Markdown Version

Session Date/Time: 10 Nov 2026 15:30

[00:01:54] Thibault Meunier: So I guess we're one minute in and we're ready to start. Because I have a small companion for, like like, on Lisa's side. We take a cat, which is joining us, but that's a a side one. Okay. I think we'll start and see how we go from there. So everyone, this is, like, the virtual interim for OCM, the one we have for November. This session will be recorded. So if you don't want to be recorded or etcetera, just, like, be cognizant of that. That said, this meeting is covered by the Note Well. Should be remote. It's still like something that that that we... Like, all participants agree to abide by. You have, like, a longer version of that document, which is available at the QR code. Essentially, it's, like, the various policies that, like, we work through at the IETF, and these concern anti harassment policies, the very select intellectual property contribution that you make because it's an IETF meeting, and so all participation that you may have like part of the IETF process. Some quick notes specifically because we have a lot of people which are remote given this meeting is remote. Make sure that, like, your audio and videos are off. It's, of course, very nice to see all of you. I think it's mostly for the duration of the meeting. During the meeting, you can request and, like, show your hand. It's at the bottom of the screen. You have, like, a... Like, something to join the queue. And when you're in the queue, you can, like, state your name. So you... Like, just at the start when you begin speaking. There's also a chat which some people seem to be already accustomed to, so you can have some participation in there as well. The agenda today is mostly focused on, like, the documents that we have adopted and, like, the one, for which we have a milestone coming soon, which is the OCM open cloud mesh document. We'll present a short status on that. And then we'll have, like, Giuseppe kicking us off, and then we have, like, a couple of points for discussion which are running, like, security compatibility, the collection work that, like, Micke has done, and the distributed directory service, which I think has been some updates, which have been part of the document. At the end, we'll review and see, like, what are, like, the next steps based on, like, the current milestone that we have. Should we, like, provide them or what the authors think on that side? I think mostly that's it. Do people have question on the agenda? Things that, like, they would like to cover. Anything else? Okay. Based on that, we'll go, directly into the document status. So we now have three documents which are working group documents. We have the IETF-ocm-open-cloud mesh, which is still at version seven, which was published in September. So there's been quite a few activities on GitHub and on the latest editors draft based on, like, the mailing list and messages from the authors. These are likely to be published next week, so we have kind of, like, more up to date version to start reviews on that. If you have not read the document, I do think both this meeting and on next week, once we have, like, a kind of more complete version, Review would be, like, appreciated, and any review you can send to the mailing list so that it gets back to the author, and we improve the document. We have tools that are documented that got adopted. We have, like, the OCM MLS federated groups and the OCM integration protocol. Both of these have been adopted by the working group, have still not seen a new version being published on the data tracker, so we're still on the zero zero. Mostly, that's it. There's been some interaction for the MLS document and some of the coordination that I think will happen with the MLS working group, and that's all on the mailing list. If you have some sort, you can participate there as well. Maybe for the authors of the document, anything that I missed, on the status and... Except for, like, the first one, which I think we'll, like, just dive into.

[00:06:38] Giuseppe LoPresti: I think it's all fine. Thanks, Thibault.

[00:06:42] Thibault Meunier: Okay. Perfect. Well, on that, I guess, Giuseppe, I will, like, give it to you.

[00:06:48] Giuseppe LoPresti: Yep. Okay. Okay. Let's see if this works. Okay.

[00:07:03] Thibault Meunier: Pass you control of the slide, and now you should have it. And Yep. You should be able to Thank you.

