Session Date/Time: 22 Oct 2026 14:00
[00:00:04] Daniel Petrie: Okay. Try to get people to sign up.
[00:01:31] Jonathan Rosenberg: Hey, Dan. Can you hear me?
[00:01:33] Daniel Petrie: Yeah. Hi, Jonathan.
[00:01:35] Jonathan Rosenberg: Hey. Sorry about this. I'm I'm stuck on a bus on the way to New York City.
[00:01:40] Daniel Petrie: Oh.
[00:01:44] Jonathan Rosenberg: Hey, Dan. Hey, Jonathan.
[00:01:46] Daniel Petrie: Hey, Scott.
[00:01:50] Jonathan Rosenberg: So, Dan, before we get started, one quick thing. From our last meeting, I still have an open action item just to send you a PR to add redaction to the course specs.
[00:02:00] Daniel Petrie: Yes. Yes. I I would love Yeah.
[00:02:03] Jonathan Rosenberg: You're gonna have that either Monday or Tuesday of next week.
[00:02:07] Daniel Petrie: Awesome.
[00:02:07] Jonathan Rosenberg: Awesome. Yeah.
[00:02:09] Daniel Petrie: Yeah. Okay. Don't don't worry about reviewing the rest of the doc. I think, you know, it it... You know, I think it's good enough for the next for the meeting, but I would really love to have the redaction tagging in there.
[00:02:21] Jonathan Rosenberg: Alright. You got it. Let's see. I've just... Otherwise occupied till Monday, but as long as that works. Monday, Tuesday.
[00:02:28] Daniel Petrie: Yeah. That's that's fantastic.
[00:02:33] Thomas McCarthy-Howe: I just have this picture of Jonathan being in the back of a Greyhound bus.
[00:02:38] Jonathan Rosenberg: NJ Transit, man. New York City is a disaster with traffic today.
[00:02:46] Thomas McCarthy-Howe: Well, that's too bad because I could've called you the rambling man. Yeah. Remind me to make an introduction for you, Jonathan.
[00:03:00] Jonathan Rosenberg: To To whom?
[00:03:02] Thomas McCarthy-Howe: To a man that we met down at the Vcon show at Atlanta who's Alright. Kind of an itching guy. I've been... Life's probably with me.
[00:03:19] Jonathan Rosenberg: Yeah. Actually love to hear how that went.
[00:03:22] Thomas McCarthy-Howe: Well, yeah. Well, I I know Dan Dan was there too, but I thought it was great. I think we counted 40 independent companies that had Vcon ambitions, announcements, opinions. There was some really good dense... I think it was a very dense crowd, like not just being stupid, but not a lot of fluff there. And I thought there was some... You know, I I thought there was some particularly interesting demos that we were seeing that are Vcon enabled. I thought we're really, really neat. Yeah. What do you think?
[00:04:03] Daniel Petrie: Yeah. No. It's good. It's... You know, man, I think it's the biggest concentration of discussion of of Vcon. Right? I mean, I think it's bigger. There's more discussion there about Vcon than there is in the IETF in terms of people. Which
[00:04:17] Thomas McCarthy-Howe: is kinda funny.
[00:04:18] Jonathan Rosenberg: Yeah. Yeah.
[00:04:19] Daniel Petrie: So anyway, which is, you know, I mean, I... When I... When any place where I talk about Vcon, I always evangelize and I always try to get people to join the, you know, the IETF list in in meetings. Yeah. But I I... You think it's... You know, Jeff's doing a great job at getting getting momentum, and it's, you know, it's been steadily increasing. You know, he did the first one last April, and and he's... You know, it's been or, I mean, a year ago, April, I guess it was. Right? And and he just... He's he's getting some good momentum, but I I I did a talk on, you know, all the changes in this last edit. Right? I went through, talked about call topologies. I talked about, you know, kind of all all the things that you... We now can enable, and then I went through the kind of three examples that I've talked about here. The, you know, the the case where the the... With the the agent and the... And and the supervisor, you know, whisper sidebar or whatever. I talked about the, you know, a compound transfer case and showing how, you know, all the transfer events that you can capture. And I talk about the, you know, the the simple, you know, kinda AI agent to AI agent and AI agent tool scenario that I, you know, I discussed two weeks ago here. And so I... It it was well received. I got a lot of comments that Yeah. People enjoyed
[00:05:51] Thomas McCarthy-Howe: Yeah. I heard that too, Dan.
[00:05:52] Daniel Petrie: Understanding me.
[00:05:53] Jonathan Rosenberg: So... Yep. Heard that too.
[00:05:55] Thomas McCarthy-Howe: And and, you know, and there's there's four phone companies of of size that are
[00:06:01] Daniel Petrie: Yeah.
[00:06:01] Thomas McCarthy-Howe: You know, waving the Vcon flag. Crescendo was. Alianza was. Gamma was. And so was
[00:06:14] Daniel Petrie: Carrier... Carrier X?
[00:06:15] Thomas McCarthy-Howe: Carrier X five.
[00:06:16] Jonathan Rosenberg: That's five. Yep.
[00:06:17] Daniel Petrie: Yeah. And and Yeah.
[00:06:20] Thomas McCarthy-Howe: And a couple resellers.
[00:06:21] Daniel Petrie: Another one. Yeah.
[00:06:22] Thomas McCarthy-Howe: Yeah. We have a couple of huge... Couple of large resellers. You know?
[00:06:25] Jonathan Rosenberg: Yeah. That was my next question. Who are the big names there? So let's let's hear that those guys were there. Yeah. Maybe the AI people?
[00:06:32] Thomas McCarthy-Howe: Yep. Yep. A good a good, you know, dozen different AI vendors and
[00:06:39] Jonathan Rosenberg: Any big names is my question that I would recognize.
[00:06:43] Thomas McCarthy-Howe: So Modulate was there, and SoundHound was there, and ConSig was there, and Guava was there, and Vappy was there. Jonathan Rosenberg was there.
[00:07:02] Jonathan Rosenberg: Yeah. K.
[00:07:04] Thomas McCarthy-Howe: And as far as resellers go, triple eight VoIP was there, and Intelisys was there. And I guess as we're as we're rolling it down, also
[00:07:17] Daniel Petrie: So dial pad was there too. Yeah?
[00:07:20] Thomas McCarthy-Howe: Dial pad was there. And I... After I burned the poor guy on stage, I I found him I found him afterwards. I apologized, and he sounded very interested and welcoming about integrating to VCONS.
[00:07:32] Jonathan Rosenberg: That's awesome. Oh, wow. It's a pretty good turnout. Yeah. It was. Well, they can Yeah. More people there than ITF because it's like cakes. More people wanna eat a cake than bake a cake.
[00:07:47] Thomas McCarthy-Howe: Apparently so. Yep. Original thing coming up that we should probably talk about is, at least mention, is a policy summit in DC in the first week, and the the minority FCC commissioner is gonna be speaking at it, you know, along with, you know, some of the usual, you know, Glenn Richards and the usual, you know, suspects. But the fact that she's gonna be speaking is, I think, is really interesting.
[00:08:20] Jonathan Rosenberg: Yeah. That's cool.
[00:08:21] Chris Wendt: Mhmm. Yep.
[00:08:25] Daniel Petrie: Alright.
[00:08:28] Jonathan Rosenberg: We get started? Or
[00:08:30] Thomas McCarthy-Howe: Let's start.
[00:08:31] Chris Wendt: I saw I saw your note, Jonathan. I guess you're connected but not... But still mobile.
