Because we'll give folks a few more minutes to join
everything good in your neck of the woods,
is a bit hectic, of course, but what's interesting
about the same. Yeah, my computer died on Monday, so it's been, it's been a bit of an adventure, but I think I've been, I've been running in a new one so it's okay.
God that would kill me I think. Yeah, it
was, yeah, there's never a good time.
No, no. Exactly. It's never a good time and God then you realize how much checks you have on
that. Yeah, no it's it's it's good, it's good to take stock every now and then I think maybe.
Yeah, yeah. Honey, I was actually going to call you the other day. Just a couple of days after, after Sonia and I haven't seen anything from you. There was no email, or anything and I thought, and I have to check if these mushrooms were okay after all.
Yeah, of course they were fine, they're fine. Well, yeah.
Yeah, that's good. I actually I just take it that part but what is consultant.
Well yeah, it's one of those things you really have to know what you're doing.
Yeah, no, totally, totally. Yeah, sometimes if you don't know what you're doing that can end up in a really fun night of what, not all of them are poisonous right.
Well, that's true. That's true, that's true. Absolutely true. So I wonder who else is going to join so I mean how's so life is good so you're in Berlin.
I'm in London. Okay, you're in Lozano,
okay, okay. All right, you were just in Berlin, at one point earlier.
Yeah that's right, that's right.
Yeah, cool, cool, anything because
I know that Zoom wanted to join but.
Oh, okay, well let me just make sure he's got the invite.
So he has it I send it to him. I'm not sure, maybe it takes them some time to figure out the logistical,
right. Okay. Well, let me see if I can wait see if I can disinvite him. Anyway, just so that he's on the, on the calendar invite and it shows up in his calendar. Now, Yeah.
Yeah, might just be you and me.
And that's actually in a way okay as well because I was, I was gonna say I mean I wanted to check what Trump's schedule is but I can only join the first half hour, then I have to jump into another call. Okay, but that's okay today because we actually wanted to ask you and obviously the other was was if next week we could maybe occupy, I think about an hour or something like that, to actually go through the use cases. Present the service model, and then really start diving into the conversation that was said to have like the intersections with the other building doctrines, and what types of wild intersections are we talking about where do we need to API's and so on.
Yeah, yeah, exactly. No, I think, I think that's fine. Anytime you want to do that welcome Taylor.
Hey guys, I've already had Maxine.
Yeah, so anytime it's good for you, um, we could do it today we could do it next week you just let me know where we can, I mean, it's something that, sorry go ahead.
Today we didn't have enough time for it, but yeah if next week we can organize them that way because I think I mean from our side I can most certainly say that we will have interfaces with payments for content management with obviously registers, and then, then it gets a bit blurry for me because I think then people probably understand better. So thinking, communications and workflow and all of these things. They're probably evolved all the way. Anyway, so for me it was just important that we sort of announced that, and then we can prepare it for next week.
Yeah, I mean, that's totally fine. So, um, you know what I, what I've been thinking, is that the, the flows between foundational authentication, and the other building blocks are pretty much going to be the same. Right, I mean, you know the timing and all of that might change but from, you know, so I think it'll be good to drill down into that so just let me know let us know when you're ready.
Yeah, so next week. I don't know what you can do about that but if you can ensure somehow that these people will be on the call, then we will prepare the documents and basically just take the time and walk everyone so
yeah well I can't guarantee. I mean it's tough to actually ensure that anyone's gonna call, I mean, maybe, would it would you rather have set up a different call and I can just make sure everyone's on there and it's just like, we're going to go through the ID spec and we need everyone from payments and every phone, everyone from registration and insecurity and just kind of like would that be better. Do you think or cuz I mean, the whole point is we can take notes, right, like I record the calls, there's like a meeting minutes. So we're trying to do it asynchronously. So it's up to you I'm happy to set up a separate meeting or, but usually I can't guarantee you like who's gonna show up on these calls
were made. I mean, rather than setting up a separate meeting I don't know I mean now that we have two other people on the call as well they can chip in there but I think the best but I just thought this, this meeting is already in the agenda of people maybe we can just prepare it with an email where we say, hey, next week we actually wants to focus on this there will be like an hour, an hour and a half, taken on that and whoever is interested, I mean it's also up to them if they're interested, they join us, if they're not interested in, there's no way we can put them on to the vitamins. Perfect. It's just from our side we have prepared it, and join.
Perfect, yeah I love it. Okay, cool. So, um, I will just, I'll send a note to everybody that we're going to focus on it next week and please join if you can.
Yeah and if you wanted to, I mean just highlight as well from our side certainly payment, register and contract management is most certainly going to be touched, but I would also expect that. I don't know I mean, did you tell me it's like mediation communication workflow I mean all of these things are probably also touching on identity right I mean we'll probably have intersections as well and that's why I started when I put it on the globe. Yeah, so that's where I want to have some guidance like those that don't do any of the other building blocks, and do we need to make that more concrete.
Okay, cool. Well, So I'll just make sure that, you know, payments and workflow and registration are available. If you know just encourage them to come. Yep. And, you know, Taylor is doing workflow so maybe he. I don't know you're super busy now tailored so I don't know if you can make it that would be cool.
I have, I have just been encouraged. Okay, awesome. Yeah, please, please do include me on the email. I, yeah, I'm running running a little short on hours these days but. But yeah, obviously dude I can be there. Also there's workflow has like, it's a phenomenal team there's like five folks who all seem to be participating. Fingers crossed. So maybe someone else could join from workflow as well.
Okay, great. Yeah, perfect. Well, that, that sounds, that sounds great. So, um, yeah. Normally we're, you know, we need to set up a separate meeting for workflow anyway but yeah if possible, that would be amazing. So I'll just, I'll just include you, and then maybe you can forward it
maybe say like allowed to invite others, or whatever.
Yeah. Okay, great. All right, perfect. So we'll focus in on that next week,
right and then just, I mean I don't know where he is right now, I'm sure he will join any second, but just for everyone's information, As I will have to drop out earlier. John, who you all met in Estonia, he's now properly on boarded into our team, and he will actually help us to put all the documents together and be sort of wrap up the identity building blocks. So I think in that respect, It's great to have him obviously, and he might be reaching out, she goes on separate occasions, just to make sure that this is aligned with the templates and the documents and so on. So there's really two you might. So if you see him. That's what he's doing and. Please welcome him. Perfect.
All right, good. Yeah, so exciting stuff so I guess that is kind of like the stand up for it, for from ID is that we're hoping to next week, be able to focus in on how it interacts with, you know registration payments, etc, and just dial in some of those use cases and make sure that we've got the interactions squared away. Okay, Yep, um, excellent. Anything else you want to cover that's related ID.
I think I'm I think that's what we'll be done next week.
Okay, great. And then, let's see, I guess. We've got information mediator. So, would you like to give an update about your, your progress.
We are in process of finishing document. Okay. And it's all Taylor we had questions for architecture as a problem is with props up functionality. Our question is, shall we. Hello, everybody, to subscribe to some events or there should be some mechanism for for authorizing.
Um, yeah. And who wouldn't ask can I add even more context around that. Sure, it basically Ramkumar bra brought up a really interesting point that was effectively that we can't do Pub Sub but from a, from a practical legal standpoint, and his, his argument was that, so, so, PubSub is a design, where you can produce events or your you can emit data, and you like by design, you sort of don't know about who consumes it. And, you know obviously within the realm of like registered valid secure applications on the information mediator, but the whole point is that you produce data and you don't know who consumes it and how they use it, and, and then you. Conversely, we subscribe to certain event types, and you don't know who produced them are by definition this is. And from this, like, Association of who produces the data and who consumes it, you accrue all of the technical benefits of Pub Sub, that's what Pub Sub is. And Ron Kumar's point was actually, no. We should not have Bob's up because no one will use it and no one would be allowed to use it. What we want instead is to design some sort of targeted multicast kind of thing. But he was basically arguing that there's no way for a global administrator of an information mediator to, sort of, authorize subscribers to listen for certain message types, and there's no government ministry in the world that would ever would ever actually emit data. If they didn't have, if they didn't know exactly who or whom they were sending it to. And so the, I mean this. I'm not the government guy, I'm more the tech guy in this conversation so I sort of felt like okay, that's, like, if that's true with what Ramkumar is saying is true then like, no, actually we need to scratch off pumps, we don't want Pub Sub. And we want something quite different, where one clinic could image information about one particular birth, and it would sort of painstakingly choose exactly who to send that information to. And this is a this is a super different thing than Pub Sub, and it might be what we need. It might not be it might be that like, no, actually, we're trying to encourage people to use Pub Sub because of all of the benefits. This was like suddenly there's really big and I thought super interesting and valid point. And so, Alexander and I were trying to sort of wrap our heads around this, but it seems pretty fundamental. And I guess the motion on the table Ramkumar, I think we would describe it slightly differently, maybe, but the from a, from the technical perspective the motion on the table is to completely remove the Pub Sub module, and to instead propose some kind of functionality that would allow an individual organization to make a specific number of post requests to a very specific set of like recipients. Yeah, and so this is this is the thing that's out there, it's a little bit of a curveball from like oh man I thought we had, I
thought we had always yeah totally nailed down, okay.
Yeah but but it doesn't make sense because I think it's, uh, you know, I, I feel a little bit of like emotional investment in what we've put together so far but like, that's not gonna help anyone. So, and I actually think that the argument is pretty valid.