[00:07:09] Giuseppe LoPresti: I do. Thank you. So, yes, this is a little bit of a a bit of, let's say, status update and and to put the basis of the discussion for for today. So let's start with the assumptions. So the assumptions are the following. We have the main draft. It's already version seven. There will be a version eight really very soon at this point. Let's say this week is... Everything is frozen. Next week is the IETF 121. So then we will publish it. I I think we can upload it at the beginning of the week so that it's there for everyone. But this is the case for the main draft. Then we have the other two satellite drafts and so to... So to what you mentioned. So there has been already some discussion on the MLS part, so the distributed groups. And, okay, this requires more work for sure. And also the other one, we know at the moment we have a a single proof of concept implementation, so wouldn't consider this this ready for being closed as a as an RFC. An important point, which I guess would be part of the discussion afterwards, is that there was a security review. We launched a poll. The poll indicates some support towards going for breaking compatibility, which I must say, I'm... I was kind of surprised to hear that. We just had a meeting, actually. It just happened an hour ago with most of the sites that are actually deploying deploying those solutions for the US Federation. So I made them aware that this is this is what's going on and what is... And, actually, many of the participants. At least some of the participants are coming straight from the other meeting. So thanks for joining, actually. So this will be an interesting part of the of the work because, I mean, if we go for a final call with the OCM 2.0, we have to do a bit of cleanup. So then with that... Those assumptions, the to do list is a little bit like as follows. So we need to make sure that we don't have forward dependencies from the core draft to the MLS and integration-protocol drafts so that we can publish the core draft ahead of the other two. This, I think, is mostly on Micke and myself. I think we can look into that even in the course of this week. This is not such a big deal. Then we have two rather important pull requests. One relatively small, but... And the other one a bit bigger. So the distributed directory service is the bigger one. And we have to also there agree on on what we want to do and decide whether this is okay or not, etcetera. At that point, we publish... I think we publish next week no matter what is the state of those put requests. So that is... And once this is done, the idea was also suggested by Thibault to engage with the group, the directory group, maybe even during IETF 121. I don't know if... I guess Lisa is probably going to attend because it's very close by. Maybe this could be one thing to be to be brought to their attention or or to be see. To be seen how we do that. Otherwise, one important information to... Just to follow-up on from the last meeting. We did get this funding approved, so we are able to work on some of those implementations until the end of the year. And there is already ongoing work on the notifications on the... On those federated groups and the distributed directory service. This is what we plan to do between now and... So it's until end of January. So this is good. I mean, then at least we have, again, a quick feedback to to have a reality check of the specification. And so... Oops. What's next? So this was there already on the last interim. I adapted it a little bit. So, of course, we are going to work on the on the implementations and fix the spec. We have to work on this version too. At least, okay, we should probably kind of stamp this decision today and then people agree that we got this direction or not. And then editorial review and all the rest of that towards the soft deadline that we had imposed to to our working group, which is in a month, essentially, which I confirm... Okay. At the time, I was saying we probably... We can probably deliver all the... All of the drafts. At this point, the target would be to deliver the main draft. And the other two, again, they they require more time, so they will need work even during next year, especially that we depend on on the MLS group as well. And that's all I have. Happy to take questions or to... Well, actually, I hand over to Thibault, and then we can discuss. Yep. And I already see Micke that he is willing to ask questions.

[00:12:49] Micke Nordin: Yes. So I think one important thing that needs to happen for the core draft is to have a few people reading it and commenting on text, finding typos. There are... Yes. I mean, we we found large duplications of chunks of text when Thibault read it the first time. So I think there may be still things like that. So it would be very, very helpful if everyone can take a look themselves and also poke other people they know to to really go through and read the text and propose changes.

[00:13:35] Giuseppe LoPresti: Okay. Thank you. Yes. And Thibault also wanted to... Yes.

[00:13:42] Thibault Meunier: Thibault Meunier. So on the... With the chat on for the HTTP directory, my direct track review, That's something which, like, I wanted to have, like, the latest version, at least, like, published to, like, have something a bit, like, more, like, more substantial track for, like, a new review. You know, like, some specific parts that, like, you would like more attention to. That's also something that, like, you can share. I think... I don't know. Like, we we... Like, we have some idea, but that seems like we'll request on the latest draft, I think. Similar to the I n r o d review, which I think you briefly started, but we could start, like, once we have, like, a new document, I think. Mhmm. I also want to kind of, like, double down on, like, like, like, what Micke said about, like, having more people review and, like, just come up in the thread. I went through the document and I pre will send a bunch of of... I didn't have the time to to to to dive straight.