[00:08:38] Jonathan Rosenberg: I'm on a NJ Transit bus. I'm on I know you...... Open tunnel. So I cannot speak too much, and I cannot speak loudly. I don't have access to my laptop.
[00:08:49] Daniel Petrie: Yeah. Can you see slides okay?
[00:08:53] Jonathan Rosenberg: Yeah.
[00:08:54] Daniel Petrie: Okay. So I'll present and and then I won't expect any comments from you. I'll do all the controversial stuff now.
[00:09:06] Jonathan Rosenberg: Oh, it's better to wait thirty minutes, but okay.
[00:09:13] Daniel Petrie: See if I... No. This is mostly summary. This is not... I don't have a lot of... You know, I I don't... I only have, like, one kinda new issue that I think is interesting, and I I hope that it's not controversial. Should we get on? Or you wanna... You know, we we need to...
[00:09:33] Henk Birkholz: Do we need
[00:09:34] Daniel Petrie: to make the the blue... The, you know, the Yeah.
[00:09:37] Chris Wendt: I guess
[00:09:38] Daniel Petrie: we normal disclaimers and stuff.
[00:09:41] Chris Wendt: The note file and all of that good stuff. Right. Thanks everyone for coming. We've gotten a big crowd. Bigger crowd than usual. That's good.
[00:09:58] Daniel Petrie: Go over it then. Yeah. Next slide. Or... Oh, I have control. Sorry.
[00:10:03] Chris Wendt: Yeah. I gave you control.
[00:10:04] Daniel Petrie: K. Sorry. So what are we gonna talk about? I'm gonna talk about... I made a few changes, not anything huge since the... You know, our last meeting. I'll I'll try to go through at least the high level what what those changes were. I'm gonna go. I I had a list of open issues last time. I've kind of... And some of those, actually, I didn't have written down, but last time, but but I think I've captured those and, you know, I've got some updates on those and a couple... Just couple new issues. And then then... Sorry. I didn't update this slide completely. You can ignore the the end there about the examples. That's just out of date. So changes. So I think it was really clear after our discussion and, you know, after, you know, going through the... Doing the the prior set of edits that we now have a clear concept of contributors versus receivers. And, you know, the contributors go in are the parties that are listed in the dialogue, and the receivers are the parties listed in the in the session. And so I made a pass at trying to, you know, really, you know, make that point throughout, where we talk about parties, in in the different context. So I... So there's bunch of edits to try to clarify that and make that point, you know, kinda very, very clear and and hard. I did some similar edits. There... You know, they... It it calls... There's a section in there that would calls out email and talks about the... You know, we've got the different SMTP fields. Right? We've got from to c c and b c c. And, you know, with the prior point of distinguishing contributors and receivers, I, you know, reedited that that section. Right? So, basically, the from goes in the dialogue parties list, and all the other ones, you know, go into the to the session. And I also talked about, you know, the prior... I'll get that in the next slide, but the... Or one of the next slides. But in in the prior meeting, I talked about... Sorry. I dropped the dropped the ball at my point. I I talked about the issue of, oh, do we have... You know, do we... You know, with with text messaging, do we do these, you know, joins and drops every time we send a message with a different receivers list? And I I talked about, you know, how we use sessions to do that. Right? Not not drop and join. It did some clarify... Clarification that there was there was... I I... You know, when reading through and reviewing, there were some ambiguities about dialogue... Text dialogues, whether they contained one message or not. There was some just weird language in there. I kinda clarified that that, you know, a a text dialogue contains a single message, you know, from from the contributor. The... And I fixed... Oh, the the JSON schema was broken. I did a a... Some cut and paste error in in the last iteration, and it was it was... It just broken, so I fixed that.
[00:13:51] Chris Wendt: Dan?
[00:13:52] Daniel Petrie: Okay. Yeah.
[00:13:53] Jonathan Rosenberg: One of the issues from last time was how do we model an email thread? And whether that was, like, if there was a session, whether each message was a session, and how to model, add and remove the people on the two and cc. Was that sorted in this in the latest
[00:14:10] Daniel Petrie: Yeah. I made an attempt at at that. Although, I... I'm not sure I totally nailed the thread concept, but I I did try try... I'd talk about how, you know, within within a thread that the, you know, the the parties involved changes, and and I I wrote on how to do that. I didn't really talk about the higher level content with thread. That probably needs a little more work there.
[00:14:42] Jonathan Rosenberg: Yeah. And especially if there's probably multiple ways to do it, and they all be valid, you know, message per message per session, session per unique set of senders and receivers, you know, session...
[00:14:56] Daniel Petrie: That that I did talk about. Right? You know, the lower level. I didn't talk about the con... You know, a, you know, potential container for the set of sessions, right, for a thread. I didn't I didn't discuss that. I I think we still... You know, we need another iteration and defining that, you know, the session relationships. Right? That I just made a quick pass of that, and I need some more thought and that needs more clarification on on the cement and and the concrete semantics of what those relationships are and things like that. So that's not done. I think this... There's, you know, kind of a quick proposal there, and it needs work. I I I just hadn't had the cycles to to think that through.
[00:15:46] Jonathan Rosenberg: Okay. That's fine. And as... Again, just to bookmark work on next, we should also think about, you know, if you email an AI agent, the AI agent does a bunch of tool calls and processing it and the process of conducting a reply, making sure that the VCOM can capture those things much like we're talking about for chat and voice AI agents. It's just a good exercise to make sure. Again, just book the smart ones and models in there. Yeah.
[00:16:14] Daniel Petrie: Yeah. I have, you know, kinda... I I I I... At some point, I wanna go through and, you know, kinda... I need to do a little more reading and catching up on, you know, the AI agent, the AI agent, the AI agent, the tool documents to to make... To get a better handle on all the stuff that we're trying to capture there so that I can, you know, come up with a few, you know, deeper examples than the simple one that I that I, you know, went... Walked through two weeks ago. So open issues resolved. So I I I already kinda covered most of this, you know, text type dialogues. I did did a bunch of clarification on this stuff. I think I've talked about everything on this slide here already. So I'm just gonna move on. These are these are the opening set of issues that that I presented last... Two weeks ago. Most of them are still open. I did add a few more things that I kinda talked about but didn't have in this list last two weeks ago. And it's the the second and third items there. You know? You know, this issue of session start and duration have to encompass, you know, that of all the dialogues. In the text, says that I'm not sure, you know, that we want to constrain it to that, and that needs more thought and and and research on that. And so... But I I just wanted to write that down as an issue. And then...
[00:17:58] Jonathan Rosenberg: Do do do we wanna discuss those, Dan, or you wanna wait till you're through the slides?
[00:18:03] Daniel Petrie: We can. I mean, if if people have thoughts, I just... I haven't had I haven't had the, you know, the thought cycles to have much thought. But if if anybody has comments or thoughts on that, I'm I'm happy to happy to to take that input at this point. But I I just... I'm not well prepared to kinda present any ideas or thoughts on it.
[00:18:22] Jonathan Rosenberg: So we're having a meeting to discuss open issues. Yeah. So my thought is that, you know, we have this use case, like the SIPREC use case, right, where we're just capturing a tap of one of the participants. And in that model, you know, we can imagine we have a lengthy dialogue, you know, that ran for a while, the comics and raging was talking. And we're using the session to just inject... To basically create pointers so that, you know, we might have three sessions that point to the same dialogue, but just with different time markers for the start and end of the session. Yeah. That indicate, you know, this is the part that was, you know, that was relevant here contributing from this user. You didn't do that. You know, it's the work that the the server would have to slice up the recording into dialogues that match the the boundaries. And so I think we we gain a gain a lot by having a nice clean separation from the the the metadata, which is really Yeah. Yeah. Session. Compassing and the raw data, which is a dialogue. So I I would I would relax this constraint. Okay.