Well I'm not sure I'm not sure, like, for a while. Yeah, okay, well, so I mean I guess I'm not sure I fully understand the argument, but it seems to me that
basically that no one would ever publish. So if, if they didn't know exactly who was listening to every message type, then it would be illegal to publish that message. So if we have a new birth event, and I'm a hospital system, and I want to notify other agencies of this new birth. The proposed architecture was Pub Sub, and that would be that the hospital would emit this new birth message, And then anybody who was subscribed to that new birth event would be able to receive it and update their insurance systems or their education systems or their whatever. And he's saying that I will not ever emit a new birth event in less beforehand I know exactly who I'm sending data to sort of like that nobody would ever help in Pub Sub. I think that the argument is that these governments they need to know exactly where every single piece of data is being sent, and they need to have written contracts with all of the sort of would be subscribers.
Okay, so, yeah. No, exactly. Okay, but, so I guess I thought we were okay with this, because it seemed like, if you're allowed to communicate through the information media or at all between two services that that sort of secured already right like the ability to secure communications right is there There already is like that contractual authorities provided by information mediator in the policy around it. Yeah. So, if you can talk between, you know if the Ministry of Health can talk to the hospital already then why wouldn't it be able to subscribe to messages.
B, because the Ministry of Health can only talk to the hospital because that hospital service in that hospital application is sort of been authorized by the Ministry of Health, and the thing that PubSub does is it goes up to a global level. So,
yeah but if only people from the Ministry of Health, right, so let's just be clear The Ministry of Health wants to subscribe to the hospital to find out about birth events to increase the you know incremental population registry like that's a hypothetical scenario right. Yeah, so, if then if they have communicate if they have the ability to communicate through information mediator, why wouldn't they, so as long as the ability to subscribe is restricted based on the same
IP address. Yeah, ability, doesn't mean that they have rights is a principle, it can but in case the minister's proper rights to that.
Right, so you're saying that it's because there's no authorization. Right, so you can't, you can't filter, the problem
is actually it's a it's a two sides of this problem first side is, what is in place, different understanding between for example, how I will expect it, and Ramkumar Ramkumar. Suppose that one type of event can be generated by different sources. Okay, and my approach would be is that by event type already define what our data and who is a maker of the data provided.
This is something different than Pub Sub though, that, that the that the, you know on evaluated on its merits, it may be a perfect architecture but it's, it's not Pub Sub per se.
Nope. Yeah, it's actually in this case, producer of event could be on the one.
Exactly, exactly. You're saying like, it's no longer an event type, it's just a message that comes from one particular agency.
Yeah, yeah, but it's in sales event it in that sense that you can ask to deliver it to you, or you're done. Fantastic. And this is one side, from the other side. It's not another side of the same problem but actually when hospital emits adverse events, it emits this, it sends information to pops up application, and this Pub Sub application is third party in this play. It's intermediate, this, this intermediate must have access to events, which I admit, as hospital on one side and on another side, consumer of events also need to be allowed to consume events from that third party. So, instead of point to point, rights infrastructure you actually you have rights between me and the third party, and I said at the end receiver. If you base it on X next.
Right so extra just basically provides the secure connection right so,
but security person between two points but right actually pops up situation, it's more than two,
because well, it depends on how you look at it though because if the Pub Sub system is living inside of the information mediator, then yeah,
okay, okay, it's not. Yes, yes. Okay,
so, so I mean I guess what I what I'm still not getting. And I you know I mean I guess we might as well talk about this because, unless you have anything you want to ask about ID from Valerie at this point like foundational ID. Um, because yeah we're gonna cover all that next week so we might as well, given that
in that in that case, let me just say goodbye them because I have six minutes left anyway so
exactly I mean, this might not be super, super thrilling for you but we might as well just dive into it and like try to try to get it. Okay, well thanks so much for joining Valerie, I'll send the email out and we'll focus in on it next week. Thank you. Excellent, thank you. All right. Have a great week. Bye. Okay, so it seems to me right where securing the communication channel between those two and if you consider that the Pub Sub system is built into the information mediator conceptually it's, it's, as secure as the information mediator, right. Yeah. So, if you can't, So A and party a has to have a contract to talk to party B and party B has to have a contract to talk to party a right, like they're two separate contracts. So, if the hospital, right, I
got your point if if subcon pops up is subsystem of information mediator, leaving Converse same set of rights as media and parties in normal communications area, we don't need additional forms.
Right, okay, so it conceptually it would have to be part of the security server which is a little weird, right. But um, yeah, so,
actually in in the beginning, in the 3d software extra we had a synchronous messages, Kind of. Kind of similar stuff. Store and for
what,
with, with insecurity server, so it was similar to this pops up
right yeah exactly so I guess what I don't get right is why that's a problem. So if you, if you're the ability to subscribe, is a, is a gated by the information mediator, so in other words if you don't have the contractual ability to talk to me. Then why does it matter. What's the difference between calling an API to get the birth data, and pulling it, or just subscribing for it, like I mean you've got the right to talk to the hospital from the Ministry of Health, right because information mediators granted you that right. But, so what's the real difference like I don't get that like, what's the, what's the security concern, I'm not really sure. I mean, is it just granularity that you can't restrict by the type of event or
okay i. If we change this approach and the way that this pops up is in every, every instance of security information made it. Then, then we don't have. There is no problems with access rights.
Yeah that's what I thought, cuz you know, it's basically so that's kind of how I imagined it would work because otherwise it's like yeah, otherwise it's like you're kind of bypassing the suit the security server and that would work. So I think, you know it's got to be there has to be, you know, which is tricky because if say, yeah, go ahead. Sorry,
in this case actually work this modals as event could be only could be emitted only from one source. In this model, we don't have a event, which could be emitted from different hospitals, I can just remit. Verse of child and it's the same event in this case it's not to say it's different.
Yeah, well, you'd have to go and subscribe the Ministry of Health if there's five hospitals, they'd have to go in individually subscribe to subscribe five sources yes to five hospitals right. And I think that's how it has to be otherwise you break the security model. So, Um, yeah. Okay, yeah. So, but, but
it's not the not the clean pops up as Taylor is speaking.
No, no, it's less convenient for sure but it's, it's just it's at least it preserves the security model, right, like if it's, yeah, so I guess I misunderstood that and my apologies I misunderstood that it was kind of based on the type and that we wanted aggregate types across hospitals and just have like a birth event that day. So I missed that detail, but I don't, I don't think we can do that. Just from a security standpoint, I think Dr Ramkumar is right because then you don't know who sent the event. Right, yeah. Yeah, I mean if it's at the security server level then it can be logged in, you know, it's like it's the same as responding to an API response like you could just log, oh I emitted this event, you know, consent conceptually just in, in terms of legally binding part. You know, so it's, it's, it, I think the trade off is is less, it's more work on the developer at the Ministry of Health, they have to, you know, individually subscribe. Right. But, yes, but they'd have to do that anyway if they're using API's right they're gonna have to go and call the API individually. So, um, you know, I don't Yeah, I don't think we can offer this convenience. Okay. All right, cool. All right, uh, glad we talked that out. Um, let's see. So, welcome, Arnold, how are you. Hi, good
morning, Max, how are you today.
Oh fine. So, I'm sorry we're just talking about some details about the publish and subscribe system. Do you want to give us a, an update on payments,
actually on our side, we don't have any new updates, maybe just to say that we are still working on the APS, and of course that's a little bit more detailed work so we don't have anything new. Okay, it's still working still ongoing. So it will take us about, maybe three weeks because we had meetings with Marcelo, and me force, and also someone working on PMS or vulture API's, so they are all trying to work on their site, see what they can put into the document but nothing is ready yet. As you can imagine it's a bit of, you know, well it's a bit of work.
Yeah. Yes, exactly. I mean, you're, you're trying to, you're trying to pull together three different kinds of API's in the one. Yeah, just I mean, yeah, so I take as much time as you need. Right, I mean to get the right result to get a good result. But I can imagine that's difficult to do across several groups. So, that's fine. Yeah, so I mean, what I would just recommend is if, as they become available, say, the BMS API may becomes available. Yeah, you know, just let us know and we can we can start looking at that. So, perfect. That's, that's really good news.
Except I think there's one more thing that I wanted to add is I wonder if someone from British she's connected on here. No one, okay. So Max I wonder way that it could be possible for us to have a meeting with somebody with registries, we're just presented next, because they're a couple of things that we actually assuming the registry is going to be doing. But, as I understand from our previous meeting greetings from the US we'll be looking at them in a different way. So perhaps we need to synchronize those aspects.
Okay, well let me just say, I mean let's talk about it for a minute, just so I can type it up and we got the context for registry so you're imagining
or identity, actually.
Actually I wouldn't know. Let me be honest I wouldn't know, but at least one who is dealing with registering information for people who are participating, people or even entities that are actually participating in some form of payment. Okay, and that could be a beneficiary, that could be some company that's being paid, for example for vouchers we are paying to schools or something like that. So somebody needs to be registering those schools without an outside we don't
understand.
Yeah, so. Absolutely, so it's definitely a registry and registration type task. And I think it's cool that we should look at maybe embedding a registration building block inside of payments, or something like that, too, to cover some of these use cases. But it's absolutely understood that. So there's two things that I understood last week from the payments group one is that payments is not going to be holding on to payment details. Right.
Yes, that's correct.
Yeah, and so those payment details and everything around them needs to be stored somewhere else like in registration.
