 Hey, Baram. Baram, are you there? Yeah, I'm here. Okay, cool. Just want to get you on the attendance. Hey, Clemens. Clemens, are you there? What about David? I'm just trying to get my audio set up. No, no problem. Bobby, are you there? Yes, I'm here. Okay. And Clemens, are you there yet? I'm going to give him the benefit of the doubt. I can't imagine him not speaking up later. Sorry, though. Did you ask my question? No, no, I just want to make sure you're there, Fabio. Okay. Yeah, yeah. I'm here. Sorry. Okay. And Stanley, are you there? Yep. Okay. I don't hear Stanley yet. I hate the fact that on Chrome, on a Mac anyway, if you use the certain finger gestures to go to like move the window left and right, if you do it too, too many times, and then things, oh, you meant to go back a page. Hey, Doug, I made a mistake. I screwed up somebody's name because it wasn't updating correctly. Baram, I overwrote his first name with my name. Are you sure he doesn't want to go through a rename to do that? Maybe, but I need to adjust it. There you go. Thank you. Okay. I'll take care of that. Let's see. Here we go. And who's that? Oh, Clemens. Okay. Yeah. Thank you. Let's see. Who else? Steve, are you there? Yes, I'm here. Hello. Okay. Hello. Yes, I'm here. All right. Let's see. Thomas. You there. Thomas. You there. What about Jim Curtis? Yes, I'm here. Excellent. Thank you. Mark. Hey, Mark. Or it. Are you there? Yeah, I'm here. Excellent. Alex debris. Yep, I'm here. All right. Thank you. Who else? Thomas, are you there yet? What about Shawn fieldman? Hi, Feldman. Feldman. Sorry. I apologize. It's all good. Let's see who else can I pick it on? Daniel, you there. All right. Austin, you there. Good morning, Doug. I'm here. Excellent. Matt Rikowski. I'm here. All right. Joe Sherman. Yes, I'm here. Good morning. Good morning. What about Dan Rosanova? Yes, I'm here. All right. There we go. You're on either. Yes. All right. Dan Barker. I'm here. All right. That's not everybody. I'm sure I missed somebody. Yeah. Hi. Hello. All right. This is Rachel. I'm here. Oh, hey, Rachel. Excellent. Just getting to you. And I'm, this is Sarah. I'm joining by phone. Sarah zoom. In 15 minutes. What about the arm? Hey, this is the arm here from Oracle. Excellent. Thank you. Let's see. I'm pretty sure I missed somebody standing there yet from Oracle. Okay. Is there anybody on the call who's not. Been. Starred in the attendee list yet. Hi, this is William from Oracle. Right. I started. Who was that from redhead? I'm sorry. Who was that? Did we just spoke? William. Oh, William. Got it. Okay. Thank you. And there was some other person I noticed close. So you're there. Yes, I'm here. Excellent. And Thomas is here. And Thomas. Excellent. Where are you Thomas? There you are. All right. Let me start sharing my screen. So we're going to get started in a sec. All right. Let's see who kind of pick on. Ben, are you there? Good morning. Good morning. Chad. I'm here, Doug. Excellent. Thank you. I think that might be everybody. Hey, this is Stan. I'm here as well. Oh, excellent. Thank you. Good morning. I had some issues. All right, cool. All right, I'll give it to four past the hour on my clock and we'll get started. And is there anybody on the phone. Who did not announce themselves. All right. In that case, why don't we go ahead and get started for past the hour. Okay. We did roll call. Just a reminder. There was several. Yeah, as they're outstanding, just try to get those and get a chance. I'm going to keep going unless someone did like specifically say anything. Okay. I see Sarah's adding some comments in there. So you guys can read that later. So just a reminder. The coup con CNC F con and EU. It's coming up relatively soon. We deal with doodle poll as of right now, we have 14 people signed up for face to face meeting. I believe that's probably more than enough to, to officially make it a, a quote, official meeting. I will send out a, or create a doc or something for people to start jotting down official topics in advance of the meeting. I think it's a little too soon to start doing that quite now, or quite yet, but just to get your heads up there. Are there. And just so you know, I did request a BoF session for us. There's not been confirmed yet. But it is in the works. Hopefully we'll get that fairly soon. Are there any other questions or comments around the. All right. Not hearing any moving on. Just a little nagging reminder. We have 15 open PRs. You can see in the notes here. Three have Travis check errors. You have unanswered questions. One needs to be signed three need rebasing. Some of them meet multiple criteria in that list. So just a nagging reminder to, if you opened up a PR, please go back and look at it. So we could try to get some of these resolved granted. So we could try to get some of these resolved granted. Some of them may be blocked for other reasons other than. Procedural things like things like they're blocked by agreeing what his source or something like that. But in terms of being able to answer questions and stuff like that, try to get to those as soon as you can. Just so we can move things along. And of course things like rebasing and stuff you should be able to do now. And so with that, we can now jump into the meat of the agenda. So Sarah, you want to talk about your PR 102. Yeah. Am I, can you hear me? Yep. We can hear you. Go ahead. Okay. Super. So I'm going to talk about the design goals that are now part of the first. The, the key thing that I think we've all been aligned on is that where the goal is to allow services to be loosely coupled during development, deployed independently and later can be connected to create new applications where those could all be done by different people in different entities. And I think the key moving forward out of the discussions on Monday and Tuesday was the second paragraph where we figured out a way to talk about what we're trying to achieve without using some of these words that imply implementation to at least to some folks. So, so the first sentence, I think it's unchanged, which is the goal of the cloud. Define interoperability of event systems that allow services to produce or consume events where the producer and consumer can be developed and deployed independently. We chose to use the word producer