to point out that you determinism is needed at the data model level as well as the serialization level, but it's trying to keep the text to be somewhat minimal. I mean, there's a lot of appendices. So and I and I didn't think it was necessary to be normative or put it in the main body of the the the text as it's not you know, it's it's just something to think about, if you're designing a protocol. And I didn't think that we needed to invent a whole new terminology and all of that for it.
[00:15:17] Вадим Гончаров: Also about the last month's discussions. So did you post to another working groups, mailing list questions about what they expect from serialization or not. Maybe some answers were given. I mean, that example of C509 from when the part begins not from the document start.
[00:16:03] Laurence Lundblade: Yes. There's some text in the latest draft that allows for that or, you know, mentions it it's sort of indirectly. It's not I mean, it didn't go on too long about it. But there is some some the intention is to accommodate that. And and, again, this is just advice to protocol designers. It's not anything normative.
[00:16:45] Вадим Гончаров: Also, I thought that such problem could probably be solved in generic CBOR libraries by providing an application. An application gives CBOR pointer to start of area and end of area to and the library should return offsets. Such solution may be better and more than I have sketched in list about some deep level about a race on a bundle protocol example. But the point is currently an expert draft. So and we have no Rohan today.
[00:18:01] Paul Hoffman: Well, greetings, y'all. Deep apologies. And so yeah. And I just got here, so I don't know what has happened so far.
[00:18:17] Laurence Lundblade: Not very much. We've had a little bit of discussion about the serialization draft.
[00:18:23] Paul Hoffman: Good. Because that was the main topic for discussion. Laurence, have you gotten sort of what you need or want from the discussion today? It seems like we're doing fairly well on the mailing list.
[00:18:42] Laurence Lundblade: I mean, there's I mean, I think the there's not been any, in my opinion, anything major or substantial. There's a few little comments about things here and there. Yes. But but that's that's it. And I, I mean, I can think of a couple of little a few sentences to to add or tweak in the draft to address a few of those things, but, they're all independencies, non normative stuff. And Yeah. So I I'm not feeling like there's anything much to worry about here.
[00:19:26] Paul Hoffman: Good. Well, again, we're we're in the middle of working group last call. We're, you know, we're not done or anything.
[00:19:33] Mikolai Gütschow: Yes. Yes.
[00:19:34] Paul Hoffman: My reading of the list is the same of the the comments that have come up on the list is the same, which is can you talk a little bit more about this, but all in the, as you said, dependencies all in the side stuff, not in the main serialization stuff. And again, I apologize for being late, but was there anything in the last twenty six minutes where people had brought up any major stuff on dependencies and such? I'm sorry. On on on the main topics? No. Okay. Good. So oh, and I am now seeing that Christian is having audio problems, which is why he was not there. And again, I apologize for showing up late. Is there anything else that we should be using this time for to discuss in serialization? I'm feeling that the list is doing fairly well both in bringing up things as well as having a discussion about them, but we are here now on this. Is there anything else on this that we should be discussing here?
[00:21:01] Вадим Гончаров: Well, I've asked about probably, I I didn't maybe paragraph or two about example for from ALDR from CD based on Telegram example. What and that generic libraries can use Cboe pointers in the API for specifying which offsets to return to application for the incursion side and but this likely is not a normative parts anyway.
[00:21:46] Paul Hoffman: And this is related to your decoder stack requirements thread from earlier?
[00:21:56] Вадим Гончаров: This relates to questions Laurence raised in list about complex cases, like, in bundle protocol or Yeah. C509, I thought
[00:22:09] Paul Hoffman: Wi Fi bundle. Back on the bundle protocol. So let me ask you a question, Vadim. Do you agree that the answers to that when Laurence puts those in are not changes to the normative text?
[00:22:31] Вадим Гончаров: Sorry. I'm did not read the last the version of draft because I thought here today will be about EDN. But as far as I know, API questions are not normative in Okay. IETF.
[00:22:51] Paul Hoffman: Great. Great. Good. Well, then let's keep that discussion happening on the mailing list, the the working group last call on on this on the mailing list. It seems like we're doing well, and I don't mean to say that, oh, we're done because not everyone has read it. I I will do a nudge and such like that. Given that, we can't switch to EDN except for the fact that, you know, we just started the Christian just started the call yesterday for controversial features and such, and we haven't seen any yet. And again, that's because it's limited time. Do does anyone here have at least a short list of the ones they are intending to bring up in response to Christian's message? Vadim, I seem you're see that you're still on. Is that to talk about those?
[00:24:03] Вадим Гончаров: Those two, I was yeah. About serialization draft, such contribution about several paragraphs My reference CBOR-pointer of draft from, but this draft is in expired state.
[00:24:27] Paul Hoffman: Ah.
[00:24:27] Вадим Гончаров: Well, the the we can discuss this at least.
[00:24:31] Paul Hoffman: Yes, please.