Yes, and from the group we did have some recommendations, they may not necessarily have to store it, but they may need to store an alias. Right. And then in the payments. Payments building block, we do have this component that can actually look at that, alias amongst the financial service providers and ask them that. Who speaks for these past four legged say Arnauld, we do have a person called Arnold with this phone number. Is there no speaks, anyone who speaks to them, okay, I know it says he's subscribed to this bank, but can you confirm that, actually, I know belongs to that bank right, and the bank will give us a notification to say that yes, we do have details for an old, and then you can proceed and make that payment. So, in turn, in terms of you know, for registry what they need to know is that on all this area so with this particular information which bank, you belong to or which bank Do you want to be paid with. And then for us we can ask that bank whether they actually owe the same details like there's an 11 account with this bank based on the alias that they give. Okay, so probably have to hold the information but they need to hold some areas or something like that,
okay so yeah that's great so. So, I mean, if they can store the payment details, alongside that alias then that helps a lot. Right. Yeah, because then all of the it's sort of becomes like a nonce, because you could, you could even have a list of aliases, right. So yeah, then, then, then it's much much better because you're not storing someone's credit card, or, you know, cell phone number or other potentially sensitive stuff in, in like a hospital registration or school registration system. If we can just store like the reference like you're saying like a nonce, yes, then I think that's much better if it as long as we can have that sort of Vault functionality inside payments, I think we're good. Okay. So what do you call that service.
So yeah, I said we're calling you the account lookup service.
Okay, okay.
Take an alias or not. I'm just gonna say, and then account details. And, and, like whether the accounts active right.
Well, they, they may or may not specify that because we also expect that for us the account lookup service can do that, confirm that it's active. Right. I always tell them back that you know, this person is afraid to just start from that financial service provider. So,
okay. All right, so great so I'm just gonna say that, that we do have a facility to take an alias or a nonce, and that we can validate the account details and confirm the account is active, so. So what that means is like registration can pass the nonce. Instead of holding a specific payment details. Okay, great.
Sorry, I think
it's a little bit, that's a little bit loud if somebody might need to mute. Okay, great. Okay. I think that's understood from last week I think registration understand that we have to figure out who's actually going to be doing the, the implementation work for, you know like, there's going to need to be a voucher management system, right, and there's, there's going to need to be some tool where you register recipients for voucher payments and all of that, like you're saying so. I think we're I think we understand that, I'll just make sure when I, when we talked to registration that they understand that. But it's more, it's kind of like a higher level thing right it's, if we need to figure out who's going to do the work because those are modules that need to be delivered as part of the payment building block for it to be self sufficient, right. So, yeah, I mean, it's probably best that somebody takes the registration building block and does it and, but we need it. Yeah, I think we do need to, it'll be good to have a more specific conversation with registration about that. I appreciate
that, thank you.
Okay, great. Anything else you want to cover about payments.
I don't think there is anything additional we need some discuss some other things but they still eternal at the moment I think he also mentioned them last time. Yeah, like Porto, is like that, but not necessarily for the consumption of this group so there's really not a lot that's new that concerns a lot other groups,
okay, you know, I think there is one detail about payments that came up in last week's architecture meeting. So we're talking about the voucher flows. And I don't know if this is a good time to talk about it but I just thought I pointed out that. Yeah, there's also, let's see. We need to check to see if we need the ability from the outside to check have an outras voucher has been redeemed.
Oh yes, and part of the API's that were giving so that should be fine.
Okay, and then also we need to be able to cancel an active voucher in case like somebody says, Oh, I lost it, then you can imagine the admin would want to be able to confirm it hasn't been redeemed and then cancel it.
And make no
is I mean I don't know. Yeah. Does that it, unless there's some reason that doesn't make sense, but it seemed like no
it's, it is very valid, I think, something that we can add to the API's so let me note it down and then I'll take it back to the group. Yeah, yeah, I
mean it's just like, Yeah, Cuz other Yeah. Otherwise, you, you, if somebody loses a voucher, there's no way to, you know, and if you want to we can make that part of a version two of the payment specification. Right
now I think we can, it's something that we can already add that shouldn't be a problem, because we do have this API that can check and update the status of vouchers, so yeah, we can just add something for Constellation there.
Cool. Okay, great, well and then as these API's get closer to ready. Yeah, I mean, you know, as far as the, the other thing that I wanted to point out is, I mean I guess this is for everybody but the open API specs really only need to cover the, the x the where other building blocks are talking to your building block. So it doesn't need to cover every single API and all of the modules in the system right. So, it should be just enough to, you know, In the case of vouchers. We're really only talking about activation redemption. Right. And then this other checking if something's been redeemed and then canceling, that's it. Right. Yeah, I mean that's it those are the only public API's we need to focus on, right, and the in this goes for all of the different modules within payments right you don't need to provide. Absolutely 100% comprehensive coverage of everything that Mojo loop does right it should be just enough to cover the specific interactions between building blocks and the use cases that we've, we've identified and other use cases right I mean you guys have identified a bunch of different scenarios that aren't really covered by the use cases which is really, really valuable. But it's just like I don't want to. Pete, we don't need part of the power of guff stack is is not going to be every single API that you can use to make a payment. Right. Part of it is that it's a small subset of those API's and that's what makes it more adaptable and nimble, and we can expand those I'd rather be able to, I'd rather be in a position to expand API's, in terms of the open API specifications over time, rather than have to, you know contract or maintain them, if they're overly complex and there's too many API's, you know what I mean. Yeah, I hate. So yeah, I mean just try to keep it as minimal the minimum viable platform, so. Okay, cool. And yeah, and then I think it. We can also focus in on use cases but I think you guys have done a great job of that as well. So, um, okay. So, um, I guess, on the architecture side, I can just do a quick stand up, we've been reviewing payments reviewing registration and registries and trying to figure out, you know, some larger questions around enterprise well not sorry, wrong word about reference architectures, right so what that means is, how can we provide out of the box, a series of reference data models and reference workflows that are kind of maybe industry specific or country specific. So that, and I think the data models is where we want to start because if we have the same notion of a patient and GM stack ships with a reference version of a notion of what a patient data record looks like, then the chances are there's one thing to be able to connect different building blocks together but the chances of building blocks. Actually being able to communicate with each other relies on them sending the data in the same format, ideally, right, so we want to try to encourage that with a series of reference architectures that have, you know, reference data models for, you know, people, places, things, actions and then you can imagine expanding from there, based on Jason schema so that we've been trying to figure out how to bring that into the system overall. Because it's, it's, we're, it feels like we're getting really close to solving these problems of being able to connect the building blocks together effectively. Yeah, but if we don't have some sort of standardization around the format of the actual data that they're sending back and forth, or at least some suggestions that we can offer people out of the box. When they're so you can imagine somebody, one thing that's great about their current registration system is that it's designed it's code free, so a business analyst or even a non technical person or even, you know somebody working at the ministry could sit down, Build a workflow that matches their kind of system that they're using, which is good because it follows Conway's Law, what's bad about that is if they're constructing the data models, or the data structures. Each hospital is going to have a slightly or each ministry is going to have a slightly different version of what say like a patient record looks like so we want to try to figure out how to standardize these data models and provide a reference architecture so we've been talking about that a lot.
And, you know, whether or not, whether or not or how the registration, like the registry system works because right now if you look at registries, it's, it's it'll expose an existing database right and but it doesn't give you any way to modify the data model, which is fine. So, but that means that which is good but then you have to then, you know, we have to ship some sort of set of reference data models or a reference architect, architecture, I think, so people can say like, oh, while I'm dealing with citizens or I'm dealing with patients and have some sort of, you know, I want to create a table of patients, and have some sort of standardized underlying data model without having to think too hard about it and still be able to customize it right but that's where it kind of gets tricky so I'm looking at these issues, and then also looking at how we can ensure that we look at this, a UI building block actually. And the idea of that is to have a customized like standardized styling across the UI across the entire government system. And then it cuts in standardized styling and in standardized set of components and standardize interaction in terms of navigating across services and navigating between services and things like that so we're trying to come up with a baseline design language around that. And some basic wireframes around, around that. And then also something else we started kicking off is just how do you, if we if we have this demo of one of these use cases like registration for maternal and child in maternal infant care, right. How do we present that to people when they go to the website, right, how does somebody interacted so we're kind of stepping back and looking at some of these higher level, user experience things both in terms of how, you know like a government could come to the website website and look at one of these as a solution and then how we presented as building blocks, etc. So, I guess that's kind of long winded, but a little bit of what we've been talking about.
Sure.
Okay, so, um, let's see. So
Max before we go on to the next thing.
Yeah.
Any questions about any of this. Yeah,
well no, just
my computer freaked out, or Jitsi freaked out or something. and I crashed out of this meeting, right as we were talking about Pub Sub. I was just wondering if I could hear what what what came of that or if there are any sort of immediate next steps.
Well, so, I think there was a realization that if the Pub Sub system is built into the information mediator. Structurally, and if there's a requirement, say for the Ministry of Health to go and subscribe to each of the five hospitals, then say that you know as opposed to just subscribing for a birth event, then it kind of solves a problem, right, because you can use this, the same security model that you use for calling the API is, is, is used for subscribing, and, and who, you know, right, like, then the information mediator can say like, Am I still allowed to talk to, you know, the Ministry of Health, and if that certificate has been revoked because the hospital got hacked or something, then they won't, it can stop right there and not even publish the message to that particular recipient, right. So yeah, I mean, that's what, that's kind of where we were, I mean, it seemed like you had this beautiful system where you could have six hospitals start up and say like, oh we're publishing a birth event. Right. And they'd all get aggregated. But then the problem is that point to point security of, you know, which ministry can speak to which hospital at a given point in time is lost. So that was kind of what we, you know, like, if you can imagine that the security servers already like the information mediator securing the request to find out about births from a polling point of view. I mean, how is it doing that it's basically securing the request from the Ministry of Health to each individual hospital, right. So, if we do the same thing with the Pub Sub, it's just, then it doesn't break the security model.