and consumer because those are not in the set and they, we felt they were generic terms and they wouldn't imply any definition. And they seem to be had sort of generally accepted. So, so the first sentence, I think it's unchanged, which is the goal of the cloud specifications is to define generally accepted high-level definitions. And then the second sentence, a producer can generate events before a consumer is listening and a consumer can express an interest in an event or class of events that is not yet being produced. It was a key, I think, expression of what many people in the working group would like to call a topic, but then seems to imply a particular implementation of PubSub or Kafka or a set of things that use topics. And then a number of other folks in the working group are consider triggers and event types and those types of things, reclassifications and want to move towards those, those types of formalized definitions. And so, but we all agree we want, we're advocating for those different things in order to achieve these properties. So I think that that allows us to kind of frame what we're doing and distinguish it from asynchronous messaging. I think one of the things that I've learned in working with these technologies over the last year is that when you say event in the industry, it can mean one of three things usually. One is any kind of asynchronous messaging. It's often called an event. And this is a specific kind of messaging where we're decoupling the actual getting something from source destination from the event generation. The other type of event, which I think we all agree is a subset of what we're doing, is any time series data, particularly most often uses logging, is considered an event. You talk to a logging person and that is exactly what an event is, time series data. And I think that this working group is looking at a lot. It wants to be inclusive of systems where things might not be ordered, time may not be a thing, and a lot of people deal time should be optional. And so that's like a completely different lack of or different concerns. It's overly specific to what we're talking about. So we tried to capture what is the essence of the kind of eventing that we're talking about. And then lastly, the last paragraph in the design goal section is, I think, just a little modified from before, which is to defend the specification will include common metadata attributes of an event that facilitates interoperability, where the event does not contain any details about the consumer or transport that might be used to defend the event. So in some ways it's restating the prior paragraph, but it's clarifying some of the things that we generally agreed to at many of our meetings and the ideas that these three paragraphs help newcomers to the specification to understand what category of eventing we're talking about. So before we talk about non-goals, I'd love to get feedback about particularly from people who weren't involved in the Monday, Tuesday. I think what you just described is perhaps the most succinct description of what kind of events, cloud events, is trying to describe. And most of the detail you just said in describing this paragraph is not actually in the paragraph. It might be useful not to have it there, but to have a more verbose way of describing all of those things that you just said that is maybe linked to from here or some things like this is the succinct version. This is the concentrated description of what we're going for. And here are pages describing sort of all of the different ramifications of this and the way it sort of bubbles out into the rest of the world in order to get a more complete sense of what you're trying to get through with this paragraph. Yeah, I like that. And I think that that could be like an add-on. I think we tried to keep this concise one because we wanted it in the spec, but also fewer words are more likely to drive alignment. And what you're suggesting is really consistent with having another and adding something maybe to the reasoning or a companion document. Yes, precisely. So Ben, is that something that you'd like to see part of this PR or as a follow-on PR? I really appreciate that specificity and laying out here are three different examples of different types of events and how they fall into this spec. And I got great value from that. I just wanted to call it out. I don't think it needs to be part of this PR. I don't think it should be part of this paragraph, but I think it will be incredibly valuable to help orient people that are coming because the flip side of being concise, yes, it becomes easier to get everybody on board, but there's also so much more wiggle room and ambiguity because there isn't sufficient exploration of all the different options. So no, not part of this PR. That was really great and useful. I wanted to say so. Hi, Sarah. I think this looks very helpful. It's concise and simple, and I think it will mitigate a lot of confusion for newcomers. And I also appreciate what Ben said. I know that there's discussion and about section and in the future could link out to the about section where you could explain in more detail all the things that you just described. I think that would be a great solution. I'll also throw in, I've been working on something that I just made public yesterday to explain these concepts inside Google because we have lots of different messaging systems. I run into this a lot. So it might, if it's useful to folks, we can take some of the language from that. But I'll just put it in because it's other. Any other comments, questions for Sarah? Of course she rolls on to the next piece of text in PR. Okay, go ahead, Sarah. So non-goals. These have just been ones that have come up in conversation. There may be more non-goals, but I think that this has suggested to be in another sphere, but we can talk about that. So basically the function, a lot of the stakeholders in this group are specifically focused on function as a service. And however, some systems don't even involve functions. You could trigger other things or invoke or your consumer would not be functioning. So I wanted to particularly call out that the function-building invocation process in a fast system is outside of the scope of the specification. Also many