[00:14:43] Giuseppe LoPresti: Mhmm. Just

[00:14:45] Thibault Meunier: generally, think having more people review to ensure that, like... Even if they're, like, a bit outside of OCM to understand... Do they understand the document? Is it readable? And does it go through? Or even if, like, they implemented OCM and they have more specific blood, like, think would be really useful in this document.

[00:15:05] Giuseppe LoPresti: Indeed. Thank you very much. So, yes, this is exactly... So, yes, I agree also that whatever we get already there, I will I will for sure publish it to the data tracker so that at least we update the data tracker to the current state, which is for sure better than the version that was published, I think, on the beginning of September. So for sure, this is... This can be done ahead of the... I mean, on... Once the data tracker gets opened. Maybe this, I can say also for the benefit of people that are not used to the IETF practices, let's say. So what the IETF does is they freeze the the detractors to the location where all the drafts are are published the week before the IETF meeting. So to give time peep... People to read them without frantic changes. But then on the day of the... When the IETF starts, then you can publish again. So I guess by Monday, I will publish the the new version. So we have Ron yesterday.

[00:16:21] Ron Trompert: There we are. Yes. Hi. About this question, Aaron, about the question about breaking backwards compatibility. Mhmm. I wonder if the the the the people who voted in favor of them realize what they voted for. Because now maybe in this situation where we're now where a lot of EOSC nodes are still in a sort of test build up phase, You you may be able to get away with it. But let's say, a couple of years from now when we have hopefully a few 100,000 users. If you then do something like this that you break backwards compatibility, then, basically, you deliberately breaking the system for for a number of people, at least for a certain period of time when... And where you have to wait until everybody upgraded to the latest upgrade version. So

[00:17:17] Giuseppe LoPresti: But this would actually advocate for doing this right now and not postponing it because eventually, the real problem here is we knew it, but, okay, now it's really written black and white. If you use the spec, you are compliant with the spec, but you use it in a way to chain together all of the bad practices that are still there, your server is open to server side request forgery, one of the classic Yeah. Back

[00:17:45] Micke Nordin: No. No.

[00:17:46] Giuseppe LoPresti: No. So we have to do something there. So it it Yeah. I think at this point, it actually makes sense to to to say this practice is now completely completely forbidden. Like, servers that do that way are noncompliant, and we say it loud and clear in the spec. And then, of course, the vendors can still do can still do backward compatibility. Like, we can still imagine that you can accept a different... Like, the the old format of a share, but then maybe you want to block basic out access, which is one of the less secure ways that we have, you know, this kind of this kind of stuff. Yeah. Micke is actually...

[00:18:31] Micke Nordin: Yeah. I I just wanted to say also that breaking backwards compatibility in the sense of... That we are talking about. It actually more or less just means that c file is no longer being able to do their authentication, but they don't work anyway. They have a broken discovery. So we have put a lot of effort into Nextcloud and OpenCloud and on cloud and certain books to make them all be on this level where we are breaking compatibility below them, but not for them. So so it's not like we're breaking backwards compatibility for recent deployments. It's for one dot zero dash proposal.

[00:19:25] Giuseppe LoPresti: Which is the version that was published ten years ago or so or whenever. I mean, yes. I agree.

[00:19:33] Ron Trompert: Okay. Well, thanks for the explanation. I agree.

[00:19:39] Giuseppe LoPresti: And by the way, I'm happy if

[00:19:41] Maciej Brzeźniak: If I may if I may share the thoughts, Maciej Brzeźniak from PSNC, we actually do run quite considerable installation of c file. But on the other hand, we are actually putting our efforts in in the in the in the spirit of ensuring our EFSS system is OCM compatible. We do actually put our efforts rather into open cloud or similar OCs based platforms rather than into c file because we we kind of consider of the adoption of OCM and c file not being that important for c file itself. Right? So we, ourselves, running these two platforms at least, we tend to think that it will not be in long term supported by C file, the OCM. Of course, this is the perspective of the end user, not the the user of the platform, not the... I'm not representing c file at all.