[00:24:32] Вадим Гончаров: It's informative reference anyway if it will be included in non normative. So okay. Let's serialization in this. For EDN, today, I all day, try to remember all those discussions happening in June. Too much Right.
[00:24:57] Paul Hoffman: Well and, again, let's have those on the list, and and let's see where where we come out with the what is the list of things that people want to remove. And, again, we just started this yesterday, so I think it's a little bit premature for us to assume that we have stuff to discuss today here on the call. The Christian did ask for, you know, how to how people should respond, and I think that that will certainly help focus. We did not put a deadline on this, and I think that that's reasonable just because it's summer and many people here are Europeans, and they believe in taking vacations, which we Americans don't. So I believe that what we should do is schedule a call again for two weeks from now, and hopefully, we will have a much better picture on the EDN document of what has been brought up on the list, or maybe nothing. At this point, now that Carson has done the grand reorganization and things are clearer, there may not be features and such. Anyone want to comment on that? Vadim?
[00:26:26] Вадим Гончаров: Yeah. I have procedural questions about is each such prop feature proposed to be removed must be sent in separate mail message.
[00:26:39] Paul Hoffman: Yes.
[00:26:42] Вадим Гончаров: So so many messages.
[00:26:44] Paul Hoffman: Okay. Correct. And that is for that is to help other people track. You might have five things you wanna remove. You might even feel that they are related to each other. But if you put all five in and then someone goes, well, I wanna remove three, but not these other two or such, It will make the chair's job infinitely harder. So the part of this is to help the chairs, like, decide what needs to be removed. And we know that some people don't want anything removed, so we wanna be fairly careful in this. So, yes, a separate thread on each part, as Christian said in his message, will make it so that even though you might know that they're all related or you want them, it will help the other people in the group be able to comment. Because just because one person wants to remove something doesn't mean we're gonna remove it. We need to have rough consensus. Joe?
[00:27:49] Joe Hildebrand: Yeah. You were asking who's who's gonna ask for stuff to be deleted. I think I'm going to start with ellipsis
[00:27:57] Paul Hoffman: Okay.
[00:27:58] Joe Hildebrand: And suggest I'm gonna start with let's remove the tag for ellipsis and make a separate argument about removing ellipsis itself.
[00:28:11] Paul Hoffman: Thank you. That would actually be good to have those. And, again, we're not in a rush. We are not under, you know, a hard deadline. So if you wanna do those in parallel, that's fine. Or if you wanna do one first, see how the reception is, and then do the second one, that's fine as well.
[00:28:29] Joe Hildebrand: I think I might prepare two separate messages and send them at the same time.
[00:28:35] Mikolai Gütschow: Okay.
[00:28:35] Joe Hildebrand: But I wanna make it clear. I'm I'm trying to separate I'm thinking about, like, how do I start a thread that has the least, like, the the most focus to it so we can make a crisp decision about just that one thing.
[00:28:48] Paul Hoffman: Thank you.
[00:28:50] Joe Hildebrand: And so I'm using ellipsis as an example of how I'm trying to think through that. Does that make sense?
[00:28:56] Paul Hoffman: Yes. Absolutely. And that's just fine for you to do that. In the same way, it is fine for Vadim to pick other ones that may be big or smaller, whatever. Let's see how the threading on the list goes. Mhmm. And if and if it starts going to hell, the chairs will say something. But I I like the idea your idea of trying to before you put it up, trying to think what will get the best focus and and then move from there.
[00:29:28] Joe Hildebrand: Right. Now there'll be some interaction between those. Like, if we Yep. If we say yes to this one and no to the other one, it won't make sense.
[00:29:37] Paul Hoffman: But Correct. That'll be okay. And and and if you say that in them, that is still fine to have the two separate threads going. Yeah. Because at some point, you might say, look, the you know, this they are both working, or one of them's not working and therefore the other is not. What would be lovely on the list, and I'm saying this not just to Joe, but to anyone, you know, who's going to be suggesting things, is if you put something on the list and it gets no traction or it gets a lot of negative responses and you realize, oh, this isn't that important, please feel free to come back and close off the thread yourself and say, oh, this isn't, you know, like, wanted this, but it seems like no one else wants it or whatever. Don't rely on the chairs to do that. I mean, we will, but I consider that to be the chairs are supposed to help determine rough consensus. We're not supposed to be driving rough consensus. So if you all, like, see that in your own threads, feel free to add to the thread and say, this doesn't seem to be going very well. Or, gee, this has a lot. You know, not to try to close out the thread quickly, but to place a marker and see what other people on the list believe. Mikolai?
[00:31:06] Mikolai Gütschow: Yeah. Just for clarification because you just mentioned negative responses. Should we voice on the list if we do think that a feature should should stay, or should we just silently not say anything?