folks in this group are working on language-specific runtime API. And we're just saying that those are outside of the scope of this specification. And may or may not be follow-on work, but certainly many of us believe could be enabled or accelerated by this specification. And then there's been some discussion about the importance of figuring out some kind of authorization thing. But the people who are advocating for that also agree that it is a... We do not want to, you know, select a single identity or access control system where aligned with the, you know, king-makers philosophy of DNCF. So we just wanted to call that out like that is not something we're doing here. And leaving the way open that we may need to deal with authorization. But that is... That is actually... That's like potential follow-on work. And sort of related to that, I'm going to zoom up to the working group process where we took out authorization because I think that generated a lot of discussion and instead put in this place like one of the things we will do is identify and resolve whatever else is needed for interoperability. So we're kind of leaving the way open that we need to have further discussions on how do we refine these goals in order to meet them and streamline our conversations. But that is as far as we got. Any questions on the changes to the README, meaning the working group process or to the non-goal section? Any comments? I put this in on the PR. I'm not sure if Sarah has seen it yet though. There's some legacy status section in there that we added in right at the beginning of this effort. And I believe the roadmap and some of the design goals kind of afflict with that section. And for clarity's sake, I would recommend probably removing that maybe within this PR or a near future one. Yeah, I had a comment back on that and I was going to tell you that as soon as this PR was merged I was going to submit a PR to pretty much remove that section. Great. Yeah, I saw that comment and didn't have a chance to comment on it. And I'm aligned with that and I think it could be a follow-on PR. Any other questions or comments on what Sarah presented? I believe the rest of the changes in the PR are just simply typographical, just extra spaces type stuff. Yeah. Okay. Any other questions, comments? Okay. I know the outstanding rule is things have to be changed within like two days. I believe there may have been some minor changes done yesterday. So if you have any concerns, don't feel like you've had a chance to fully digest this, please speak up. But I am going to ask if there's anybody who would object to adopting this, because I don't think it's gone through any substantial changes within like the last day, day and a half. So it's on the border for timeline. Or if you want to speak for someone else, you think may want to re-emerge. That's fine too. I don't want to lay unnecessarily, but I don't want to force it either. So I'm going to go ahead and ask, is there any objection then to adopting the PR? All right. Now hearing any, thank you guys very much. And thank you Sarah very much for taking the lead of offline calls on this one. I appreciate it. Thank you sir. Great job. All right. So the next topic, I think is what is source now. And don't, we don't have to do it this way. It's my initial idea when I was thinking about this on the agenda was I'm not sure how much we can get into on this particular call right now. And so I was kind of wondering what you guys thought about having additional meetings, the same way we did for the goals section, where a group of people go off and work on a proposal to bring back to the group. I was wondering what you guys thought about that process rather than trying to hash through it right here right now. Maybe we could bundle also the event type and source into like the same discussion. Because it's sort of correlated. That is a good point because I think there's something, there's overlapping concerns in all that. Yeah, we had already, I think we had already touched on some of those aspects in the, in the call, in the calls this week, especially when we were talking about the class, the class of events and basically the, the fact that a classification is required for you to be able to subscribe to it. And if we, if we take, if we do another round of discussions on that, I think we're going to get clarity. I think ultimately what this touches on is that we have, we have the source, we have the source ID, we have the source type, and then we have the event type, which means we have four qualifiers and, and the goal would be to rationalize all of those and see how many of these we really need. I know that there are some people who are, who believe that there should just be one. There's some people who believe there, there should be three. And we should try to figure out with all the people who are passionate about, you know, that classification about how many there really need to be right now. There are four, but my personal taste is. Yeah, and there's also the namespace one, which I'm still not clear about. The, the namespace one, I think we can go and add that as the fifth if you want. I think that's like, if the, if the source is namespaced and the event is namespaced or there's another third namespace. Yeah. So on that particular one, going too much into the details because that's a discussion I can probably also talk about for like half an hour. I think there's a, there's good reason to go and disambiguate the event all up when you get an event in your hand to have a clear notion that this comes from Azure or this comes from Google Cloud or this comes from AWS having a clear indicator of whether what, what, what system realm that event originates from that gives you good scoping and understanding how you should interpret all the other fields. So I still find this notion of namespace as the, the, the scoping for, you know, what kind of system realm does that come from? You still very useful. Yeah. We want to solve it right now, but you know, what you just said was source because you said, where does it come from? And if the source has a quality, you know, so, but let's just table it to the source discussion because. Yes. It's. I have an idea. This sort of came up in a few different. Maybe what we have to first