[00:19:36] Daniel Petrie: Yeah. My gut feeling is that we wanted to allow it, but I didn't have, you know, you know, concrete. That's a great example of where that's useful and and meaningful. So so, yeah, I I I... My gut feeling was that we should allow it. You know? It's not how I wrote it, you know, originally, but I didn't... You hadn't thought through the examples when I when I wrote the text for that part. And so
[00:20:03] Jonathan Rosenberg: Yeah. And that's... So... That's great. No problem. That's why we're having a discussion today.
[00:20:07] Daniel Petrie: Yeah.
[00:20:08] Jonathan Rosenberg: The other thing that means is, you know, there's now a possible corner case where the dialogue times have no overlap with the session lines. I think that's okay. You know, it's not useful, but it does, like, have nothing to do with them. I would not put them... I would, like, avoid a a must strength statement that says you must
[00:20:27] Thomas McCarthy-Howe: Yeah. Yeah.
[00:20:27] Henk Birkholz: Add some more up
[00:20:28] Jonathan Rosenberg: and and say, yeah. You should... But this, is still compliant. It's not true. And then there's basically a couple of things that are
[00:20:35] Daniel Petrie: Okay. Yeah. I... I'm... I I... I'm... I think I agree with all that. If people think that's that's cool, I will try... Attempt to try... Make a pass at that edit before I submit. Anybody have any contrary thoughts or complementary thoughts? Okay. So I will I will... Hey,
[00:21:03] Thomas McCarthy-Howe: Henk. Yeah.
[00:21:04] Henk Birkholz: Yes. I'm here. I have complementary thoughts, but this is a little bit dependent on the timeline. Of course, we wanna ship something for SF, but that doesn't have to be the final version. Right? Because we could use the headphone time to walk through the whole CDDL. I'm just making now a comment that the quantity of an extension point is not actually really necessary, but that is just it. So my my complementary thought is can we do some of that strict checking that is mentioned here in... At the hackathon and and and check out if... Because I only can test against the corpus, really. And and maybe something that is not allowed is also allowed, but I will never know this without the, test case where it is, instantiated something not intended and the CDL passes it through. And so so this would be back... Like like, a extreme programming side by side review for, I wanna say, two or three hours. And then we do this online or maybe at the at the IETf meeting if you're there. I don't know.
[00:22:14] Daniel Petrie: Yeah. I'll I'll be there. I'll be at the... You... You're talking about the the the CDDL, being strict, point. Is that that... That's where you're referring?
[00:22:24] Henk Birkholz: Okay. Exactly. It is strict, but only to the respect of the corpus, I know. And and so, there might be things allowed that are not intended to be allowed, but I have no test case that would show that. And thereby, yeah, I really have to, we have to walk this through at least once. And and then we're very good.
[00:22:46] Daniel Petrie: No. I mean, that's... I I think that's where we are for a large portion of the draft now, right, with all these changes. You know, I... I'm not comfortable making this, you know, any... Anything. I mean, obviously, there's still text that needs to be edited. But even, you know, if there wasn't, I I wouldn't be comfortable at this point saying, you know, work work of last call because I haven't implemented this myself. I haven't... You know, I don't know when else I... Has. I don't I don't believe. And, you know, until we, you know, kinda try this, implement this, I... This is, you know, this is still a work in progress. Got it. Much so more though than... So than it was, you know, six months ago. So I I totally agree. We need we need to bang on it. Okay. Thanks. Yeah. Thank you. Thanks for coming, Henk.
[00:23:39] Henk Birkholz: Sorry. Maybe a meta question. Do you need me for any of the upcoming items really? Because I have a severe conflict. I am in a full year of a of a conference center, and and and I'm I'm bailing. So is there anything that you need feedback on directly, or is this all list in online IT... In online IT effects?
[00:24:02] Daniel Petrie: Yeah. I mean, offline, it would be great if you could go through the current c... CDDL. I I made a bunch of changes. I had discussed this two weeks ago, and you weren't here. But I made a bunch of changes. I took... I I merged your PR in, and then I made some additional changes kinda, you know, extrapolating from what you did. You've added these sockets, and I went through, and I I did a a more thorough pass of adding sockets where I thought extension points ought to be. And that also, you know, kind of, you know, got me into the thought process of, right, I need to mention where those extension points are in in the draft text as well. And so I I did an iteration of that as well. And and then, you know, that that also stimulated, you know, my my thoughts on, alright. I need to write this up in how to write how to write a a VCON extension in in the, you know, in in the draft as well. So, I mean, I I I think, you know, review if you... When you have some time, you know, re... Review of of of all that, it would be it would be extremely helpful. We don't need to do do that here.
[00:25:15] Jonathan Rosenberg: Okay.
[00:25:15] Daniel Petrie: There was... They said there was one other point that I shoot. I dropped it. I had another point I was gonna ask you about. I forgot what it was. But No worries. Connect with me on signal so that I can I can I can talk to you?
[00:25:35] Henk Birkholz: Yes. We can do this via email. We can connect via email to signal and
[00:25:39] Jonathan Rosenberg: walk off
[00:25:40] Daniel Petrie: I I... Well, I I I I never get a response from you on email.
[00:25:44] Henk Birkholz: Oh, yeah. That is that is my email. That's not me. It's 1,000,100 emails in my email box. So... Sorry.
[00:25:54] Daniel Petrie: Okay. Let's go from here. Thank you. Yeah. Thanks. Oh, damn. I know what it was. I was gonna ask him about the
[00:26:06] Henk Birkholz: I have a look. Yeah.
[00:26:07] Thomas McCarthy-Howe: So Henk is fine. Okay.
[00:26:08] Daniel Petrie: Good. Good. Yeah. So so I was gonna ask you about the the charter edits, and I I just... And then I had just a general comment about the the the charter. I would like to push that the charter mentions not a, you know, AI agent... The AI agent extension, but a set of them. I would much rather see a bunch of smaller extensions than one huge one, because I think the pieces of that are useful. Right? And then... And it's... We can cut that into pieces.
[00:26:43] Henk Birkholz: We can cut the the the existing thing into pieces and add items to the pieces where necessary because I think that we go into different layers of the of the whole structure Yeah. Then that wouldn't make it feasible to make it one realistic in any in any case. So we... I can only... But my assumption was always we can only really cut them into pieces when we have the extension points. We will wanna put those in and then... And go from there.
[00:27:07] Daniel Petrie: Exactly.
[00:27:08] Thomas McCarthy-Howe: Yeah. Would support that. Yeah. I'd be able with that. Yep.
[00:27:10] Daniel Petrie: Yep. Okay.
[00:27:13] Jonathan Rosenberg: So for the the wording, would say one or more extensions as needed. And I I hate coding exact document structures into Charter. Give yourself flexibility so you can see other docs. Yeah. Yeah.
[00:27:22] Daniel Petrie: Yeah. Exactly. Yeah.
[00:27:24] Thomas McCarthy-Howe: Actually, the chain of thought was one of the... That really jumped out at me as a thing we might wanna have or not. So
[00:27:31] Chris Wendt: Great. Create a PR going into SFO. K.
[00:27:36] Henk Birkholz: On the charter, you mean? Yes. Okay.
[00:27:38] Daniel Petrie: Yes. Yeah. Yep.
[00:27:40] Chris Wendt: I assume Daniel do that.
[00:27:43] Daniel Petrie: Yeah. I can I can I can I can do do that?