[00:20:40] Giuseppe LoPresti: Yep. Thank you. Now just for the sake of getting organized, I'm also happy if Thibault wants to share the rest the discussion, or I can keep following the questions. So this is totally... I mean, of course, this was already one of the points. Like... So... Okay. Shall I gather that out of this? It's probably good to do this change to version two now and get it in. Also, I can add for for Maciej. I was in contact actually with the c file CEO. It's already many months ago, so we should update this information. But back in the days, he told me, we don't have enough effort available now to follow you, so we will implement OCM whenever you publish a final a final specification. So they are kind of standing by. Yeah. So with that, in fact, then it's what Micke said. So the other ones that are following, they are all up to date. So they are already able to do so... Yep. It makes actually sense to to do the cleanup now at this point. Now do the chairs see any problem in this extra little bit of work that we would like to do? That is... We'll do... We'll go through everything that is should and make it must, and whatever is deprecated, we really drop altogether. Shall we do that and then approach it a little bit there? Or maybe it's already the time to do this in parallel? What do you think? I'm really open to any option.

[00:22:28] Lisa Dusseault: Yeah. That sounds like a reasonable editor decision as long as we review all of the changes, which we're gonna do.

[00:22:41] Thibault Meunier: Yep. I think just to, like, like, Monisha, like, my impression is that ideally, like, the more constraint you can place, the better. And, like, if, like, we foresee that, like, the next version of OCM just... I don't know. The product might be backward compatible for, like, reason, but, like, in terms of the protocol, if we foresee that, like, the best version is the one which is more constrained, which, like, seems to be the sentiment among the the editors and, like like, to be seen, like, making those changes before publication would be, like, ideal.

[00:23:13] Giuseppe LoPresti: Mhmm. And also then before handing this over to, for example, or to anyone else external, let's say.

[00:23:22] Thibault Meunier: I think, like, it depends where you want the review. Given that it seems that the musts are going to be in the specs down the line, it may make sense to, like, have them before requesting the review. It does not... In terms of, like, ordering, like, what could also happen is we could... Like, the group and the editors could, like, update the draft based on, like, the various, like, discussion that happened in the group and then make a new update, like, version nine with, like, these changes and say, okay. We would like version nine to go through, like, HTTP directory, review for the draft. And that may also be, like, to provide, like, some visibility about the current work to the working group and at the same time, keep on working such that there's a version that can be reviewed.

[00:24:08] Giuseppe LoPresti: Okay. Yes. Thanks. It's okay. Yeah. We could already indeed start to do this this clean up and publish an investment. And then, of course, there will be even more of these kind of cleanups. But, okay, then we can already approach HTTP or whichever other external body or person or whatever that can go review. Okay. Then should we go back to... Thibault had had a nice list of items that we could discuss. So one was the security part. I think we kind of agree. We have a kind of... Can I say rough rough consensus, or is that or is that really completely reserved? Sorry for the job. But it's that indeed we go ahead with with version two and clean up clean up the spec. Yes. Thanks for the, indeed, the list of items. And so the other parts are the collections and the distributed directory service. Those are the two still ongoing things that are are happening. I don't know. Do we want to add to have... Like, I think you can maybe mention where we are with that.

[00:25:31] Micke Nordin: Yes. So for collections, I think we discussed in the last, interim, and we have, talked in the chat and on mailing list and things like this. So it seems everybody is in favor of adding it, and there is a PR, which, Giuseppe and Mahdi, reviewed. And, there there is one final thing, to do, which is the... It's not related to the text itself. It's the open API specification part where I had duplicated the envelope for the notification. So, hopefully, I was able to fix it with the help of GPT Soul, but Giuseppe promised to take a look at the open API specification to make sure that it was good as well. But I think the text is there, and the story is that we want to be able to have some... The possibility to perhaps advertise some kind of datasets that you have in a repository or something like this. You publish a collection ID for this, and then a user can request a share using this collection ID. The collection ID can also be used to collect different shares into the same... Like, have a have a way of connecting different shares, and you can have perhaps multiple... In a multi share, you can have the same collection ID and, yeah, a lot of things like like this. So I think it is good to go once we make sure that the, spec.yaml, is correct. I don't know if anybody has anything else to add about collection and collection ID.