align on is what is the criteria by which and how do you use inner out required. I think that we'd all like to align on the required things before we go through a backlog of optional things. And I think if there is some descent on what some of the things that some people think are optional and that ends up generating lots and lots of. And, and then we also have a sort of problem where we have different. I'm sorry. I'm having a lot of trouble catching most of your comment. It's, it's very faint and slightly blurry. If you, if you have a. Closer Mike. Can you use that better? Oh, so much. Thank you. Okay. Thank you for saying that. So what I was suggesting is that we. Come up with some Chris criteria. For what is in or out. Right. Some people have proposed that this be a set, a small set of use cases, right? Like IOT and. You know, the cloud service, you know. You know, the sort of AWS. Google Cloud Functions kind of use cases. And I'll AWS Lambda. Let me say, because I know Amazon does lots of things. Google also does IOT. And then the like, the like, you need to be able to have a gateway and we need to be able to support protocols like this specific. Context. Where we, we think that this. Specification is applicable. And then other, you know, we also had a whole conversation about user stories like that. Maybe the user stories is the, if we can come up with a set of user stories, then we can attack them one user story at a time, perhaps. And at the, I'm not like, believe it or not, I'm not a huge process long, but in this case, I think that we could align faster on the criteria than we could on the set of attributes. And that might help. So for the, the user stories is something I'm going to have next week because I'm going to compile them. So I think that would be great if I could get, get more feedback on the, the proposed set that I sent around to the mailing list than just from IBM. And, and I hope they will help and drive it in terms of what, what goes, what the attributes get included or not. I think there's a, one of the criteria is which of the attributes is actually useful for the majority of the people in the committee, you know, what, if there's, if there's any doubt whether the attribute needs to be preserved or property needs to be preserved or not. We have a process to go and figure this out. So what I would suggest is that we first come up with the set of user storage or use cases or however we want to express them. And maybe we can have like the same kind of a couple of live meetings where a few of us get together and hash out. Right. So that we have some alignment and forward smithing in advance of the call next week. And then we first align on the ones that are where we think we could not implement those use cases unless we have these attributes. Right. And anything that is not required for one of the use cases or stories then becomes optional. Right. Yeah. I think it's also, we need to think that the envelope or the context is not a replacement to the event data itself. So the, you know, you can have the use case by simply looking into the event. So the real question is what do we take out of the event into sort of generalized context. Right. So that may lead us faster to the question of transport than, and we made bucket things that we believe these will, we have to first do the required use cases. And then we do that. I mean, the transport is going to like, how does this get transported? It's going to be necessary for interoperability. Right. I believe that even for optional fields, there still needs to be a majority need, which means there needs to be, there needs to be majority consensus that the field that's optional is so important that you may occasionally leave it off for some use cases, but for a majority of the use cases, it's, it's, it's important enough to have it. Yeah. I mean, I do think. Yeah. I mean, we talked about like, you're an, I talked offline about timestamps. Right. Like that's like, it may be optional in some cases, but like it's really, really high value to when, when a sub is a sub is a sub. And then you could, you can see that the sub is a sub. You could see that the sub is a sub. It's a sub. So you can see that the sub covers multiple use cases actually picking the same name. So yeah, I'd be aligned with that. Okay. So I think what I'm hearing is agreement to potentially have the working is this extra little working group go off and look at two different things. One is the combination of at least those four attributes, you know, Sores, event type, new space. It came over the other ones were, but then also potentially come up with some guiding principles Is that a fierce? Yeah, well, actually, I'm proposing, I think Clemens and I are both talking about, like, let's first come up with a guiding principle slash criteria for deciding, and that may lead to a companion's proposal of the attributes that would support that. But doing them as separate PRs, I think, then if it becomes contentious, we can align on the criteria in the next meeting at least. Well, so I don't disagree with two separate PRs or pieces of work. I think that's fine. What I'm a little nervous about is saying that we're going to serialize it, because I don't think you necessarily have to get agreement on the guiding principles on when something is in or out before you start the discussion of how to reconcile what is source. Okay. We can let the working group decide what they want to propose. So we can just say, basically, we have a certain amount of time, and if certain things are contentious, then we won't align before Thursday, but if we do happen to align, we will propose attributes. That's fine. Okay. So now the biggest question. Did you also suggest a group split off to do the user stories, or Clemens, are you just collecting that on your own, and we'll propose, we'll send a PR or something? Well, I've sent out, I think, 11 or 12 candidate statements. I would love to get some input from people. If I don't get any further input until, well, end of the week, I'm going to go and start probably condensing them on Monday and turn them into a section here for the document. Sounds great. The more input I get, which is either adding to those statements or saying, my product goes to the following things, we basically