[00:27:47] Jonathan Rosenberg: Listen, while we have while we have Henk and before he drops, one one other thing, little sidebar, the scheduling. I guess, I don't know if you guys noticed. This thing is stuck on Friday, this Friday afternoon, which is, the loser spot. Cool kids, man. Yeah.
[00:28:06] Daniel Petrie: We're the dedicated.
[00:28:08] Jonathan Rosenberg: We put... What did we put in a request to move it? Chris, did we hear back on whether it's getting moved? We have much drama on moving. There's so much drama
[00:28:19] Henk Birkholz: on moving things. I would be very careful with that. If it's just... Yeah. The loser spot, I don't think that will move a lot.
[00:28:31] Thomas McCarthy-Howe: We worded it.
[00:28:31] Daniel Petrie: Well, Henk, we
[00:28:32] Jonathan Rosenberg: we put in a request to move already, so I'm just asking Chris if there was an update.
[00:28:36] Henk Birkholz: Okay. No. There there No up... No update yet.... The map. We
[00:28:41] Thomas McCarthy-Howe: we have had our... A couple of couple times we've had that spot.
[00:28:45] Daniel Petrie: Yeah.
[00:28:46] Chris Wendt: Yeah.
[00:28:48] Henk Birkholz: I don't mind unless someone wants to leave early, which then is a real problem. I don't know.
[00:28:54] Thomas McCarthy-Howe: It's because they're it's because they are so badly behaved. That's that's the that's what I'm making.
[00:29:01] Jonathan Rosenberg: Actually, Henk, you'll be there on Friday?
[00:29:04] Henk Birkholz: I will be there for a long, long time. I think I have almost three weeks at the West Coast in SF... Ietf full coverage. Yes.
[00:29:13] Daniel Petrie: Yeah. Awesome.
[00:29:15] Jonathan Rosenberg: Cool. K. Thank you.
[00:29:19] Chris Wendt: I I... I'll try to follow-up, but I I I know there's... I've seen a a bunch of other requests happen too.
[00:29:28] Daniel Petrie: Yeah. I mean, I think, you know, if we can get more time too, think would be... I think just get a lot to talk about here. You know, lot lot to socialize in terms of what what has changed and a lot to discuss in terms of, you know, a bunch of new issues.
[00:29:45] Chris Wendt: Yes. And I included that in the request as well.
[00:29:48] Daniel Petrie: Awesome. Thank you.
[00:29:52] Thomas McCarthy-Howe: Always nice to have a little bit of hang.
[00:29:56] Daniel Petrie: Shall I keep going then? Are we... Any other issues?
[00:30:00] Henk Birkholz: Okay. And that is... I have to do that for now. Sorry. And, yeah, I'm I'm I'm over overbooked, but this was important because last time I should have been here, I regret it, and this is me trying to
[00:30:11] Daniel Petrie: make up for it. So Thank you, Henk. Okay. It's good to have
[00:30:15] Jonathan Rosenberg: you. Coming.
[00:30:16] Thomas McCarthy-Howe: Accomplished. Yeah. See you.
[00:30:20] Daniel Petrie: So as I I added these two two additional issues that we... I had discussed last week, weren't on on this list, so they they they are, you know, kinda not really new. And then the... So the the... We talked about the one about the the time overlap between sessions and dialect. The other issue was the... You know, which session and event references, you know, when the party or dialects spend more than one child's session. You know, basically, you know, is there an inheritance of of events, and do they apply to the children or, you know, whatever that relationship is? So that's the... You know, that's something we need to explore and think about more, or at least I do. Maybe other people have thoughts on that. And then let's see. I think the rest of this is not new, so I'm not gonna take time here to to go about that because I just... I I have no new new thoughts on those. I don't have anything to interject unless somebody else does.
[00:31:29] Jonathan Rosenberg: Just, Dan, as a general rule, it's it's hard to make decisions on data modeling questions like like this one without, like, concrete use cases that we can Yeah.
[00:31:37] Daniel Petrie: Yeah.
[00:31:38] Jonathan Rosenberg: Throw the proposal against. I'm I'm sitting here trying to dream up what's the use case where
[00:31:43] Daniel Petrie: Yeah.
[00:31:43] Jonathan Rosenberg: Event is ambiguous in terms of which one it might work, and so we can think about it. Yeah.
[00:31:49] Daniel Petrie: Well, I... Actually, the one... This... The the the one example that I had, you know, like, was the... Actually, I think I have it here. Right? And in in this case, you know, this was the supervisor whisper sidebar scenario. The join... Session zero is the, you know, the the parent session, And then the the join events for party zero and one are on session zero. Right? And they apply to the other... To the to the child sessions. That was that was where that came up in in my mind, where that issue came from.
[00:32:46] Jonathan Rosenberg: Oh, so this is like this. Do we have the agents, the coach, or the agent join each of the two sessions separately, or is it inherited? Is it separate?
[00:32:56] Daniel Petrie: Yeah. Yeah. Exactly. So the... I mean, that... That's the issue. I don't have I don't... I I haven't thought about that enough to to have a a strong opinion. I I I kind of, you know, faintly feel like we ought to be able to have the inheritance of the of the of the events just because I hate duplicating stuff. But I haven't thought through through the consequences or, you know, the issues with it.
[00:33:27] Jonathan Rosenberg: I think the problem is if you... What... If this data model says inheritance means the events always apply to children, then you can never have a case where a sub session has a different participant set. So is there a case where we want, you know, a trial session having different participant set? I think the answer is definitely yes.
[00:33:49] Daniel Petrie: Well, no. You can you can still have a different participant set because each session has its own set of participants. Right? It's got... So you can have a joint event on on a... For a party on the on the on the parent. You know, the event is on on the parent, but you could have a child session that doesn't even include that recipient. You know? And so you
[00:34:12] Jonathan Rosenberg: So then the parent inherited No worries. That's what we're talking about with a certain party. It's not the question.
[00:34:19] Daniel Petrie: Yeah. I mean, so it can inherit the event, but it's kind of nonapplicable because that's... That party's not not not a participant. It's not much different. I think conceptually then, you know, we've got a session that refers to dialogues that aren't part of its its time period. Right? I mean, it's it's kinda... You ignore it because it's
[00:34:39] Jonathan Rosenberg: Yeah. Up with the...
[00:34:41] Daniel Petrie: It's participants.
[00:34:41] Jonathan Rosenberg: Then how do you know... That makes the rule about ignoring events event specific. So we have to sit here and we're at least in systems that might not have no participant. So... Yeah. But if you can train up an event that has a participant ID that doesn't make sense, like, enter DTMF. I'm sure that's a few sample. You know, I'm a participant in the in the main conference, and then there's a breakout, and I press DTMF. DTMF is audibly injected into one of them. But I'm a disparate in both of them. So how do you disambiguate where was the DTMF at first? So in that model, I don't think you could have the... You don't want the awareness.
[00:35:25] Daniel Petrie: Well, I mean, I think in that model then, you don't put it on the parent. Right? You put that event on the on the specific session.
[00:35:31] Jonathan Rosenberg: Okay. So what if I press the VPN that's in the parent?
[00:35:38] Daniel Petrie: How do you do that? I mean, it it... It's it's an event, and you you you get... You choose what session it applies to. Right?
[00:35:51] Jonathan Rosenberg: I thought... Yeah. Maybe I'm confused. I thought the question is, if an event happens in a parent, so is it an event that inherited by the children and vice... And or vice versa? That's the only question. Right?