Okay, okay, so it's almost like, and we talked about this briefly with Ron Kumar. Yesterday it was almost like it is, you know, if, if Pub Sub has sort of at its very most basic implementation, it has like two keys right there's like an event type and data event type and payload or whatever, we're basically saying, we can keep the security model where you have to provide individuals sort of signed access for all. For anything produced by or going to any of these individual agencies, so long as we add like a third mandatory attribute in any Pub Sub notification so event type payload and, like, producer, or something so it is like something that actually identifies who produced it and that will allow us to like keep the same security model is that, is that kind of the idea.
Yeah, yeah and you don't even need the producer because the producer is employee so the, the, the person who
event type contains all the producer so each event. Event Kobe generate that by only one producer. That was solution. So it's new
birth, hyphen, is
the hospital a new birth or hospital be your worst from hospital see. And the second decision was our proposal is just we this application pops up application which we saw, it's some where it will be actually in every producer co ops something.
Right, so each producer so each. So essentially, we're talking about the embedding it into the security server, right. So, if you want to. And then that way it uses the same security model that you use to just call. To call regular API's now, right. Um, and it has the same kind of logging and, you know, data integrity protections that right that that that information mediator gives you.
Yeah, every security server has instance of pops up running application on on this secure.
Right, exactly. And so, so you basically can only, you're only going to be allowed to subscribe to something. If you have a contractual relationship with the hospital and the hospital, you know, so you have it solves these contractual issues right because you won't even be able to subscribe. And then, you know, If we want to make it really really secure when the event is published, it can check to see is that recipient Do I have permission to send to talk to that recipient, right. So if the, you know, Ministry of Health revokes the certificate that allows hospital. A to talk to it. Then Hospital A can find out and, you know, just drop that on the floor and stop sending those messages right. So kind of ensures that the, you know, that knowing who sent it is, is still kind of explicit.
Right, right, right, and then, yeah, and then, then you cannot. So if I'm, if I'm insurance ministry, B, I cannot sort of unilaterally, choose what I subscribe to. Rather, I have to attend sort of authorization to subscribe to, you know, hospital A's events, is that right and then and then we're sort of all, all happy, think, right,
I think so yeah so exactly so you have to get contractual. I mean, it would be the same. Probably the same contractual relationship that you'd have with Hospitalet that allows you to call API's would also allow you to subscribe for events. Um, but it would just point to point like that, right. And then, likewise, when you receive an event from Hospitalet in the future you know it's it's coming directly from them, right. That way, because you subscribe directly to them. So you kind of implicitly know who's sending it.
Yeah, it's a bit different to hear when, when you, when you are asking something active, active site is you. And this external access system is very good. In that case, receiver of request. Check. Does the party who, who initiate this request is, is this party allowed to do this request to him or not, we have this, this, this scheme in extra, but in pops up, it be different active party is is emitter of data, and when the hospital sends data to ministry in normal situation on side of ministry we are checking his request allowed to this hospital or not. But now it's a bit different. Now we at hospital need to check. Can we send to ministry or not. Yeah, exactly. If this is different. Right, right, metrics, then in basic Excel. Right, yeah. It's something conditional. Right,
Right. Um, yeah, so, um, yeah, understood, but I mean I think you know it should be doable, and if we do that then we have like the Pub Sub and security, it's just not quite as convenient as having a single place to go and register by event type you have to go register by event type for, you know, in, you know, in different places, right. Okay. So, um, what else should we cover any anything else you wanted to say sorry about that. Sorry you got dropped there
is my browser. Yeah.
Okay, cool. Does that make sense. I think that kind of solves it because I don't want to get rid of the Pub Sub, you know we can't get rid of it. Like,
right, right, so it's just, it's, it's sort of it's, it's, it's, I guess what I had heard what I had sort of, You know we were working on like single, you know, single ministry applications. The advantage of Pub Sub is your publishing to a effectively to a to a, you know that the subscribers are like real authorized good actors and I guess what we're trying to do is assume that they're all bad actors from a technical perspective. And because it's across different ministries and so is basically just, I, I had thought of requiring knowledge of the sender and requiring knowledge of the receiver as being equivalent to saying no Pub Sub. But I guess what we're doing is saying, no, no, let's keep. Let's keep thinking of it in terms of Pub Sub it's just like gov Pub Sub, right, It's like, you know, we're adding a little limitation on Pub Sub where we have to know the provider, and we have to know the receiver, and we can, we can still call it Pub Sub, we just need to make sure people know that it sort of got this asterisk and we think we can probably implement something that does what Pub Sub does but has, it has like a binding between a publisher and a receiver so it's like a, we can call it Pub Sub but it's more like this cognitive multicast thing rather than publishing and subscribing. You're like, you're specifically signing up and agreeing on these like one one connections, and it can be facilitated through this like one to many interface or something.
Well yeah, I mean, you know, but keep in mind that information mediator has like a directory of services, right. So you could, you know you could alleviate, or at least have some sort of sugar around that that fact that you have to subscribe to five different hospitals by just having some sort of, you know, having the SDK, look up in the information mediator and say like okay well who are the five different hospitals. And then, you know, just get back a list of, you know, URLs, but yeah, You do have to subscribe to a specific place but then once you've done that, they know who you are, effectively, and they're going to basically post back the response directly to you anyway. So, yeah, it is a little bit different because you do kind of at least have to. It has to be a point to point relationship but I mean, you know, conceptually, maybe not right, like the information, if it's built, if the pump, publish and subscribe system are built into the information mediator. I mean you could still have it. That you're subscribing for health events and then it goes and looks up at all of the all of the endpoints and it individually subscribes you know to all five of them on your behalf, right, it depends on where that's happening and then in that case, then it is kind of true Pub Sub. Right. I think the difference is that instead of having like a centralized hub. It's kind of distributed amongst all of the all of each security server has its own, you know, each information mediator, has its own publish and subscribe system, but I still think we can meet, you know, and what that does mean is like at some level, the mystery of health is not subscribing to a birth event, they're subscribing to a birth event on at 10 different hospitals individually, but who actually does that I mean that could actually conceptually be done by the information mediator. Right, based on some directory lookup, Yeah. You know, it's still, there's, there's still a lot of issues with that right because then what happens later if there's a 13th hospital is added right to that does that automatically they get automatically subscribed or, you know probably they'd have to right otherwise this would not work so it adds it adds a bit more bit more complexity. So to start, I mean, it's not true Pub Sub in that you do. No need to know who you're subscribing to the events,
you know, like who receives the messages and who's producing
them. Right. And, and, but we have to do something like that I think just to preserve the security my security model, right.
Makes sense. It makes sense. Yeah, I think I was just, it just took me a while to wrap my head around it and just sort of see that it is in a sense it still bugs out it's just got this interesting sort of Gov, government, government, you know, either enhancement or limitation or asterisk is just a tweak to it rather than Yeah.
Yeah, I mean, maybe we shouldn't call it Pub Sub. We could call it something else like, you know,
but, you know, throwing around this term targeted multicast, But I don't think that necessarily is going to sell a lot of T shirts.
Yeah, well, you know, yeah, we can think about it, I mean, you know, if you conceptually though if the information mediator had some sort of, you know, way of saying hey, You know when a server starts up, or when it, you know, it could say, you know, send a broadcast event and say hey you know I send birth events. Right. And then anybody who was, you know, subscribing for birth events could, you know, then subscribe to birth events from that specific, you know, information meet that specific instance, in addition to the ones that they already are subscribed to. So there is a way to make it work but for now like that that's super complex and hard to get right. So for now, I think we should probably just expose, you know, put the onus on on the developer of knowing exactly who they're subscribing to me kind of like they have to they have to know exactly who they're calling an API on so we can add Right, right right. But I think if we can add some syntactic sugar in the future that would kind of, I think it is technically possible. Yeah. Yeah, so, um, yeah, to make it nicer, at least from a developer experience and still have this, you know, but it's, It's more, it's more tricky, right. Yeah. Cuz we want to keep it decentralized, that's the key, and point I think
it's I think it's still, it's still can. It's just, you know, there's a, there's a permission and checking step right like and that's, that's, I think that's totally cool though because that there's still, you know, there's still going to be a sort of a Pub Sub management portal, and maybe it's part of the general like list of who has access to what services which applications are authorized to use which services provided by other applications, and there's just sort of an extension of that so it's like okay so you know, either. That is literally the way we authorize events subscription or there's just another thing that looks exactly like it that up at the global level is like registering, these contracts have been made. I think it's, yeah, I think, I think it works.
Okay, cool. Well,
I guess we'll have to Yeah, I wish, wish rom Kumar was here it'd be good to hash out with him as well but I can follow up with him. Okay, cool. Uh, so. Alright, so yeah, I guess it's Yeah. Point to Point publish subscribe or something like that,
something. Yeah.
Yeah, okay. And I mean, the nice thing about that is, uh, you know the legalities can sort of be addressed. I mean the way they're addressed by by expert or information mediator is by logging every request and response. So, you could also log every like event that was published and who is published to write on both ends, right, you could say like on the center, you could say I publish this event to these five people and then on the recipient, you could say I received the good log, oh I received this event from these, you know, this source. Right. Yeah.
It will be over the log, my underlying system.