can get some ref cons on them. Those will, that will help me to solidify that the user stories are really user stories. I mean, I have, as you see in the email that I send, already have support for it from multiple Microsoft products if I look at the services that we run, but that should not be sufficient. And even if you think you're, so everybody on the call, if you, even if you think that your contribution would be redundant, you should not think of them as being redundant but helpful. So I would appreciate a response from everybody. Thank you. Can we incorporate things from the serverless white paper? Because I think we've had use cases over there as well. Keep it, keep it as, keep it as, I mean, the, I think the candidate things are a template of, of this, how short I want that I'm thinking about the statements. So yeah, I think it would be a good idea, you know, if you want to like review the white paper and see if there's anything that is critical, but I love the idea of like, let's try to come up with a crisp condensation of that, right? So we did, we want to have it as short as possible in order to express these cases that are critical. Yeah, feel free to write three pages of footnotes, but keep the statements short, please. Yeah. But yeah, keep in mind, this isn't, you know, the only time we're going to get a chance to add things to this. Right. The list, I assume will probably grow slightly over time. So. Sorry. Go ahead, Kathy. Yeah, I have a question. Um, last time I presented a use case, I think, you know, we said there is a, there's a place on the guitar place for us to put in that use case. Is that the way, is that the place to put it or should I think to read, to review Clemence user stories and see if those capture enough detail to cover your use case. And if they don't suggest an addition, yes, or multiple additions. Could you all paste a link to that, to that group so people can look at that from the notes? That's an email that I sent to the, to the, to the mailing list. But you can, but you can still link to the mailing list from that. Yeah. I think if you can, just so people have an easy way to go find it. Okay. So while someone finds that link, um, do I have a volunteer who would like to organize and get this meeting going? So I'm happy to associate the meeting at the same time as we did this past week, which would be like 8 a.m. Pacific Monday, Tuesday. Okay. Where can I find that reference for that meeting on Monday, 8 a.m.? Is that in the, I just missed it in the, in the notes. So that we would meet in the same Zoom location that this meeting needs it. Got it. Got it. So Sara is a meeting the, I mean, the same, um, what's the link to the meeting? Sorry. I missed the last week's meeting. Yeah. So it would be at the same, at the same call, virtual meeting location at 8 a.m. Pacific on Monday and Tuesday. And the idea is that we would, you can, you can participate asynchronously. And in fact, we, it would be really, really helpful if in advance, if today, tomorrow, over the weekend, whenever people have time to add in to, um, you know, the, the discussion about your use cases so that we have collected as much as possible before that meeting. And then we'll have notes after the meeting so that people who can't, um, come at that time can participate asynchronously. And just so you guys know, I did try to put the meeting minutes into the same dock here. So you guys can read it if you missed the call. So to call in the meeting, it's the same link as, I mean, put into our, the meeting was set. Yes. Our group now, just click that on the same. Yeah. It's the same Zoom meeting. Yes. I see. Just 8 a.m. Pacific on Monday and Tuesday. Is there any way we can move this to 9 a.m.? That's up to you guys. Yeah, Sara. So, um, I have to look at my calendar, but, um, I will, I'll move one of the meetings to 9 a.m. I will note that the CNCF talk meeting is 8 a.m. on Tuesday. Or it's, um, somebody else is sure they're available at 9 a.m. on one of those days. Maybe we should say 9 a.m. Monday and 8 a.m. Tuesday. Yeah, 9 a.m. will be better. Okay. So 9 a.m. Monday, 8 a.m. Tuesday, Pacific time that is. And Climans, can you make those times? Uh, yeah. If I don't. Okay. So if I have a conflict on Monday, then you can facilitate the Monday meeting. Yeah. Don't worry. We'll figure it out. Okay. Super. Somebody will facilitate the Monday meeting. Yeah. Great. Okay. So Tuesday 8 a.m. is a TLC meeting, no, no, right? But yeah, Tuesday 8 a.m. is the TLC meeting. Yes. Maybe we can flip it. Hold on. I have Tuesday 9 a.m. Monday, 8 a.m. That is true. Okay. So, okay. Hold on. So we're agreeing to 8 a.m. Monday, 9 a.m. Tuesday. Is that what I'm hearing? Yep. Any objection to that? Done. All right. Cool. Thank you guys. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. I didn't have to ask, but I got to do it before we move on. Are there any other sort of clarifying questions people have relative to this topic? Otherwise save any deep discussions for the calls on Monday and Tuesday. But just one last chance for someone to speak up in case they may not be able to make it. All right. Cool. In that case. There are a number of the PRs that are out there, as I said, have a little bit of work that needs to be done. But I did think there was one that we might be able to discuss with you. So this is a pretty long term call. It's been a while and uncommented. And it's adding a log level to the attributes. Oh, I don't remember who this person is. Let me just see. Malhar. Are you on the car? Call. Okay. Anyway, you guys can be, it's fairly short. They just want to add a logging level or log level attributes where it has to be debug info, warning, error, or fatal for value. at this point in time since it has been uncommented for quite a while. So I think like this is like this is just my opinion. It might make more sense to align on what we think the criteria should be for when an attribute is in or out before we decide on this. I'm a little hesitant to block all forward progress on that but if people want to I'm okay with it. I just like I said I thought since this one has been basically uncommented for a while and maybe people can look at it independently of those criteria being defined but if enough of the group want to wait I'm okay with