[00:36:05] Daniel Petrie: Okay. Yeah. No. I I guess I can envision a, you know, conferencing situation where you have breakout session or something like that where it it goes into the... Well, I don't know. If it if it goes into the parent, it's part
[00:36:18] Jonathan Rosenberg: of the...
[00:36:19] Daniel Petrie: Oh, but it wouldn't echo into the... Yeah. No. I I I get it. Okay. Yeah. My gut is we
[00:36:26] Jonathan Rosenberg: should not we should not... Inheritance, if you put it in, you can never get it out.
[00:36:30] Daniel Petrie: Yeah. Yeah.
[00:36:31] Jonathan Rosenberg: And since you have one count, for example, where you don't want inheritance, the whole thing breaks. Whereas if you leave it flexible, you get the drawback of duplicating events. But so what? You got plenty of bytes on the Internet.
[00:36:45] Daniel Petrie: And and another way... Yeah.
[00:36:47] Jonathan Rosenberg: It's more flexible.
[00:36:48] Daniel Petrie: One way we could solve that instead of duplicating is... Right now, I think an event reference is one session. It could have, you know, an array of sessions. Right? And
[00:37:00] Jonathan Rosenberg: Oh, well, that's an interesting idea. I really like that. I think that handily sells a problem. That's good.
[00:37:08] Daniel Petrie: Yeah. Okay. Well, thank you. I I hadn't had time to explore that, but this is this is good good input. And
[00:37:16] Jonathan Rosenberg: Well, that's what these are for. Right?
[00:37:17] Daniel Petrie: Yeah. Yeah. Yeah. Yeah. Cool. Alright. I'll go back to... Where were we? So I think, like you said, nothing... I I didn't I I didn't have time to think about any of these things. You know, I can walk through them if you want, and people can, you know, shout out opinions if they have it, or we can just move on. Alright. I'll move on. New issues. So almost everything in the draft that has, you know, kinda enumerated values or object... I mean, our parameter names or object names and and things like that are are all... There's... I am a registry for everything. Don't have them for disposition and encoding values. I, you know, I guess the the kind of question is should I should I add them for that... Them as well? They seem like kind of things that we that we might not be... Well, at least disposition seems like it's kind of an open thing, and I I I don't know that we necessarily wanna mandate, you know, registering all all those values. But in... Maybe including... I don't know.
[00:38:40] Jonathan Rosenberg: So just to be just to be clear, the the question is, should there be an for this? Because we have for each other parameters. Right?
[00:38:51] Daniel Petrie: Yeah.
[00:38:54] Jonathan Rosenberg: I see. So maybe a little bit of a tangent, but, you know, one of the things that struck me is, like, a little weird with the draft is there's a fuck ton of registries, man. And I wonder, do we need so many registries? Can you just have a single registry that instead has, like, parameter, you know, field. What you call it? Like, you know, one one string, which is which thing it extends and then another string, which is the new thing. So you would just register... If you wanted a new disposition, you'd add an entry to, you know, you... I wanna register, extends disposition, new value, you know, customer didn't purchase. If you wanted to extend tools, you'd... You know, new... You know, extends tool name, tool actions, and the new tool type is extended reasoning or whatever. Like, this way, it's one registry from the whole doc rather than tons.
[00:39:52] Daniel Petrie: Mean, a bunch of little ones....
[00:39:53] Jonathan Rosenberg: Doesn't work. Yeah. Instead of a bunch of little ones. Yeah. There's, like... It's a lot.
[00:39:58] Daniel Petrie: I mean, the the... I I I had to kinda... Rowan brought up that comment, and I I I brought that up, I think, at the last work group meeting. And I didn't I didn't get any feedback. So I I... Or I didn't get any anyone who objected to the large number other than Rowan who wasn't there to discuss it. So
[00:40:21] Jonathan Rosenberg: so... I think it's gonna bite us it's gonna bite us hard now with this new... Once we stick to AI stuff and we're gonna be in extension land.
[00:40:28] Daniel Petrie: Yeah.
[00:40:29] Jonathan Rosenberg: And this is stuff, like, every week, there's some pioneering innovation in AI, you know. So...
[00:40:34] Daniel Petrie: Yeah. Yeah.
[00:40:34] Jonathan Rosenberg: I I think having it easier to just have some, like, a two or three level naming structure for these fields, then we're probably fine. Again, I haven't done an analysis, but my gut says that's way better than tons of tiny registries.
[00:40:49] Daniel Petrie: Okay. Alright. I'll I'll put some thought into that. Thanks. The other thing, another issue is do we... I... Oh, yeah. Sorry. You already mentioned the redaction tagging. So I'm gonna have to I don't have to go through that. I had a couple discussions with folks about... You know, I went... Walked through all the kinda topologies for recording that we now support and, you know, the new ones that are enabled with session. And I had someone point out a new scenario that we don't cover. And so I kinda... I I... I've got a slide for that. Sorry. Rather than talking to that, let me bring that up. So, you know, we we talked about this this scenario. You've... If... You know, the recorder makes a bunch of recordings based upon talks parts. Right? So you got... You get a... End up with a bunch of these little recordings or whatever, and we support that. Right? We... Yeah. You know, you can... You have a dialogue for each, and then you have a session that groups them into into, you know, the the call concept. But there are... I I talked to to some folks, and they didn't like this because you end up with, you know, lots of little recordings. And and they didn't... They they wanted... They... Their their conference recorder worked differently or a little differently. And so...
[00:42:33] Jonathan Rosenberg: You don't have to do that. You don't have to do that at all. No.
[00:42:36] Daniel Petrie: I know you don't have to, and I... I'm worried that. But they pointed out, I thought, a a a pretty interesting, additional approach that I think we don't cover today with with with what we have, and I and I think it's worth worth covering. So, you know, some recorders do that same kind of thing as they only record Toxperts, but they they they do this in a single recording file with, you know, some number of channels. And so they record, you know, three channels and, you know, the three loudest speakers or whatever, get recorded and everybody else gets dropped. Right? Or probably the same thing in terms of mixing in in the conversation in terms of who gets heard. So the the the idea is that, you know, we've got a three... You know, in in this case, an example where we're recording three channels and, you the mixer and recorder are deciding who the parties are that get to get to be heard and recorded. And what we have right now is we have the ability to say, you know, what party or parties or none parties are being recorded in a specific channel in a multichannel recording. What we don't have is the ability to change that at some point in time in in the recording. And so my proposal here is to solve that, we invent... We create a new event type, which is, you know, dialogue recording channel party change. Right? So that you can specify that at at some point in time in this channel, in this recording, the the, you know, the party or set of parties changes to a new set of parties. Right? One or more or none. You can specify. And right? And so in this example, know, the white white spots are no... You know, there's no recording or no... Nothing recorded and no parties. And so these these half moon dots here are representing party change events within the recording. And so that's that's really the change here is to create this new event. I have text for this. I didn't... Sorry. Didn't bring it here, but I wanted to kind of present the concept. And if people are okay with that, I will add this to to to draft.
[00:45:02] Jonathan Rosenberg: I think this is a... That's a really great use case. I I had not thought of that one either. Just a question. I like the general idea of the events for party change. Are the events events on the dialogue or on the session? I think the dialogue...
[00:45:19] Daniel Petrie: The dialogue. And this... And, like, this is the first case where we have an event that is dialogue specific as opposed to session specific.
[00:45:28] Jonathan Rosenberg: Yeah. Yeah. Fascinating.