That's true, that's true. Yeah. Yeah but I mean that's, that's my point is, like, because we're doing this point to point. Then we get all of the data integrity and logging and all those benefits that make it kind of more legally binding that make, you know, enforce that make the, you know, date, solve these data integrity issues. Okay, cool. Uh, well, let's see. So, um, I guess I wanted to ask people, maybe we should. I'm gonna just go through last week's key decisions. So everybody should everyone knows about the new architecture blueprint PDF.
Whoa, cool, cool.
Yeah, so I'll post that in here in the chat. So I'll just you know reminder
PDF by
version and section number. Oh it's it's it's a bit rough like the logo looks bad and I don't know it's missing some stuff, but um, it there's a version up there. So at least we can you know start referring to
this press kit there. Yeah, the numbers. Yes,
exactly by number and version number. So, you know, say like version one Dotto what out one, you know, right section two to three. I think is enough. So reminder about that. So the other thing is we talked about terminology last week, and I don't know if everyone was here, but I'm just share my screen for a second. I'm going to. Yeah, I'm on a new computer, and I have to authorize access to share the screen. And so what that means is I'm going to I'm going to have to quit and restart so maybe I won't do that, I'll just do that later. But anyway, um, I just wanted, I'll post the link here instead. Um, Let's see. To this terminology section so it's just a reminder that groups should definitely feel free to update your terminology here in any terminology that's needs to be understood across groups, please go ahead and add it directly into the architecture blueprint. And then we can review it, right, these are missing from the PDF right now cuz I didn't review them, but now they're reviewed so the next version of the PDF will show up. The cool. Yeah, so let's try to hash out terminology that should be understood across all groups there. So reminder about that. Okay and then another thought, I guess, is just how do we deal with and this is a question that Ramkumar brought up last week, how do we deal with minimum performance requirements for each building block, right, and maybe there should be some in the blueprint as well. And I don't really know how to I mean, it seems like it's so specific. You know, I mean, we went ahead and added one. So I'll just tell you what we added to the architecture blueprint, just as like a baseline, you know, requirement. And that was that. Let's see. Cigarettes API's must return an API response within 60 seconds. I mean does that seem reasonable to people or is that too short.
Sync synchronous API's
are just like API's, may it should just say API's must return a response within 60 seconds but I mean, how does that strike you all,
I mean that, like,
literally there's a request and the response to that request yeah I mean 60 seconds is feels very long.
Yeah, it's super long right.
Yes, that would be, that would be like the case where I would be advocating for a 201 accepted right so like, in, in, in 500 milliseconds saying, Okay, I gotcha. I'll come back later and like, Here's your ticket number or something.
But it depends actually on request if you are requesting the whole database in one file. It's not enough even 67
Yeah, exactly right. So, you know it's a thorny question. Yeah, but I thought we should have some sort of at least baseline established for the entire system. Right. Yeah but I mean, I think, you know, I think, I think, I think this is exactly right, Um, you know the the that this notion of a ticket with a 201 or something right.
Maybe, maybe we put here. This way, you response to the request will be longer than five or 10 seconds, then you need to do it in two phase tickets, and then based on this ticket to get
actual results.
Okay,
ticket that is resolved by polling or something.
Or it's like guideline, I don't know how to how to.
Yeah, you should you should return you. I think this one, I mean this would be like, every response will take longer than five seconds you should return a ticket that's resolved by polling separately. You know, A ticket with a suggested callback time that's resolved by polling so, you know, I don't know I don't know. Right. And then you can say, Okay, try back, you know, try back at this URL in 10 minutes, you know, and then if it's still not ready, you can say okay try back in two minutes, right, or something like that. Okay, I mean, I don't get
this feature I think we, we fell say that it's not provided automatically, it's its responsibility of each, each request itself to this request should be designed this resist configuration.
Yeah, exactly.
So nobody can can can implement it if it's initially done differently.
Yeah, okay, well, so I'm just gonna throw I'm gonna just say, you know if the response will take longer than five seconds you should return to a ticket. With a suggested callback time that's resolved by polling or something like that. Because I don't know how we can enforce that right I mean, but at least we've got to, you know, kind of plan there, so I'm going to resolve, I'm just going to put a comment in here I just want to.
Okay. So yeah, it's always good to review these key decisions we made. Let's see what else you know but also I guess this applies agent each building block right like each build at least this way it shows up in the cross cutting requirements and each building block can override that if they need it, right. So maybe there's some sort of like payment that always takes like you're doing some sort of Bitcoin payment and it's super energy intensive and slow and it always takes five minutes, or takes up to, you know, right, for the actual. So I think we covered that. If we just have this pattern. But I think we have to kind of spell it out pretty clearly like exactly what the responses look like and whatnot. Okay so, and then another thing we talked about last week is that we need to figure out how to Mark Fields in data models as sensitive, so, you know, we want to prevent them from leaking via fraud or you know even leaking personally PII, personally identifying information. So this would be like, for allowing extra caution downstream. So, you know like credit card field could be marked, so that it doesn't get logged, or a password field right so they get blind by the logging system. For instance, right. But, uh, you know, this is sort of mitigated by having a secure information vault. Right. Like we're talking about with payments they have this ability to look up, you know to validate given a given a handle or a nonce or, you know, some sort of unique identifier, an alias. Yeah, you can look up, you could, you could say like okay I want to make a payment on behalf of this alias, or this nonce. So that kind of mitigate some of this but you still, because then you know you should be able to print the nonce out to log in, it doesn't matter, right. But still, so like and this came from this discussion last week around voucher payments right like we can't be leaking the serial number two logs, but we have to have the serial number and the transaction, you know, like a couple of numbers to print out an actual voucher right so like the information system has to get that voucher, all the these kind of sensitive details but if they're logged in some way that some, you know, Tech Support Assistant can see them then there could be fraud. Right, so I you know, just a reminder on that I don't know if we have any sort of, if anybody has any particular knowledge or experience with that but it seems like to me the way to do it would be to somehow in the JSON schema Mark Fields as sensitive, and then just, it's kind of up to whoever's doing logging to, you know, be mindful of those fields that are marked as sensitive, which as I'm saying it sounds like a long shot right like who's actually gonna do that. Okay. Any questions, comments about that.
From from mediator point of view, actually, whole content. All payload of messages sensitive. So from that perspective, there is no reason for marking marking some fields more sensitive than others. I would say, if it's some mother loving. Yes, maybe, maybe it's reasonable. Yeah. Do we do we plan a long building block or something like that.
Well, we're gonna need a monitoring building blocks. And we said like there might be a log sync and a metrics and can then
log see, we have locks in. Yeah, of course, for lock sync yes, we need this sensitivity
flag. Yeah, exactly. And it's not just a log sync I mean you can imagine it were one of the things we're talking about is like for payments they need some sort of registration tool that allows somebody to provision a payment or provision a voucher, right. And if that's a paper voucher then they need some way to generate and print out a barcode that has all of the sensitive detail that you could use the cash and the voucher. So certainly that tool is going to have a log of, you know, the fact that some admin created a voucher. At that point in time, right. So, later, when the person comes back, They can say, hey, you know, go to a customer, or call the customer support person and say like hey, you know, I got a voucher and that person could see conceivably a log of all the vouchers that were issued. The problem is, if they can see the sensitive numbers that were used to create the voucher to print it out, then they could then there's potential for fraud. And it's kind of related to this issue with, you know, PII as well, right. Like, you don't want to leak something personally identifying and, you know, some sort of sensitive medical data either, right. So, so you can mark all of the fields that are sensitive, and then maybe, you know, we've got a better chance of this not being as much of an issue. Okay, yeah, no, that's a good point. I mean information mediator, yeah so but it's sort of like at the record level or field level. Does that make sense. So, okay, and then, oh, the Reply All issue has been resolved in discourse. So, and everybody's got their own email addresses that you can use I just posted those. So, supposedly, you can reply all to the same topic, or you can reply the same topic more than once, like that was an issue that you were you reported. I believe Taylor's
i Yeah, that was my issue and thank you so much for fixing that it seems to be working perfectly now
yeah I don't take any credit. It's Steve.
Oh, yeah, sure, Stephen.
Yeah, I'm not gonna take, I'm not gonna, I'm gonna steal his sunshine. But ya know so it's extra thanks to Steve for that getting that set up so, but just so everyone's aware, we don't have a good solution to give groups their own meeting notes topic. So we suggest that in for now groups use an email address, above and then use a subject like meeting minutes with the date in it, and then it'll have a unique subject line. And then, you know, but the problem is that there's only you can only have so many levels of topics and discourse. And because we've got each building block under a building block category then it's too much somehow to have a meeting notes inside that, you know, some topic, I don't know, I don't really fully understand it. But the goal was to have an email, folks wanted a separate meeting minutes, email that they could use that would go to a separate meeting minutes topic, but we haven't figured that out. So we could just suggesting that groups instead use the email Aedra email addresses there, and just use a special subject line to separate meeting minutes out. Okay, so it's all going to be kind of one flow. And so let's try that if that's like a big problem then, You know, let me know. Okay, so a couple things that you might have missed from the past, all the open API specification should be published as JSON, and they will be merged in the master repo. So we want to structure them so that you're forking the main gov stack repo, adding your API, open API specifications there. And then those get merged into master and that's just so that we can have this sort of centralized approval flow for adding new versions of open API specifications and versioning them as a group, and things like that. So we can say like this collection of everyone's, you know, specifications, regardless of the work that you're doing in your own branches, is the official gov stack version 1.0 of those open API specs and, you know, And it's an entire collection of everyone's specs right is the idea. So we can version them that way as well. Using this approach, and all your resource models, must be done in JSON schema format, so that's open API 3.1 Slash, and one of the weather changes and open API 3.1 Is that all of the data models are using a JSON schema format, so you can actually combine your, and this is one of the nice things about procrastinating because the tooling should finally, you know, we, you know the tooling should finally be catching up to open API 3.0
Yeah, I'm so cool. All right. That's about all that I have, I guess I wanted to ask, Does anybody want to publish a draft PDF specification. Like, it can be rough, it can say alpha on it, it doesn't matter. If you do,
when we when we get to the bottom of our current round of reviews, I mean Alexander, this is, you know, ultimately your call on this but I feel like info med when we get to the bottom of this round of reviews, we could maybe throw something out there that says,
I mean if you look at the PDF like my PDF my first publication was pretty embarrassing like the logo looks bad. There's a bunch of stuff missing, and that's okay. Right. I mean I think the sooner we can do it the better even if they're a little bit rough and raggedy and not complete. So just let me know because then when other people are referring to your you know your specification, then they can refer to it by version number, and by heading. Yeah, instead of some random google doc link. Okay, cool, um, anything else folks want to cover Ramkumar you yeah I'm glad, glad you glad you were able to join.