[00:27:48] Giuseppe LoPresti: Maybe just to say, I agree. It's really just a matter of... I didn't attempt to to review exactly that part, but I think it's it's a it's a kind of easy and non contentious change,

[00:28:01] Micke Nordin: to be honest. It's like Maciej has a question perhaps.

[00:28:04] Maciej Brzeźniak: Just a comment that pro... Probably to my understanding, this feature is quite research use cases oriented, but perhaps there are also some business cases which could be, you know, relating to that feature. But I I I I I just wanted to make sure that I'm understanding things properly that this proposal comes from the use... Research use cases.

[00:28:27] Micke Nordin: Yeah. So it comes from thinking about connecting repositories to file sync and share systems. Yeah. So... But it's quite general. So it's basically just a a separate ID which you can use to identify a resource before it it has been shared. So the current ID... IDs that we have is the provider ID, and the provider ID identifies share. So if you share the same document with the three different people, it has different provider IDs, and it's not allowed to have to have the same provider ID for separate chairs. So the collection ID is a way of identifying a resource or a collection of resources ahead of time. So... And you can use that to request a share, basically. So you could think about a lot of use cases for this kind of ID, I think.

[00:29:41] Maciej Brzeźniak: Okay. Thanks for this explanation.

[00:29:47] Giuseppe LoPresti: Okay. And then the other one is a little bit more complex. Is what we call the distributor directory service, which... Okay. There were a couple of conflicts which which were really stupid to fix because, meanwhile, we added things to the changelog. Nothing really critical. Maybe I can say quickly what it is about, but then, I guess... I don't know. We could also also go in the in the discussion of... Because while we are there is I addressed the comments that were put by Richard, by Viktor, and in principle, it should be good. Then I will still need another review maybe from Viktor because there there were a couple of comments from him that I'm not sure they were all addressed. But for the benefit of everyone... So the point of this of this change, addition is to expand what used to be this appendix c, the directory service, which is provided externally, and make it such that you can have either that, I mean, that stays as a possibility, or different servers can agree between each other by sending notifications and by having the administrators of the servers to agree that they become part of the same federation. So this is vetoed by people, by administrators. And we kind of introduced now the concept of the administrator of our NCM server to to make a server part of a game federation. And there is a third version, but that would be part of the MLS specification, but use the MLS groups. So this way of sharing groups across servers with a cryptographically guaranteed way to distribute information, and we would use that also for for the federation of servers. So the third one, that definitely requires more work because it first needs the the MLS federation to work in the first place. The middle one would be the distributed way to do a directory without relying on a central location that holds all the all the servers of a game federation, which goes a little bit like what... I mean, what we have demonstrated... This was already about a month ago for the use symposium. This will be for the benefit of the chairs. We demonstrated a full federation in full... In... Fully working in production, but the list of sites was served from one of the one of the nodes kind of out of band. And we just agreed that, okay, we... For the moment, we confuse it that way. But, of course, this is not very satisfactory. So this can be improved. And the poor request is all about trying to detail this this protocol. There are a couple of contentious things in this one, and, actually, it's also a bit longer. So I don't know. We can go into into into the details of of this one in particular, or we take this also for for afterwards. The last comments were all about already... We use the word federation. This I I agree indeed. I I add the idea of overloading the word federation, which used to be a federated group of people. So people belonging to groups spanning multiple servers. What eventually wants to do also for for the use federation, so having some kind of distributed group or distributed identity identity provider. But then we are also talking about a federation of servers. So then the the idea of this was Micke that proposed it, I I agree is to have federation refers to the server federation, and we need another term. And... Okay. Then this should... We can even see what you think live for for a meaning distributed groups. So they distributed some in the sense of a group of people, a group of users. So the poll is open. I propose to have it short, but I really didn't like too much, and Micke doesn't like it, so I'm happy to change it. D group, just to shorten distributed group. I'm not very good in names. So anything would do. But, okay, we need something there. Yes, Micke.