[00:45:29] Daniel Petrie: So... Yeah. So I... This is what I'm proposing in terms of the event parameters. You know, we have a type, right, which says that this is a... You know, I I don't, you know, here's a name we can talk, you know, discuss name. I don't I don't have big buy on it, but we have a type. It's it's a channel contributors event. We have a time stamp. We have a session, maybe. I don't know. Maybe that's optional. But we absolutely need a dialogue because we're talking about a specific recording, and and and and we have a channel index as to which channel it is and a new set of contributors for that channel in that recording at some point in time.
[00:46:13] Jonathan Rosenberg: I think this looks good, but I would avoid this. I'd skip the session. It's redundant. Right? I mean, the session will point to the dialer. Okay.
[00:46:20] Henk Birkholz: Oh, I will
[00:46:21] Daniel Petrie: drop drop the session.
[00:46:23] Jonathan Rosenberg: Yeah. I agree.
[00:46:23] Thomas McCarthy-Howe: Yeah. Otherwise, I like it too.
[00:46:26] Daniel Petrie: Cool. Alright. I will add text for this in... Yeah. Yeah. Before before I submit.
[00:46:32] Thomas McCarthy-Howe: Yeah. That's that's good use case.
[00:46:34] Daniel Petrie: Any any more discussion on this before I move on?
[00:46:38] Jonathan Rosenberg: Mm-mm.
[00:46:39] Daniel Petrie: Actually, I'm not sure if I have much more in here. Where am I? No. I don't. This is... I just threw this example in here in case I wanted to pull up an an example. So that's that's basically all I had to talk about today. And let people wanna go back to any of the other points or discuss some of the the the issues we didn't discuss in in detail. It's all I really wanted to cover. K.
[00:47:13] Chris Wendt: Thank you.
[00:47:14] Daniel Petrie: All I was prepared to cover. Maybe it's a better statement.
[00:47:20] Chris Wendt: So I don't know if anybody had any other detailed discussions we can sort of summarize where we are going into San Francisco. I guess, I assume you're gonna try to submit that before the deadline. Right, Dan?
[00:47:37] Jonathan Rosenberg: Yeah.
[00:47:38] Daniel Petrie: Yep. The... Oh, it's... So, I mean, this is, like, topic change, but let me just put a a a tag in here when we wanna discuss it. There was... I I... We... There's a there's a PR on the on the charter that I I don't know if we wanna discuss about adding that here or we just wanna wait till San Francisco to to talk about that PR. But I I didn't have a chance to really digest it. I don't know if anybody had a chance to look at it or or not, but I I Could you... You know
[00:48:14] Thomas McCarthy-Howe: what I got? Which PR is this? Just
[00:48:17] Daniel Petrie: There's... There is... The only PR on... In the in the repo right now is the... Is on on the charter change that... I don't even remember. It's some... I I I I don't even know. Let me pull it up. A farmer Moretto.
[00:48:51] Thomas McCarthy-Howe: I think he might be here.
[00:48:54] Daniel Petrie: I I don't I don't remember. He said something about talking about my invitation. I I don't remember talking to who this is, but but it it was at the...
[00:49:06] Chris Wendt: Is it Francesco? I think he is online.
[00:49:09] Thomas McCarthy-Howe: I think he might be here. Francesco.
[00:49:12] Daniel Petrie: Hi. Hi there. There. Hi.
[00:49:15] Francesco Moretto: How are doing? I I was I was muted. Sorry. I was just listening for for for the moment. So it's Yeah.
[00:49:25] Chris Wendt: Yeah. No worries. Yeah. Do you wanna talk about
[00:49:29] Francesco Moretto: the Yeah. Maybe it's the the idea is to have something that glue together. You know, my... I've just published it last Wednesday, not this one, the the the Wednesday before, the map about what what is going on about the media centric commerce, all all the tables, all the all the standards, you know, venues, and
[00:50:03] Jonathan Rosenberg: a
[00:50:04] Francesco Moretto: lot of people are talking about how to put the the the... To to to make the standards about the the agent e commerce, UCP, and their IAAF, w three c, IATF, and and and so on and so forth. And so I've I've I've been very, very interested by the... This new, you know, momentum of of of Veecon in in in the way to to use Veecon for for basically one one of the means. Okay? It's it's it's... So so one one of the of the points of my my of my paper that I've presented just last week. Okay? Busy busy weeks about how to have a minimal record and and and a minimum viable record in order to... To have a a a proof and have something that it's usable, for example, in in Europe as well with the regulations that we have in in in in Europe. I mean, basically, Milan, Italy. And it said this PR is is on this on this main idea, and it's and it's something that, okay, I'm I'm re... Really open to to to discuss about and whatever. But please please make any any, you know, considerations or even close it. It's a... It's... If if you think it's it's not it's not in the in the in in in in the good for the... In in in the the the main main... It's it's something that makes some kind of noise or you if you prefer. No problem.
[00:52:21] Thomas McCarthy-Howe: Yeah. Well, you know, so so, Francesco, from my perspective, and thank you for this, this is... Think it's absolutely part of our charter, but but hits against what we expect audit... The audit team to be able to use. Right? So I think I think from a charter
[00:52:44] Jonathan Rosenberg: Can can we look at the PR on the screen maybe? Someone Yeah.
[00:52:48] Daniel Petrie: I was just going to send a link, but but
[00:52:52] Chris Wendt: Yeah. Let Let me see if I can do it.
[00:52:55] Daniel Petrie: You you have the link?
[00:52:57] Thomas McCarthy-Howe: Yeah. Bring it up.
[00:52:58] Chris Wendt: Yeah. I I just
[00:52:59] Jonathan Rosenberg: found it.
[00:52:59] Francesco Moretto: Only one.
[00:53:04] Thomas McCarthy-Howe: This is this is exactly, you know, the the boundary of what we have to represent, I believe, and our requirements into the upcoming audit working group if it ever gets chartered.
[00:53:18] Chris Wendt: Yeah. You can see that okay. Right?
[00:53:22] Daniel Petrie: Yeah. Yeah.
[00:53:25] Jonathan Rosenberg: I Yes. So this was the Yeah.
[00:53:28] Thomas McCarthy-Howe: So it adds the sentence instructions and approvals given by the person or organization on whose behalf of agent acts in references to the records and other systems that resulted from the conversation such an order or a payment. That... I'm not sure if that if that beats up against audit or agent proto, but it definitely... You know, it beats up against one of them, and I think it's in our charter to represent that.
[00:53:53] Francesco Moretto: Well, it's... I'm... Let let me check because it's... I can put in in the chart here. Or... But, well, I'm I'm really open. I mean, it's... If if you think this is something it... That it's going to hurt, but you can also see on my... Let me see on the map. In in reality, it's something that it's it's it's one of the scenes potentially to be come back to... Let me check. In in in reality, it's it's something very, I mean, small for for talking about what can be also done with Vicon as a as a mean in in order to have... To be extended and to be adopted more easily into the other... Into into the other tables, into the other working groups, and and and and so on. So... But, yeah, I repeat. I'm I'm not really... I I I... My main point is to have a discussion about that and to understand whether this this group is... Has has a willing... Has a has a will to to to to see also what is... You know, it's it's really close and really useful as as well, my my humble opinion.
[00:55:53] Thomas McCarthy-Howe: Oh, no. I I think you're absolutely right. And maybe Chris or Jonathan could step in and maybe give a little more fulsome, you know, information if they have it. But there's two other working groups that we coordinate with. One one is a working group. One, we expect to become a working group. The work work group that we we are coordinated with is agent proto. The other is audit. And in both cases, our our role there is to be the observer, to direct this... The system of record. This is actually what happened. And this is... I think this is clearly within that that that that, you know, that that place. And because this is just the charter, not the implementation, I think it's appropriate to be in the charter. I think the implementation of that has yet to be understood based upon, you know, what what the other two groups are doing in... As an implementation. But I definitely think this is part of our our scope. Chris or Jonathan, is that how you understand it or something different?