Yeah. Max, I just wanted to check one thing today, I see the slot at eight o'clock right i mean another hour or so is that meeting there.
Let's see. Yeah, yeah, yeah, that's right, yeah 530 Yeah. And I think I you know I don't think anyone's gonna make it except for you and me and Alex. Select only be accepted it so it, it might be too late for people, right. Um,
yeah, so, so we'll see
how fast. Yeah, we'll see. I mean it's a little bit late but I was trying to do that. What was my rationale, oh my goodness, I thought, Steve was going to be joining somebody from the US from the west coast was going to be joining but maybe that's not happening so we should move that to an earlier time probably it's pretty late.
Hi, Steve, it's not even in the in white.
Yeah, yeah.
His name doesn't appear when they might also,
yeah exactly I'm just looking at it I don't see his name in there so you know it's probably too late, um, I mean I guess the question was I didn't want to, you know, this meeting is scheduled to go for it can go for quite a long time like next week we're going to be doing the IDX working through all of these IDX things so they can be a long meeting so I mean I guess that's why I scheduled it for 530 so it didn't collide with this one. But yeah, It's still on there. So yeah, yeah,
most political person I think is creased on this list.
Yeah, exactly. And he I think didn't he just decline. The, he just declined it. So,
even later meeting, when we had it that at eight o'clock, or, or after chain exam time. Yeah,
no after changing the time so, um, yeah, So and then he declined the cross team building but yeah so he declined it, so he is not going to be there. You know, I think I actually think it's okay, I mean I know what he wants, like I'm kind of like, representing him it would be great to have him there. I'll just, I'm just gonna like. So let me just ping him. Tell everybody what this is we're trying to do this, you know, just talk about prototyping and how we do prototyping and this kind of rapid prototyping stuff that's been buzzing around. So just so everyone else has context on what we're talking about here. So I'll just, you know I'll ping Christo and just see if there's another time that works but you know I think we can still do it. You know,
I think. Yeah, I think we
should. Yeah, I think we should just still kind of move forward.
Yeah, yeah, because otherwise, it's only you can elect, Alexander, we can still continue with it.
Yeah, just like, you know, okay, so, yeah, I'll check in with him. Um, alright so anything.
The other thing, yeah. I just think that we, as part of this meeting meeting agenda max that is review the Excel chart updates of milestones checklist.
Oh yeah, last
week, you know, right now we have only done that done kind of thing. Last week we added saying, Okay, what's the current activity. So I think, in the standard meeting we're also capturing that aspect of what's happening right now. Right. It will help to actually fill that in. So what if we just review that every week, so people can actually post. Our their status this.
Yeah, yeah, no, that's on the agenda, we just haven't gotten to it yet. So But absolutely, we can do that.
Okay, so if it's getting reviewed here then there is like you know there's a discussion around the topics that have been pushed there, so it's easy that way. I think it will, it will, it will come in the minutes in your meetings if we do it, but you know it's cumbersome to pick it up from the minutes and add it to this Excel charts and so on.
Yeah. Okay, so yeah, I mean, so, so it's supposed to be something that people kind of update on their own, but let's just go ahead and is there anything else. So yeah that was actually the next thing was to look at the milestone checklist so you're we're right on schedule. So unfortunately I cannot share my screen, Because I'm my old computer died and I'm on a new computer and it doesn't have permission without dropping the call.
I can share it. Okay, cool. I
mean I think we've got, well we basically Arnold, Arnold, very tactically dropped off. So, we just have very politely and very tactfully and very tactically, so we just have you and we just have architecture and information mediator. Okay, so that kind of like, Yeah, but I think everyone was updated. So,
yeah, so So let's send a note, just so that people do update this next week because we will, we have been sitting with Alexander and Taylor, so we have the same forum here so it just a bit, but at least let us take a look at this list, the shape is one share.
I was gonna say we may know our status but I don't think we've updated this Excel sheet in a number of weeks, so if you think it's useful as an exercise we maybe the three of us could just run through it super quick and make sure we're not missing anything. Yeah, let's do it, because we haven't, we definitely haven't touched this sheet in a while. Yeah,
no Greg's I need the sheep because I'm recording our bases this sheet, too, and people are comparing this sheet to whatever overall progress they're looking at. So this is some kind of, we agreed on to so to stream like this. Okay, share screen screen are able to see.
Yes, inside a group.
Okay, so let's go to. Where are you based.
Okay, so this kind of. Dun dun dun dun dun in progress, if we just see what we haven't read this first let's see if they're already done then let's mark it up.
Shree, should we run through from the top real quick so identify use use cases that may use this PV yes we've done that new use cases if any found for this BB defined in template and linked to this VB. What is that exactly that, that doesn't have a done yet.
No, I see we haven't discovered any new use cases.
No, not no one.
Yeah, So, so, let's leave it blank.
Do you want to just mark, like, at, like, I don't know, NA or something.
Yeah, no,
no, no, no I maybe it's better to leave it open, never know what use case.
Yeah, so what should our, what should our target date be then maybe something.
I don't have a target date but we can just keep it. That's what it is. So, be open.
Okay, identified BBs capabilities needed for each use case,
I think. It
can also can I think we have to take a call because it's better to put what use cases we have covered, rather than saying, Oh yeah, each use case, because I don't know what our data is so fast we have to. So you have we addressed all the use cases. So far as information media is concerned, there'll be no other new answers. If we
Ramkumar Auerbach, don't depend on use cases at all.
So yeah, so it's kind of done right or you could almost take these out because they don't apply.
Okay, so let's mark here.
Say that you want to get something else. Okay.
In progress we need to put reference numbers so they're referencing to global costs, broadcasting requirement.
Okay so in progress in progress, so can I say current activity.
Formatting.
Exactly. Date is gone. So,
put one month, one week from now,
let's say
we max. Prepare a PDF so we can now proceed with
it. You don't see the sub items there functionalities required to realize all capabilities, identified. We are still discussing the failsafe and error handling. Right, right,
to realize our capabilities are identified, AI
capabilities are functionally.
Yeah I mean I think
we can find them fair but we have
proposals for, you know, extending those things right so I think we're sort of in this current round of review that we're going through. We're sort of I think there's more to discuss on error handling and things like that, Is there not.
Yes, yes. So maybe in progress. I mean maybe you identify it, but we have not defined it. So, let's say in progress.
Yeah, I mean there's also like this Pub Sub stuff that we have to, you know, Take these new ideas. In,
for example, conclude
your next setup. So this unique set of functionality basically meant that you are identified in different use cases, the same functionalities. So out of that, all the common stuff. See what is the unit. That becomes the API of this. I think they have done that, Yes. Yeah. Yes.
And the brief description of functionality. Yes, done.
Yeah, we added those descriptions up at the top,
any quizzes for use this functionality. Okay so this is the precondition stuff, depends on specific use cases. So you have to
look at it from your consumer point of view, do they have, is that something readable that they have per day that we have given how to use this functionality.
I think we've certainly specified. The, the, sort of requirements around being able to, you know, just fictional stuff we have. Yeah and like being able to make a post request and things like this. There's also a bunch of implied prerequisites around administration. Knowledge of SSL certs camp certificates knowledge of encryption and things like this that are more like on setting up the actual servers, but I don't, I don't, I'm not entirely sure that like if we spent if we spell out those requirements, I'm not entirely sure we then need to sort of explain for each of those that like you need someone who understands this domain, etc.
No. So have we listed the prerequisites to use this functionality in a way that people are aware with before they use it, they have to have all those prerequisites ready. That's what we did have a section right in all these definition documents prerequisites is on the top of that list. If we are done with that list let's say
I'm trying to find it right now.
Maybe copy request a precondition to get the exact name.
Sometimes we're calling it assumptions. So I tried different things in different documents. But Max had marked it out as pre requisites in the basic template.
I think we only have one mention of wreck quiz. Yeah, we don't we don't have it, we may have, we may have actually forked our IMDb spec from an earlier version of the like architecture spec. So I don't see a, I don't see a prerequisite section anywhere in our document I our table of contents is is auto updating, by the way. And so you can pretty quickly look through and I don't I don't see anything in there. Prerequisites should should we add that.
Yeah, I think we should review this. Okay.
Open progress.