[00:34:21] Micke Nordin: Yeah. I I mean, I don't have... I don't think it's too bad to have a d group inside of the JSON file, but it's not it's not great. It doesn't roll off the tongue. So if somebody could think about another term for for a distributed group, I think that would be very well... Welcome. We were... I mean, there there are other things which sounds weird too, like cluster or syndicate. You could think think about other words, but none none of them are really used in computer science or in this way. So maybe coming up with a novel term to mean distributed group is not that easy, and it's fine to call it a d group inside of the payload and then just call it distributed group in in text.

[00:35:14] Giuseppe LoPresti: In the UI. Yeah. Because, anyway, this... We are always talking about internal payloads, server to server payloads. So whatever goes to the to the UI of the different systems, it's a completely different business. Actually, Nextcloud, this is called the circles. Yeah. Okay. Using circles to me doesn't tell much, to be honest. So I wouldn't really reuse the Nextcloud terminology here.

[00:35:38] Micke Nordin: They are rebranding to team, I think, as well. So

[00:35:41] Giuseppe LoPresti: Oh, wow. Okay. But then team... Yeah. Okay.

[00:35:47] Micke Nordin: Yeah. Think if you can come up with a good term for distributed group, you would be my hero.

[00:35:55] Giuseppe LoPresti: Yes. Any any suggestion is good is good for me, to be honest. But I think we all we all agree. I mean, we all... At least there is some consensus that we prefer to have different keywords at the payload level to indicate the federation of servers, and this is where we use federation, versus a group of people spanning multiple servers, and this is what at the moment is called b group. And having said that... So indeed, out of the different observations in the pull request, the part that I did not address is how do we deal with the institutional problem with this kind of distributed distributed systems. It's how to guarantee consistency if you re... If you revoke membership. And for that, I think we will need to have a dedicated discussion to see how to represent this in a way that makes sense. So this is still pending. The rest is good. Okay. It's also a relatively large change. It's about thousand lines of text in the pull request that is in the agenda. I'm going to it in now. So then I'm... Yeah. I'm happy to have any comment on this, but... Okay. I can imagine also some people never saw it or or seeing it for the first time. So you need a you need a bit of time to digest it or even to parse it via via your favorite MLM, if you want. Yeah. And if we go back to the agenda... Let me see. I think at least in terms of the main items, yeah, Thibault.

[00:38:00] Thibault Meunier: Yep. So just want to interact, at in terms of, like, the items. Do people have specific feedback, opinions now? And if they don't have them now, feel free to let go, like, on the list to discuss, I guess. And, like, for... Like, the naming, I guess, it's totally fine if, like, the editors have an opinion because, like, this is probably not the the most picky point. But, definitely, like, if you have some more comments feedback, my understanding is, like, the PRs on GitHub are likely to be merged in a reasonable time frame and and and be part of, like, the next version. So if you have any comment, just go to the mailing list and that that that that that should be the best place. This meeting is kind of not a decision meeting. It's really to help foster the work that goes down.

[00:38:53] Giuseppe LoPresti: Thank you. Yes. Indeed. That's exactly same also for the other for the other things to be to be resolved, let's say, this pull request. Okay. Maybe one point for everyone is still that I think we are... We have some consensus to go ahead with having this also part of the spec in first in first go, like, in the in the... Because the... What what I call the peer announced. So so the the fact that we use just notifications, which are part of the call spec without depending on the MLS part, this can go ahead and it... There is... It's not depend on the other part of the spec. Then the other part of the spec will bring the MLS level federation. Okay. And this will come in a few months time.

[00:39:42] Thibault Meunier: Yep. I guess that's a similar comment to everyone we made last time. Ideally, the document that, like, we push does not depend on document which have not been completed or submitted. Yeah.

[00:39:54] Giuseppe LoPresti: Sounds good.

[00:39:55] Thibault Meunier: I put my hat off on the personal comment on, like, this one for, like, the addition of MNS. Does it make sense to add the directory service if we don't have the spec for MNS yet? Is it something that you would like to push maybe down the line and integrate kind of a... As a standalone document down the line? That's TBD. I don't know. That's, like, really, I think, something that might be useful to consider.