[00:56:57] Jonathan Rosenberg: Go ahead, Chris. You go first.
[00:56:59] Chris Wendt: Well, I I was just gonna say I think we wanna be careful not... Like, we wanna record these types of events, but we don't wanna necessarily define the mechanisms for them.
[00:57:10] Thomas McCarthy-Howe: Oh, no. That's correct. Yes. Agree with that. Yeah.
[00:57:13] Francesco Moretto: We we just we just I would
[00:57:15] Daniel Petrie: Scrivener.
[00:57:15] Jonathan Rosenberg: I would like Yeah.
[00:57:16] Chris Wendt: Yeah. I I think it's good to make sure we're as general language as possible to make sure that, you know, we can give the flavor of things we wanna do, but at least in the charter. I don't know. Jonathan, you have any...
[00:57:30] Jonathan Rosenberg: Yeah. Yeah. So a couple thoughts. First, there's there's two things and we should separate them. One is, you know, can we... Can the Vcon reference complicated objects that someone else defines? Do we want that? And a comp... Here are complicated objects, say, is the thing that captures the approval. And I think the answer is unambiguously yes. Like, no brainer. That has to be possible. The second question is is does the Vicon group itself undertake the work of defining the object format for how one captures human approval of something like like an ecommerce purchase. Mhmm. That one, I I get squishy, and I think it's a little bit too far afield from the current scope of the group. So I think that, you know, part a is definitely as part b, probably not. Where might that live? Well, I I don't... It's not agent proto, by the way. Agent proto is more about session protocols, not object formats. It could very conceivable be the audit. And in fact, this came up in an audit discussion as there's been interest in capturing human approvals of of AI agent actions. I have a draft that does something like this also called check, c h e q, and I didn't think it's in scope for Vcon either. Maybe it goes to audit. So I think that's all sort of still up in the air, but I definitely don't think it's an econ. So that's that's...
[00:58:59] Daniel Petrie: Right. Yep.
[00:59:01] Chris Wendt: And I assume that... I mean...
[00:59:03] Jonathan Rosenberg: But I but I am... But I do think there's super interesting... Like, I think this is a very important piece of standards work. That's all, Francesco. But thank you for coming forward, and I think I think we're gonna need a lot of this. The the world's gonna need a way for the... For us to capture and memorialize and secure what... Whether a human approved or disapproved or what exactly they said or didn't say. And so to the extent that it's more than just the conversation, which Econ captures, then we're gonna need to standardize some stuff for it.
[00:59:31] Chris Wendt: And I think this is perfect thing for extensions where we can either be more prescriptive or less prescriptive prescriptive depending on what the the thing is. But I guess in terms of the charter, I don't know how... Do we want to capture this maybe in a better way that... Or or or or or do we already have that as part of our charter, I guess, more generally?
[01:00:00] Daniel Petrie: Yeah. That that that would be my... I think my opinion would be that let's keep the charter more general and so that it's open to this, but not maybe specifically call it out.
[01:00:16] Thomas McCarthy-Howe: Mhmm.
[01:00:18] Daniel Petrie: Because, you know, that gets into the... You know, how do we package this? How... You know, what are the extensions and things like that? And maybe keeping more general to AI and and and and, you know, AI agent AI agent and AI agent tool, you know, conversational capture or observation, and maybe not specific features or functions within that scope.
[01:00:50] Chris Wendt: I mean, maybe the best... I was just gonna say maybe the best way to think about this is does the current charter prevent us from having these discussions about these topics? And and I Right. I sort of think now. But Fair.
[01:01:06] Thomas McCarthy-Howe: Yeah. You know? And also say, Jonathan...
[01:01:09] Chris Wendt: Oh, go ahead. Yeah. Thomas.
[01:01:10] Thomas McCarthy-Howe: Yeah. I was gonna say, no. I don't think it does. We should double check that. But, also, to Jonathan's point, you know, whether or not the work itself lands in our group is really up... Really, I think, coordination with the audit team. And even if it did, we have extension mechanisms that we could use if we had rules. Right? So it wouldn't hold back the core. We could keep keep going.
[01:01:34] Jonathan Rosenberg: So as it as it relates to this specific PR, I would say this... The second, looking at the stuff that's in, you know, highlighted green, which is the core change, I think. Right? It's the, you know, instructions and approvals given by a person or organization on whose behalf in h and x and references to records and other systems that resulted from the conversation such as an order of payment. I think I think the second part is really good. Right? Because that captures the idea that we need to be able to store references to stuff that someone else's problem. It's the first half of this that I I think is is more fuzzy, but I I think... So I would sort of concretely propose we delete that first part and keep the second part of this PR. That... That's that's... But that would be my concrete proposal. And then, you know, there's nothing that... Obviously, you can talk about anything, but, you know, I think that that... Yeah. Because I think the question is how does one codify an approval outside of the conversation itself is a profoundly complicated question. And that's the bit I don't want to have this group explicitly, you know, trying to do right now for this recharter.
[01:02:45] Thomas McCarthy-Howe: So Oh, I just can't clear it out clear it out of our scope.
[01:02:48] Jonathan Rosenberg: Yeah.
[01:02:49] Thomas McCarthy-Howe: Think to record it is not, but to represent it, which requires recording, I guess. Yeah.
[01:02:54] Jonathan Rosenberg: Yeah.
[01:02:55] Thomas McCarthy-Howe: Yeah.
[01:02:55] Jonathan Rosenberg: Yep. Does that make sense to you, Francesco?
[01:02:59] Francesco Moretto: Yeah. Does make sense, obviously.
[01:03:03] Thomas McCarthy-Howe: Thank you.
[01:03:03] Francesco Moretto: The the idea the idea is to have something that it's smooth enough to be also acceptable for every parts and every COGS that it's it's involved.
[01:03:20] Thomas McCarthy-Howe: Yeah.
[01:03:23] Jonathan Rosenberg: Yep.
[01:03:25] Thomas McCarthy-Howe: Very... By the way, I I think PAP and and and PACT are really interesting developments, personally.
[01:03:34] Jonathan Rosenberg: Sorry. Which ones?
[01:03:36] Thomas McCarthy-Howe: PAP and PACT.
[01:03:39] Francesco Moretto: PAP PAP and PACT. Yeah. Yeah. You can you can see
[01:03:43] Jonathan Rosenberg: I don't know what those are. Sorry.
[01:03:45] Francesco Moretto: You can you can see a lot of informations inside the the the map... Inside the link that I've posted into the into the chart. There's there's... You know, my my agents are particularly powerful.
[01:04:01] Thomas McCarthy-Howe: Yeah. It had... And has a has a significant amount of industry support.
[01:04:06] Francesco Moretto: Yeah. Absolutely.
[01:04:07] Daniel Petrie: Yeah. Okay.
[01:04:12] Chris Wendt: So what's the path forward on the PR?
[01:04:15] Henk Birkholz: Do you
[01:04:15] Chris Wendt: want Francesco to modify it? Or or, Dan, do you wanna
[01:04:21] Daniel Petrie: Yeah. No. I mean, I I... That's what I was gonna ask it. Francesco, would you mind, you know, retracting this PR and then, you know, proposing a new one with the... Based upon what we talked about today?
[01:04:31] Francesco Moretto: So, basically, leading only this second part. Is it correct? So age... Agent mediated transaction such are purchases. Is it correct?