Yeah, and then maybe an activity notes is max, could you drop a link to just, just to make sure I get live in the wrong one to the most current building block specification template.
Sure thing.
It haven't changed, I hope. Well,
we did, because it has this prerequisite section, and
at least I know, for example, I can look at payments payments, guys. The only
place. The only place we've got prerequisites is for workflows. But here's the here's the, you know, and we're not asking people to update all of you know, update your templates to this new format because that's just crazy. Um, but, but, I mean you should look at it so that you know like, What the What's, what's in it. I'll just paste in the chat.
Because it's like you had the section called key decisions right added like that this was added.
Oh yeah, that's right, yeah there's key decisions yeah there's key decisions and then if you look at that link, um, there's also a, you know, a key decision log at the be at the ending but you guys already have that. And then, at the beginning there's a version history from that but that's for publication, right. So,
just the max just check what two payments guys have given. I put a precondition, a prerequisite to something one one. Okay. I mean, then, in any case I think that should be bad because you are assuming certain things will be there for us this
way Yeah but I mean, in terms of in terms of workflows Yeah There absolutely are prerequisites like, I mean if you look at the, the use cases right. And the workflows, those, those all have, you know, so maybe you're thinking about this use case template that we have. But that's for workflows.
Are you wondering the context of this saying.
But yeah so and when I look at payments though, um, the only prerequisites I see in their spec are for workflows, like, you know, if you're gonna do a payment we were there, their prerequisite is the person has a bank account, or, you know for stuff like that, but those are workflow specific. So I'm not exactly sure what this means.
Yet, this is like that, see prerequisites to use these functionalities, right before that we did a unique set of functionalities and now. Prerequisites may be that. Yeah, so to use this functionality which means basically one of the aspects of this building block, we want to use like for example if you want to use Pub Sub if you want to use service orchestration, how do you use this, what are the assumptions before which you get to use this thing.
Yeah, no I think it makes, I think it makes a lot of sense. I'm just saying like I don't think there's a place in any of the templates, other than the use case template for prerequisites so. So I mean just, you know, maybe just, you know what, maybe it's maybe it's not just the functionality, it's like to use the building block right.
Yeah, yeah. And when you write it out, you're actually writing in how to use this building. And maybe there are some unique requirements based on that functionality. Okay, some, some things may not be necessary for some functions. So if I'm using information mediator only for a service orchestration and I don't need the pops up in my implementation. I don't know what assumptions I should have to use the service exchange. I mean we are discussing this when we are when we are having these technical specs, I think a lot of discussions happen around it but we are not spending yourself. I'm just looking at it from external perspective, somebody comes and says okay guys I want to use this building block. By the way, and then he goes ahead and then, oh, I got to set up all these things before I start using your building blocks. So what are those assumptions.
So, maybe therefore, one of the one of the places where we have to have this is to map it to use cases right. Instead this document. Like Ingmar has a table where it's saying that this particular functionality maps to use case number two, Blah blah blah. In his template. So I just, what do you think you need this, don't need this prerequisite, he goes in for immediate.
But if it sink in general, from general perspective, actually, there is prerequisites to use media to the blog. Better off system.
Yeah,
yep and there needs to be like certificate. Certificate Authority and all of this stuff.
Yes, yes, yes.
And then this we discussed about creating this in our explanation we have put there that there should be,
and we have have explained it in our paper so maybe put down
here in. Yeah, this will make it very crisp, because it is all over the place, or wherever we are encountering such thing we have been explaining it, but there's no place that says hey, only use this, please check. There's no checklist.
Yeah, there is no checklist. You're right. Maybe, maybe. Okay we can do this. In that case it's in progress.
Yeah, so let us say we have to pull it together.
So I'll add a checklist. Yeah,
from across the document.
I mean, another way that we could consider doing it rather than moving stuff around is just like if you highlight it and put the word, you know, a prerequisite.
Yeah, maybe put it in a link it. If you're explaining it somewhere in the document,
or something like that rather than move, You know, I hate to, like, try to restructure the document I mean it seems like you could almost have like a pull out that's like prerequisite, you know, and it's, and then later you could have a table of those free work prerequisites or something.
Or yeah, so add to that in this just need a list.
Okay.
Okay, so in terms of functional blocks defined. We haven't done that internal workflow diagrams laid out. I still think pops up is not normalized. Yeah, in progress in progress. So web sequence diagrams let's say are in progress.
Interface definitions.
I already corrected. No, sorry.
You're saying the current activity can conversion to web sequence diagram, converted into USD.
Yeah, vacuity consolidate
interface definitions interaction with external applications.
Open API. Calling convention
progress. We will create integration schema.
So this, whatever is in in progress, let's just give it some date here, this one with activity stuff. For a quick word where we start to
really
defining these things can we do it move like one week or next week or corner take a little
break though two weeks for permission.
Yeah I mean Taylor has been asking me to take on some of that and I've. I'm like, I think we're both struggling to find that time so we got a, uh, yeah so I you know that might it might take, take a little while, um, just because there's so many you know Taylor spinning up worker the workflow and there's all these other stuff going on
with like, Are there any are there any sort of shared resources like I know that honey was saying he's just started working with a new developer based in Geneva, who, who is quite technical. I wonder if like this, this kind of thing. Like if diagrams were provided in Lucid Chart, or in PDF, if there was somebody who sort of is cutting across these different groups but isn't doesn't have their hands completely full if they could, like, you know, just chunk those things out and
he was mentioning this guy can call for you guys.
Yeah I mean I, I think, I think that but I think that it's a good point, we need more resources. Um, so yeah so I know, I know somebody who could actually probably help us out quite a bit but I'm with with with this, this, this, this, you know, I mean we almost want somebody who's like, you know, a technical but you know kind of focusing on like the documentation side of it, not the coding side. Right. I mean so what do you what do you what do you I mean let's just have it, what do you need support with like you know what I heard is, so there's the diagrams, right. You need somebody to harmonize
outputs, some of them are PDFs, some of them are amputees, some of them are coming in some other. Yeah.
So there's normalizing the outputs, there's also there's, then there's the open API specification generation right
yeah it's it's in my mind it's three things, it's conversion of Lucid Chart diagrams to web sequence diagrams, It's. It's. Production of the JSON schemas and its production of the open API specification rows are the three things that will take big, big meaty chunks of time.
Okay. Cool. So, you know, I think we just need to get some more horsepower there.
So Max I let me take this action item because they're almost two months ago we we requisition for a tech writer to support this exactly this is this thing. If they have done something from procurement side, There's somebody in the pipeline, they can straightaway accelerate that. Otherwise, then we'll find some resource to do
that. Well I know I know somebody who could do it. It's just a question of like figuring out the logistics of getting them paid. So,
so there is somebody ready to jump in. When we just linked up. Yeah, see, I know in general they're not for the pipeline.
Okay, well, yeah, I mean because we have a similar need coming up, we're like we need a, you know kind of UX rapid prototyping person too. So I'll just I'll just talk to honey, and just say like we need, you know we need these resources now and how can we get them basically.
Okay. There was a decorator, it was called about recursion.
Yeah, so, yeah. Okay, great. So, um, but you know I guess I'm just I'm just standing out and saying like maybe we should not put these like super like ambitious dates because it's probably not gonna happen in two weeks, based on my knowledge. So let's just be realistic about that
which one
they open API specs right for you guys, like I don't think I can commit to that and tailor I'm not sure you can I mean tell me if I'm wrong,
mark here in this chat now tell me, what do I do for this.
Well, I mean I think we just say like we need to hire we need a tech, we need technical support.
But but that's fine. We're opening up late middle school to count those lists down somewhere. Yeah, but we're adding this checklist, blah, blah.
So the adding the checklist, I mean, that's fine, you know whenever you guys decided
this is simple, like really simple
this checklist.
Let me put like next week, two weeks I think. And consolidation to web sequence I think this is where you're saying, this is technical.
So I'll reach out to honey, after this call and, you know,
I think our models here in this API definition.
And I would say a month pending we're pending that we're able to hire somebody, right. So, just to make so that when honey and whoever's looking at this, and they can be like, Oh, this is actually pretty urgent. Right.
It's not tech writer, it's not a programmer,
it's an engineer, like but I know we need somebody who's like a technical, you know somebody who's, you know, yeah, an API does specialist right
technician or
engineer. Yeah.
Or do they do, the brakes on program. It's,
it's similar to writing a program.
Yeah I think so I think you really need an engineer.
Also,
I mean it's like it's sort of like we need it, somebody who's really good at, you know API design as well. You know, I mean it's, yeah, so, engineer, engineers, and engineers. Yeah, so we need a fairly experienced engineer, I would say like an engineer slash architect slash API designer. So,
same for an extra.
We might be able to, I mean I'll do what I can but I think just I think Taylor's idea of just getting some additional horsepower is gonna be really helpful.
Yeah, cuz I, I don't know, my sense is that the resource that is in shortest supply right now for gov stack is time, rather than rather than money. And I, you know, I obviously don't have full visibility there. But I would, yeah I would think that's one of the dials that we can turn and the time dial is harder. Yeah.
So we'll start called sets, it's open,
It's open. Returns. Oh, this also, this is also part of the same API definition right.
So, partially Yes. Okay, in that case, the same rate.
All listings are now becoming dependent relevant standards. See Taylor's comment, Okay, what is Taylor's comment.
Where's that
it's done actually
relevant standards referenced. Oh I think Hold on I can find something on this.