[00:31:22] Paul Hoffman: It depends on how strongly you feel. That's a very good question. So for example, if someone has proposed to remove a thing that you are sure is fundamental to something you are doing, not just, oh, it might be nice in the future, feel free to come in and say, I hear that you wanna remove this. This is why I want it here in the current case. There are a bunch of things that I think even I mean, Carson's not here, and I don't feel like speaking for him, but I think he has said in the past are would be nice to have features as compared to are really necessary. And someone might say, well, let's take this feature out. It's just would be nice to have. And if you have a direct need for it or something where that if if EDN the current EDN doesn't have that feature, it's going to negatively affect you. You can say that as well. Don't go ahead and don't start the conversation that way, but read so those of you who are suggesting, I think, to take this thing out, you can also say why. You can say, this is this seems like it's just a nice to have and we don't need, or this actually seems harmful because it interacts with this other feature this way. You know, please do say why you are suggesting something. And then to go back to Mikolai's question, if you disagree with somebody who's asking to take it out on the level you know, again, nice to have is just nice to have, but if you have a real use case for it, please say so.
[00:33:31] Вадим Гончаров: These procedures so that Yes. We must only answer to this
[00:33:38] Paul Hoffman: Yes.
[00:33:38] Вадим Гончаров: Message and only
[00:33:42] Paul Hoffman: So let's keep it framed that this way for now until we have figured out what is coming out, if anything. At that point, the the discussion will open again for the draft as it stands at that point. So let's say we take two features out, four features out. Carson will then take them out and issue a new draft, and then we will have a complete draft there. And then it will be open again to especially, again, I don't like to speak for Carson when he's not here, but he seems to really wanna touch the ABNF. Those discussions will then be open again for that in now that we have a reorganized draft and then in the future with the limited. So that should get everything that then we will open discussion again. Joe?
[00:34:38] Joe Hildebrand: Yeah. I'd first off, I do think that there needs to be a few, at least, minor tweaks to a b and f before we're done. So we'll we'll get to that. Let's figure out what features we're supporting and then mess with the a b and f. Makes sense to me. Secondly, I'm planning on trying to reiterate as many of the arguments for keeping something as I can in my message
[00:35:04] Paul Hoffman: Okay.
[00:35:04] Joe Hildebrand: In order to forestall, endless back and forth of but you don't understand why this is important to me.
[00:35:13] Paul Hoffman: Okay. So you're saying that you will put those in your message to remove. You will say, I am proposing to remove this, and I've heard these things before about keeping it, but I still want it removed.
[00:35:28] Joe Hildebrand: That's right.
[00:35:29] Paul Hoffman: Excellent. Thank you. That that would be super helpful, again, like you say, to reduce traffic, but also to make sure that, like, if someone says, yeah, I said that I want it before, but now I see your argument. I am willing to lose it or whatever. I think that that that will be again, maybe I'm being a bit selfish here. I'm trying to help the chairs determine rough consensus, and having you all do that would really help.
[00:35:57] Joe Hildebrand: Yeah. And I did also the third thing I wanted to say here, which is that in, Christian's message, he said refrain from pointing out disagreement on the list or any corresponding PR. The arguments have been had, pointers to feature interactions that were missed or appreciated as our clarifying questions if interactions are missed or be unclear after removal. So what you just said slightly in contradiction to what the message said. So I'm just, you know, you can change. Just wanted to make sure that you guys were aligned.
[00:36:32] Paul Hoffman: I read that as please don't just reargue. I read it more so I I believe that saying it or the other way I read it was, please don't start a thread saying, don't remove this. It's important to me. So Okay. Let's yeah. And and but I Joe, I hear I hear that you have that that what we're saying may be slightly different. Christian and Barry and I will try to get together in the next day or two to, you know, do that. Keep go ahead and start with the messages. We will figure out more about that. But thanks for pointing that out.
[00:37:16] Joe Hildebrand: So yeah. Just as long as you guys are on the same page and have a a a plan in place, I'm happy. Yep. Thank you.
[00:37:23] Paul Hoffman: Yeah. It we have somewhat of a plan in place. Yes. Vadim asked, was anyone taking notes? Well, I certainly wasn't taking notes because I wasn't even here. I think we will probably rely on the AI automatic transcripts for now. As much as they make me gag, a lot of people really like them and they do seem effective in places. So I think we will go with that. Any other questions for today? Not seeing any. And, again, I just wanted to say I am very happy with the the how the working group last call is going on serialization. It's not silent, and it's not rehashing old things. It's definitely bringing things into focus for a couple of people. Let's please keep that up, and let's see what the next few days or weeks bring with the questions on EDN. Thanks for coming. And, again, one more time, sorry about spacing it out. My calendar is not the best. I am putting it on for two weeks from now so that I will see it ahead of time. Thanks. Take care.
[00:38:56] Вадим Гончаров: So Alright. Anybody will write down today's summary to know IETF or or or not?
[00:39:05] Paul Hoffman: I will I will try to do that, Vadim. Yes.
[00:39:12] Joe Hildebrand: Okay.
[00:39:53] Вадим Гончаров: Okay. Bye bye.