that too. But this is a log entry and I can think of very very many events that are not log entries. It's not about log entry. We have such we have that attribute in Nucleo that allows us to flag that we essentially want to create debugging info across the entire lifecycle of the event. So we can turn it on and off or change the logging level while things are you know are in production. Yeah so why if this is specific to your solution? I don't think by the way it's specific. I think that Lambda has something similar. So log level is not necessarily specific to a solution. We see it in different scenarios as well but it's when we're crossing different solutions and trying to have some sort of some compliance between the different types of logs and that's where when we end up using log level we see other people use it as well. I'll comment on this particular item. I just haven't done it yet. Okay so it's like a verbosity level of the reporting of the sort of end-to-end system. So I understand but it's it says see my objection here is that it it even says log and so it is it is apparently related to a logging to logging. This if you do if you use this for if you use events for I don't know a condition monitor. You're Clemson you're confusing two things. It doesn't say anything about the fact that the event has something to do with logs. The logging of the end-to-end sort of messaging system or eventing system. Yeah but but why would you trigger that with an untrusted item that runs for your system? It's an audio running aspect oriented thing right. It's kind of crossing a cutting across everything. This is just a piece of that domain. It's not the whole domain. Yeah I I have to agree with Clemson here. The the idea for example that I could send an event to Microsoft Event Grid which would trigger Microsoft Event Grid's internal logging level uh before going to something else is just it kind of feels very very weird. So I don't I so I want to clarify something it's my interpretation of this is not this is telling it what logging verbosity should be engaged. I'm interpreting this text here saying this event was generated because of this logging level. I think there are two different things you can always interpret this. The way we we interpret it is what is my desired logging level. It doesn't necessarily mean that someone have to respect that but let's assume I'm throwing an event and I want to start looking at why things fail. I can sort of I don't want to do it necessarily on all the type of events because certain event has failed. So I want to be able for you know for example think about serverless framework when you're throwing some a call into the the function and you want to test this specific call because all the others pass and I want to be able to uh I don't want I have system in in production. I don't want to turn debug level on all the different types of events that go into the system. I want to turn it to a specific system a specific event seem to be failing but again it may be something that people don't see a lot of value and we we definitely see a lot of value. I find this I find this rather this is unusual I think that's a very and I'm not sure I have seen this at all ever. What you're trying to do here is you're trying to you're trying to control the diagnostics pipeline of infrastructure that you don't own using a left message level flag. That's something that I so I can tell you that flag if that shows up in our in our system we would um at best ignore it we would probably reject the message because yeah that's that makes sense but that flag doesn't necessarily have to come from the origin it can come from the intermediate level but but it's this is this is something that this is something that is so unusual that I'm not sure this belongs in the standard right I can't I can't see I can't see any system in Microsoft observing that flag. So I think that I don't have the strong opinions about well I think that it's also um at minimum we need to clarify that this is not um this is not really a property of the event as I'm hearing it right this is not like like somebody said that like there was a confusion about is this specific to the event is a log event versus any event like I am interested in this particular event event type or what how class of events that um that I want to apply logging to so it seems like it's it doesn't belong with the event specification but rather some other layer that would be in the event systems right yes it's all right I'm agreeing that it shouldn't be in the spec and it should really be if we have extensions at some point we could propose this as an extension uh that just so that there'd be commonality for people who want to adopt that extension and we do have extensions we have a PR about it too yeah so it seems like there's two things here one we need clarity is this some sort of flag that it tells the system whether to turn on debugging or not or is it a property of the event itself meaning this thing was generated as a logging message have a thing we need that level of clarity and then I'm what I'm hearing is almost either way it's either no not in our spec because we don't want to control the infrastructure from outside or it may be an extension if if it falls into the other category that's what I'm hearing so far so it sounds like um you're on maybe you could take the AI to respond back to the gentleman who opened this up and ask him for clarity on what exactly he was proposing this thing be and then based on that we can we could recommend the next steps yeah I don't know actually who did it maybe it's me but I'm I know that you sort of outlined here which first it's it's uh it's the auditing thing not it doesn't say anything about the event itself and second that it's optional or extension at best so we don't have to have it here right so maybe you can comment in the PR and they get asked for that clarity first then we can figure out the next steps that's actually one of the reasons why I wanted to bring it up because I actually did not think we're gonna get agreement on this I figured it was gonna cost some some some some chatter and I wanted to get that chatter out of the way since no one's commenting on the PR so I appreciate you guys taking for indulging me on that yeah