[01:04:46] Jonathan Rosenberg: I I would I'm saying yes. Exact... Exactly right. Yes. References to the start... Yeah. That's the part. So just delete the stuff before the first semicolon.
[01:04:54] Francesco Moretto: So the the first... The the deleting the first part in... From instructions to payment and leaving only line 55. Is it correct?
[01:05:05] Jonathan Rosenberg: No. So I... I'm I'm... And, also, I'm sorry. I'm on a tiny screen, so it's a little harder for me to read. I'm saying the the part that you keep is... And references to records in other systems that resulted from the conversation such as order or payment. And you delete the part that says instructions and approvals given by the person or organization.
[01:05:26] Francesco Moretto: Okay.
[01:05:26] Jonathan Rosenberg: This way Okay. It was good. Understood. Alright. Thank you. And thanks for participating and Thank you. Contributing, Francesco.
[01:05:34] Daniel Petrie: Yeah. Got
[01:05:35] Chris Wendt: it. And and I think there's also at the bottom there, the line 55, you added a bullet...
[01:05:41] Jonathan Rosenberg: Oh, I missed it. Let's see.
[01:05:44] Chris Wendt: But I I I guess that's sort of...
[01:05:46] Jonathan Rosenberg: Oh, no. We should delete that. No. Because this is... Hold on. Wait. Hold on. This is So
[01:05:57] Francesco Moretto: what about line 55?
[01:06:02] Jonathan Rosenberg: Oh oh, I see. Okay. That's not... Sorry. Missed it. It's a little hard for me to read it here. Yeah. I think it's fine. Yeah. It's good because, I mean, I think this is an important use case. Yeah. I mean, I think we've already talked about humans talking to AI agents, and a lot of those are gonna be doing, you know, commerce. So this certainly fits the bill, and it would make sense given the reference. Yeah.
[01:06:22] Francesco Moretto: Going going going to be... Going to modify the the the PR.
[01:06:27] Daniel Petrie: Yep. Awesome. Got it.
[01:06:32] Chris Wendt: And then we can we can have a brief discussion about that and accept it in the meeting.
[01:06:39] Daniel Petrie: So you you don't want me to merge that that PR then?
[01:06:44] Chris Wendt: I don't know. I guess I guess we can as long as Chris Francesco. Alright. Is everyone okay with that?
[01:06:51] Jonathan Rosenberg: Let's not make this complicated. Merge the PR and then and then Yeah. Dan, like, right now, submit another PR just deleting that that extra bit we wanna get rid of and then merge it, and we're done. We don't need to talk about it again. Yeah. But that's... And and
[01:07:02] Daniel Petrie: if if everybody's okay, I'm gonna... I'm gonna change the wording so that it's, you know, one or more extensions for for AI to AI. Sounds sounds great. I'll merge I'll merge that in, and then we'll have just one document to discuss at at San Francisco. K.
[01:07:21] Jonathan Rosenberg: So I have a question for you, Chris. So does this mean we're not gonna... We're not planning to submit the the recharter request in advance of San Francisco? Is that correct?
[01:07:32] Chris Wendt: I I talked with Charles. I I think the plan was that we were gonna discuss it and then submit it after the San Francisco meeting.
[01:07:42] Jonathan Rosenberg: Okay.
[01:07:44] Chris Wendt: If that's alright with everyone.
[01:07:48] Daniel Petrie: Yeah. Sorry.
[01:07:48] Jonathan Rosenberg: That's what I explaining,
[01:07:49] Daniel Petrie: but I didn't I
[01:07:50] Jonathan Rosenberg: didn't Let let let me make one counter case. Alright. Let me... And, you know, this is a little bit different thinking, but, you know, there's a lot of people doing ITF because because of interest in AI. And I think, as general, they're probably... They look at the current chart, and they wouldn't even think about participating. One of the reasons I sort of like the idea of doing the recharter before was it puts the email on the ITF now in the same week, the econ group is gonna be working on this. Therefore, letting people know that they should probably attend the econ meeting. So it's like... It's a marketing comment, lack of a better definition of the term. Not quite marketing, but you know what I mean. Yeah. Advertising, if you have a better word, or getting the right people in the room who would be the right people to participate is the is the real way to think about it. So that's the only downside to to totally waiting. So just... That that was my thought.
[01:08:46] Daniel Petrie: There's there's something in between that, and so we could announce the, you know, basically, the proposed charter to, you know, all the AI emailing lists. Right? That's... We could propose it to try to get people's attention.
[01:09:03] Jonathan Rosenberg: Okay. That's a good idea, actually. Let's do that.
[01:09:06] Chris Wendt: Okay. Yeah. To to which list are we talking?
[01:09:10] Thomas McCarthy-Howe: By the way, I think... To say, I think it's been the lack of marketing that's allowed us to be safe in our little hole.
[01:09:16] Jonathan Rosenberg: It has. I mean, these meetings have been the most effective inter meetings because nobody shows up at the... Like, the the handful of us here. It's it's great. Great progress. Yeah. Normal meeting work.
[01:09:28] Daniel Petrie: Yeah.
[01:09:32] Jonathan Rosenberg: So the so the the AI the AI group now is... I think, Francesco, that that that doc you sent a link to is very handy. Right? It enumerates them cleanly. Right? It's agent proto, audits. I wouldn't bother with AI prep, but... And Vcon. But I think if you you sent a a note to agent proto and audit, you would cover it.
[01:09:55] Daniel Petrie: I I might as well hit Don as well.
[01:09:57] Jonathan Rosenberg: Don. Yeah. Sure.
[01:09:59] Francesco Moretto: Yeah. Yeah. Yeah.
[01:10:00] Jonathan Rosenberg: This is definitely not AI prep and not web bot off. So I think those are pretty
[01:10:08] Daniel Petrie: decent. Yeah. Yeah. Okay.
[01:10:15] Jonathan Rosenberg: So Chris, do you want do
[01:10:16] Thomas McCarthy-Howe: that or do you want
[01:10:17] Daniel Petrie: me to do that? Yeah.
[01:10:18] Jonathan Rosenberg: I was gonna say that's... No. Think Chris Chris Chris has a work group chat. I think it's Chris' job.
[01:10:22] Daniel Petrie: Okay.
[01:10:23] Chris Wendt: I can do that. I I
[01:10:25] Daniel Petrie: might... I'll let you know when when all that stuff's merged in.
[01:10:29] Chris Wendt: Sounds good. I'll I'll I'll I'll just... I'll validate that with Charles as well. Okay. Just to keep him in the loop. Alright. Any other topics for today?
[01:10:52] Daniel Petrie: So that's good. Prepared?
[01:10:55] Chris Wendt: Yeah. So I I I definitely agree that all these interims have been super helpful and and appreciate all the work that went into them and the the discussion. I think we've... We're coming out with a a lot better document and it... And it's been great. So appreciate all the contributions, Jonathan Yeah. Dan. It's been great. Thank you.
[01:11:19] Jonathan Rosenberg: Yeah. Awesome, guys. Awesome. So, Dan, I owe you... I'll have a PR early next week, Monday or Tuesday for the redaction stuff, and I'm gonna send you a separate note on the AI agent stuff too. So I'll probably do that tomorrow. Okay? Perfect. Alright. Awesome. Thanks, guys.
[01:11:36] Daniel Petrie: See you
[01:11:36] Jonathan Rosenberg: in San Francisco. Thanks. Alright.
[01:11:38] Chris Wendt: Yep. Thanks, Francisco. Take care. Bye. Bye.
[01:11:40] Daniel Petrie: Bye. Bye. You. Bye bye. Ciao.