[00:40:19] Giuseppe LoPresti: Totally valid question. Indeed. Let's let's maybe think about it a little bit. My my point was more like I kind of don't like to have the spec as it is at the moment where we only have this little appendix at the bottom and we rely on this central component. So having a... Having something which can already provide a way to do... To to to distribute the the service seems interesting to me, independent from the fact that you can also implement it with MLS. And then you have the two possibilities that are both valid. Micke?

[00:41:02] Micke Nordin: Yeah. I think the other way to do it is to have a single MLS flavor of a federation and push it to the MLS document. But I think it's it's fine to have a more low... I mean, more flavors of the federation, let's say. It's it's okay, I think. And... But make your make your voice heard on the mailing list, I guess, what you what you prefer. I I think we can move the directory service somewhere else. We can keep it. There... It's a lot of the questions that pop up with the... Just the distributed vanilla flavor is about how do we make sure that everyone agrees, And all of these questions are solved in the MLS case because we have a a better foundation there to know when we agree about... On something, and this is trickier when we're not using MLS. But... Yeah. I I don't have a strong opinion myself if we keep it in the course pack or move it to its own document or to the MLS document or however we want to to do it. I think we can get it in in the shape now before working group last call, and then, it should be fine. But, I'm also fine with dropping it, out of the core and moving it some... Somewhere else. It is a very useful feature to have some way of doing this federation. So we we should definitely have it at some point, at least.

[00:42:49] Giuseppe LoPresti: Thanks, Micke. I'd concur that having multiple flavors also may help, especially especially given that in reality, I mean, this also for the benefit of of everyone. We know very well that we have the US Federation, but we have a soon it is itself a federation of how many 50 something nodes?

[00:43:07] Micke Nordin: 54.

[00:43:08] Giuseppe LoPresti: Serve is another federation of many tens of nodes. So it's something that that kind of is intrinsic in in OCM to to run a federation.

[00:43:25] Micke Nordin: I I have a a question for the shares, as well. So about the reviews and the next steps. So we... Currently, I think we have one milestone for publishing the document before in December of this year. So can we open new milestones for OCM MLS and OCM IP in next year? And how do you view the continuing work on those?

[00:44:08] Giuseppe LoPresti: Indeed. I let Lisa and I mute myself.

[00:44:12] Lisa Dusseault: Yeah. So, I'm not forgetting, am I... That the, putting out the MLS, as a separate piece, it seems like a good idea, but we haven't called consensus on it. We we can certainly make that part of a minor recharter. I don't expect it to be very difficult, but we'd It was adopted. Propose some text and and ask for consensus.

[00:44:42] Micke Nordin: Yeah. So it was adopted already by the working group on the last interim.

[00:44:48] Lisa Dusseault: Then we... Okay. So it's...

[00:44:50] Micke Nordin: The... Think it's right.... Matter of May... Creating milestones, I think, in the data tracker and deciding how we should work with them.

[00:44:57] Lisa Dusseault: I mean, we don't technically need to have milestones to make... Continue making work. So maybe we just continue working and focus on that.

[00:45:05] Giuseppe LoPresti: Yeah. Okay.

[00:45:21] Thibault Meunier: Yep. I think I would just add on, like, like, what Lisa said. I think for the current milestone that we have, my understanding of the, like, current meeting that we have today mean that, like, pre will, like, postpone a bit. Like, like, milestone given, like, end of the year seems a bit early, as far as I understand, to, like, have the document in a good shape. We could postpone it, let's say, like, March or April. I think for that, it would be good to have, like, the input of the editor and see, like, how much, like, reviews we see and how much work, like, you foresee on that side. On the OCM MLS and OCM IP, I would, like, concur, to to to to to to to what Lisa said is my impression is for now, like, the group is working on it. Are there specific milestone that are defined? If you have some specific one that, like, you're thinking of, feel free to, like, comment in terms of, I don't know, like, the momentum or etcetera. My impression is for now, like, the group is working on them and we'll... Can define the milestone when we have a bit more clarity on these ones. And maybe just to add, if at some point you feel it's useful to, like, gather, like, either more input, you can go down to the mailing list if, that would benefit from on a broader design, like a shared discussion with, like, MNS if, like, that into it. Again, we could have discussion on the mailing list, discussion, like, as an interim or, like, the... Like, some discussion like that. I think all of that is something which is possible at the moment. The the the the thing to be... That the work is continuing, but we don't have a clear milestone to assess it.