and if it's to set a log level in some uh receiving consumer system and that I would expect that would be a security issue as well for an attack vector yeah I would agree all right so thank you it would be great to capture that concern in the PR if somebody wouldn't mind commenting on it I'll add that comment then excellent thank you and so with that just a reminder please do look at the open PRs and comment on that we shouldn't have to wait until these phone calls to remind people to comment because as I said that one should have been an easy one for people to complain about I thought I thought it was going to generate a lot of talk like that hey Doug I've got a few thoughts I want to run by you and everyone else first off in the specification there is this extensions property where we could test out experimental attributes and see you know the market will tell us what they like and what they don't like and that hopefully will help us figure out what to add in to the core specification over time just in case anyone wasn't aware of that additionally I believe that we need to get this thing to version 0.1 and at this time should we be focused on adding more attributes or just settling on the the few minimum amount of attributes that we need to actually get a version of this out the door comments before I state mine I'm certainly for focusing on what we got because I think we already have too many and then we can go and see where we put expansions and additions for certain use cases and I think having them as extensions first and then for extensions to graduate into you know main proper properties I don't know how to call those I think that's that would be a good thing I think the core set that we have that we have right now that's enough enough debate let's get that debate out of the way and ship an initial set and then we can go and expand that's my opinion yeah I'd like to build on that one too I think I would like to be able to have a phase where we're focused on just the narrow use cases and then even like look at all the properties and say is there anything we can cut and before we do that perg I'd like to not let a whole bunch of new things in I agree yeah I tend to agree and but at the same time in fairness to people who do think that they have an idea for an attribute that should be part of the core immediately I think we do at least need to have an initial discussion and even if the outcome of the discussion is interesting but let's defer it and move it into extensions for now I think that's a valid outcome but in fairness to people we probably do at least need to have an initial conversation about any proposal put forward we can't just summarily ignore them for these types of PRs can we tag them as potential extensions and that way we can at least bucket them definitely but we have to have that initial discussion yep agreed I'm really looking forward to walking through all of the use cases and identifying which attributes are not necessary for maybe any or at least most of them yeah just just to follow up what mark was saying there if you come across like for example this PR or this attribute where you think it's it's it should be at best pushed out that's an extension comment on that inside the PR and if enough people say yes let's talk about this later or let's defer this or put as an extension of something like that that then becomes the proposal that we can very quickly in essence vote on during one of our phone calls and we can move it very quickly we don't have necessarily have to go into a deep discussion on the calls but get that get those comments in the PR in terms of what you want to see the resolution be for that particular issue for for the user stories and use cases thing I have one one more request that I'm just thinking about based on that comment and that is if you're using so there are a few companies here which are building infrastructure routing event all around infrastructure from the ground up but I don't think everybody does that so some of you will use existing stuff that is out there in open source and if you do that it would be very interesting to also learn about what you're using today that would be very helpful just for me to kind of get an understanding what the lay of the land here is thank you all right I think we sort of beat this log level at PR to this so please everybody comment in there before I forget I did want to make one comment back on Sarah's PR 101 because we talked about the text changes itself but we didn't we did not necessarily talk about the process comment that she had in there and I wanted just to draw people's attention to it to make it clear to everybody that the PR is not the end of the road in particular with Clemens taking the AI to do the user stories and stuff like that I want people to be clear that there is some additional work that we agreed to work on like writing up personas user stories prioritizing those user stories and then making sure that all the attributes that we add in the future relate back to those user stories to make sure we don't get scope creep and stuff like that so I just wanted to make sure people are aware of that comment in there and it wasn't missed in our discussions okay it might be helpful to put some of those use cases or user stories in our roadmap too maybe just to help rain in focus I'll leave that up to Clemens to figure out the best way to PR I can get that into our set of documentation cool all right all right with that as I said I don't think we have any other PRs available to actually talk about you can see down here is a whole bunch of them that have little comments that need to be made or little fixes that need to be made Mark would you be okay with bringing up that topic we talked about last week at the conference relative to interop sure sure at the open source leadership summit Austin Doug and myself got together and we were just chatting and we talked about at some point we'd like to see implementations have have an interoperability between them the there's an open source project that we put out called dispatch which is a framework for doing functions and as part of that we've actually implemented part of the cloud event spec I we know that it's not a full specification yet but it was something that we felt was a directional that we want to at least start that implementation with a with cloud events so this is an open discussion around when are people going to be implementing something around cloud events and then are there things that we can do to start looking at the interoperability of these services to make sure that we have a spec that actually works in practice as opposed to just in theory I'm a big fan of us trying to starting to put some heads together in terms of Dev work and trying to make things interrupt and so we're we're obviously always keeping an eye on how practical the the specification is so we're playing and you know making one thing and making another thing talk together would be great I think one the thing we're missing so if we go in and say we're going to do an interop group that just kind of explores how to go and make things plug I think I'm a fan of just you know implementing writing some code first and seeing how things work best and then come back from that code and and then write down what you did so starting so we I think of us here defining the the properties more or less in the abstract so it works for many different you know transport scenarios and that we then probably in the first stab at implementation go and do something around htp that's also why I wrote that htp strawman in the as an issue as an example and then we can probably get together and see how we can go and hack something together that's a proper htp mapping and then I'll back from that group and propose that as the the first binding the first transport binding being htp so we would begin to play a map Microsoft yeah so for that we we do need the definition for you know all the headers we need to definition to the mapping of those headers to the protocol and we may need some definition of the api semantics so mark what would be your recommendation in terms of next step here is it still sort of in the thinking stage or do you want to actually take some next some concrete actionable next step that's a good question I think we need in order to talk about the interoperability we need more than one implementation perhaps what would make sense is to have some subset of people look at are there open source projects that we could convert to omit or consume cloud events and some of my team is probably up to help them to implement that in order to show that they are interoperable yeah we are open to do a version of nuclear that supports that so and you're on we we've we've also talked in the past about having dispatch support nuclear as well yep so should we look at possibly seeing who would like to participate in a in a side group discussion on sort of putting together some of the you don't need it to make this happen so so before that i myself is this implementation um about the event producer or event consumer or both or is it like a serverless platform so what is this open source we are trying to um the open source project we are trying to work on I guess my my ultimate goal is that anything any service should be looking at cloud events as a way to communicate with other services yeah right in which case it's producer and consumer or even the you know in the case of dispatch using it as the middleware bus to move things move events between services yeah I think of this I think of this as a all interoperability projects have plug fests yes where where you come with whatever you got and then you make sure that it all works works together and I think we can go and um I would probably stage this and have first aside group discussion to see hey we need to have to do communication we obviously need to have a protocol choice and then we need to go and see what the protocol what the base protocol is going to look like for the purposes of trying things out not for the purposes of really writing a normative document in the beginning and then go back and and figure out how you're going to go and plug this in your products and then we can figure out a date and time and location to figure out how we can go and plug stuff together that could also be virtual but that's kind of the progression I think it makes sense right since we're running low on time mark would you take the action I have to set up like a doodle poll or something to see who'd be interested in and and what time I'd be able to meet in the upcoming days or weeks to have a discussion about this yep we'll do it excellent thank you very much and with that I believe we don't have time to start anything deep so let me before I adjourn go back to the agenda and the list of attendees just to miss some good I think I'm missing people eat it are you there eat it yes I am you say I'm saying okay thank you okay um I'm on my cursor's in the wrong spot here uh Grish hey I'm here yeah excellent uh Louis yep I'm here David Lyle yeah this is David okay is there anybody on the call who is not dude who does not have an asterix next to the name in the attendee list all right cool with three minutes left is there any other quick topics someone would like to bring up since we are technically at the end of agenda this is Rachel I need to apologize everyone whose name I continue on this call especially you're on I'm with you I don't just miss spell I mispronounce so I'm with you all right any other topic all right in that case oh I'm sorry Kathy were you gonna say something yeah I'm thinking you know probably know people have some use cases I'm thinking maybe we should add use case that is there centralized places for to add those use cases yes we have a use case doc go ahead and just submit a pull request to add more to it okay okay and then and of course once Clemens gets his PR in there with the user stories you can obviously add more to that by submitting a PR once his PR goes in okay so there is a use call use case document yes yes there is yes okay I'll check you that thank you okay thank you all right it's linked if the email is linked from these notes now I think yeah that's for your user story one right yep yes and and so you have all of you have that as an email if you were subscribed to the mail all right okay all right any other last 30 second topics all right with that I believe we're done thank you guys very much we'll talk next week or I'm sorry some of us on Monday and Tuesday okay thanks right