[00:47:05] Micke Nordin: I was also... Ron, may he suggested that we request early review on OCM MLS from the MLS working group. So I think we... When we publish the next version, we can send it to them. But I don't know how... Is there a procedure for requesting a formal review, or is it just a matter of writing to their mailing list? And

[00:47:28] Thibault Meunier: I think you just go and say, hey. The OCM work group is working on that. We would value input from the MLS people and just go with it.

[00:47:36] Micke Nordin: Yeah. Okay. Thanks.

[00:47:45] Giuseppe LoPresti: Okay. That's also good that... I mean, those things go in a very informal way. This is this is great.

[00:47:52] Thibault Meunier: Just one thing before the... Like you said, sorry for interrupting. Sure. It seems like Maciej had his... Like, hand raised. Just want to to to change...

[00:48:02] Giuseppe LoPresti: Ah, yes. Sorry. I didn't I didn't --... Or... Remark should

[00:48:05] Thibault Meunier: I have.

[00:48:12] Giuseppe LoPresti: Or maybe this was forgotten from before. The hands raised. Maciej, are you still around?

[00:48:18] Maciej Brzeźniak: Oh, that was just accidental. I'm sorry. I'm not good. I'm not familiar with this tool. So ignore my hands, please.

[00:48:25] Giuseppe LoPresti: Okay. No problem. Plus, okay, I must say that myself, I I unfortunately have leave a bit earlier than I expected because there is something totally standard that is coming up. But, okay, fine. I think as we stand, I don't think there's anything really big or still outstanding. So from your last message, Thibault, you are also kind of advocating it even for the main draft. We delay the... Starting the process of the last call anyway by a couple of months. Is that right?

[00:49:07] Thibault Meunier: Like, my impression, at least from what was discussed in the mailing, is, like, the editors. And, again, that... I think something that, like, you will, like, either discuss in the mailing or, like, send to the chess is that the document at the moment is... Might not, like, be ready to to be... Like, might not be ready for last call by the end of the year. So if you have, like, a better idea about, like, the timeline, I'm happy to to get it, then we can, like, update the milestone accordingly.

[00:49:36] Giuseppe LoPresti: Okay. Thanks. Let's see. I mean, in principle, we might still get there. Then also, I have no idea what is the typical time for the last call itself. In fact, actually, Andy Newton told me back back in ITF one twenty six that this might go relatively fast depending on... So we we don't want to rush it. That's for sure. Right? Okay. Okay. Thanks, Lisa, for for the comment. Okay. Then it's also fine... I think it's also fine with us in the end. Like, we we delay a little bit more. The December deadline was a bit... It was a soft deadline. There is a... And especially, it's true that if we start to embark now into this cleanup of the legacy practices and editing cleaning, We can focus on all of this in the coming month and then start to do a proper review after that. Okay. So with that, indeed, I won't have anything more, but then I would hand over to the to the chair, especially that indeed, unfortunately, I would have to quit a little bit earlier. So thanks everyone for this attendance, which was at this point a little bit also wider than... To defend it more than last time. So that is appreciated.

[00:51:05] Thibault Meunier: Thank you. Yeah. Thank you all. And, if you have, like, any comment on people that you know will comment, feel free to, like, prompt them to the mailing list to to bring feedback.

[00:51:18] Lisa Dusseault: Yep. Exactly what I was about to say.

[00:51:25] Thibault Meunier: On that, yeah, see you all online. Thank you.

[00:51:29] Giuseppe LoPresti: Thanks. Thank you. Everyone. Bye bye. Bye.

[00:51:33] Micke Nordin: Bye bye.

[00:51:34] Giuseppe LoPresti: Bye bye.