And I have a comment from July 28. The following standards are applicable to data structures in the BB. We have one it says description of services should follow open API 3.1. But I think. No, we make reference so many more standards here. But this is, this is in a comment that we just haven't gotten to yet on this round of review so I'd say in progress and see Taylor's comment is perfect because we'll, we'll get to it when we get to that part of the document.
Yeah, what's our, are we, can we, right now, should we, should we try to commit to getting through this round of reviews by next Friday. Maybe we need to meet twice next week for an hour each or something.
Well, we can try.
Reverse like, what,
what would that be October 7 is October 4 14, maybe October 7, the next next Friday.
I think it wants why why first round Cool, okay.
Okay so the next one is done. We finished resource mark.
No, it's in progress, it needs to be written down as schema.
So the activity is
designed. And again you will land up here, because you need a drying.
It would be very nice to have engineer here but
it just was hard for us to start picking for this guy.
Oh,
he's we know, I mean we can sit and discuss what resources are required just develop a model and put it in. Open API. Whenever, question three. Once we have engineer.
Like, what's the timescale.
So it looks like we're doing everything in the next one week max two weeks. Perfect. Fan assumption for this as well. One month.
I mean a one month is is is the end of October right and we're. Gosh, I mean we're supposed to have a graph to the workflows back by December, 31
talking about on parallel getting things to a demo level right, so if we're going to create any anything. So can I get an input like 332 weeks. For example, take your tablet, now they
can, or you can do it this. Okay,
It's time in the past.
Accent Jack smart,
something weird happened with that last one hold on, go up to the cell above to the 17 Yeah,
I can I can clean, I can clean that up and as I get confused. No stress, let's go on to the next one I can clean that one up. Yeah.
Next will be same, I think it says one activity. You say, same activity. Yeah. Okay.
So there is some comments coming you folks can just check that later. See Taylor's comment logged out Alexander. Okay there's notes referenced, we're coming back to this is a repeat right, this is a repeat. No no no this is standards for data structures.
I think it would be the same as, as a standards from definitions.
Do we line.
Yes.
No, No,
not this one. This one is.
So this is. Okay,
okay, is done.
We have comprehend comprehensively logged key decisions.
Yes we have.
Yeah, I've been, I've been staying on top of that.
And then we have future disregard considerations also locked
with some of them we brought a workflow group others we've raised them with the cross team block.
One is activity to win it down.
I mean, it says done but it's one of these things that like, we need to continue to do it, you know, I, I'm happy to be
able to ongoing activity.
Yeah.
No, I'm
staying on top of that
activity.
Oh well, it's, it's logging them as changes are made, and it's bringing up the future considerations not to go bring them up to the workflow group and with the crossing Limbach.
Now that we're back there.
What is extreme weather, cross, cross, then this is not done, should be.
We can have done it before. This other documents are available
for doing it this is not ongoing right these key decisions stuff is done.
Well we're
going when we, when we get our next time and time. After that, we still will maybe make some decisions
are open all the way to something new.
lectern links to populate other building blocks. Architecture rabbika other documents. Is there any cross references given to staff in other building blocks are they all up to date on layer. Yeah, for example, cross cutting requirements you might
lay on all ID on on line five
reason why you re canceling. I guess we have done that right so far.
Yeah, I mean it's, Again, all of this stuff is like, you could mark it as done, or because we're doing another draft or because we're doing another round of revision you could mark it as in progress. And, but, but most of these things sort of maybe continue to be updated as changes are made. It's a little bit of a. Add it feels like. It sort of doesn't track state over time. But it is, you know,
so let's let's kind of put it as current activity. So we have missed out some internal
links populated to Lafferty. Done.
submitted
to review done. What I don't know what what about this discussion
question, discussions are submitted to architecture for review and review points, if there are any and I think we are going through that I put all of my comments. I think Matt's going through them. This is in progress,
okay in progress.
Let's say we nail it down maybe next two weeks. Yes.
Ramkumar October.
Okay. So we have completed in progress. So when you say PDF version. Max will mean this point, or you mean in Academy some forever. Any,
any PDF version, like it can be an early version a draft version just any version. Okay, I think it's important thing like especially if it's early in rough I mean we want to. Otherwise other building blocks can't refer to information mediator.
Right. So let's take it like once we finish the review can we publish it. Yep.
The question is, when's the review going to be done.
It's 15 here. Yeah. Then, Taylor will press another button, and that limit from the word pilot will generate a PDF file. Yes.
Okay, so first update Table of Contents button and then I'll press that generate PDF buttons.
So let's go to that is done, then the rest of us the personnel in this publication is facing out.
See what else.
Mike Brown is kind of a thing for example decision and unresolved aspects cleared with help of advisory neither tech advisory there are unresolved aspects at that level of spectrum right to premature.
Please take advisory but are we tech advisory what is what is tech advisory,
It's this theoretical new thing I think. I think it's us right now, but the idea was that
we just gonna say it feels like. It feels like we're kind of tech advisory,
well there was, I think so. I mean there's this theory that we'd hire this group of separate, independent, you know like third party who are less biased, who can look at us or stuff with fresh eyes and say like, oh, I believe that or I don't believe that.
Got it. Right.
And there are things that are, that we cannot resolve in our closed group meetings and no two groups, having their own views, and it's getting to a roadblock is created. The guys, somebody has to take a decision.
Well, I think, I think we I think the architecture group makes those decisions like where they're just if there's like a roadblock, we'll make the decision, but the technical review committee is more about having like a neutral unbiased set of eyes. You don't have a dog in the fight like to review these documents in it, and I don't think we have one yet.
Yeah. So that's to be done is not known.
Not buying a ticket.
Okay, so then we have all this to be done. So this first portion of this PDF, can also be put on discourse, right, with links to bla bla bla. Yeah, yeah, next week, same day, as this will be 15.
I still get so worried when I hear people cough. It's so weird.
I got so worried because I came back from Estonia. They did a RT PCR test in the airport before they left me out and all that, too, two or three days. I just get a call from these guys to come back for another test. Oh no. I was just coming back from the test when I got in one of these meetings with Max I told him a test I'm so worried. Yeah, I think, I think CDs and
DVDs okay for now. Yeah. Okay.
So the stuff. I think we all want some 111 or two ones or probably anything new to be added or updated stuff that you all give to honey.
Maybe we can just take a look at it again. Okay, let's put it at 15. If you have no other. Unless you're pulling in.
Oh, what's up Jeremy integration. Suddenly we have come to the demo stop now. It feels like the mall. So, Max, we postpone this part, until we have some clarity on the Mall roadmap might wait one month out there, we should have. Wireframes
near me you know I mean maybe you know that that that you know I'm hoping we can we can make that happen. But,
yeah,
yeah I don't know.
I mean, let's say, three months we should have something demonstratable.
Yeah I mean ideally that would be great, you know, it's just I'm not sure. You know how that how that's going to work with all of the procurements and all of that stuff right
from, from this group perspective can we simply say guidance will be internally ready by 31st. That's something I can put out we don't put anything.
Yeah, I would just leave it blank because it's pretty easy for just
saying, yeah,
yeah.
Hey, publication published for RSP inputs. Okay.
Are we ready. Yeah.
Let's see if it was worthwhile going through this.
I think I honestly, yeah. It's obviously it feels like a bit of a slog but it's like, it's. If this is what we actually have to produce, then it's good to check touch base on it every once in a while.
Yeah.
Okay.
Good. Okay.
Hey on our side just Ramkumar. I think we should definitely try to meet twice next week. Do you guys want to, I need to drop off this call in the next five minutes is the top of the hour, basically, it should be like up over email or do you guys want to pick a couple times next week, or max is there anything else you want to talk about while you have all, all three of us on the call.
No, I think, I think we're good. So, yeah. Yeah, I think we're, I actually have to go in about five minutes as well, so. Okay.
Okay, so now we can do today's next week only taking my next week. Next week.
Ah, that's this demo meetings are happening now, on a weekly basis, is there any day fixed pearls.
Uh, well,
we decided to try to make it mean I thought we decided to make it. The on Wednesday, like an hour later, right. So, I need to go in and create some creative meeting for that. So thanks for reminding me.
Also that will be because that'll be fun all the time that is. I'm kind of seeing that workflow meetings are continuously clashing with.
Uh huh. Yeah,
I think it clashed. I mean this this week is
well that's okay, I mean Taylor probably doesn't need another set of meetings to go to anyway, so kind of gives him cover. Yeah. Right, so that's probably okay.
Yeah, I mean as long as I can, sort of stay abreast of the outputs and then if there's anything that you guys particularly want my input on you feel comfortable reaching out, and I'm I'm fine.
So then fine, that's a pillar can Alexander can we do and day by day,
Monday, it's okay but what time
4pm India time whatever that means for you
or for the planet
okay for me
buying those sneakers slots in between 4pm 4pm Monday 4pm Friday.
Friday me.
Yes, at the same time Monday and Friday. Yes, it's. So Taylor, I can
do it. Yeah, sorry I'm looking at so so what what time is this Monday and Friday
4pm India time.
Okay. 4pm, India time is, is 11:30am London time 11:30am London time Monday and Friday. Yeah, I can do it. Shall I turn around the invites right now 11:30am infrared BB. Send, and Friday. And, baby. Okay, Brad, sent.
Thank you guys. Okay,
so let's meet up and see,
okay, we can do is
find them quite a bit. I'm gonna sign off. So Max are we jumping in at eight o'clock. I mean, one hour from our half an hour from now. Yeah, let's reconvene.
Okay, I hope you guys get to stretch your legs and rest for a moment before your next meeting.
Thanks. Yeah, you too.
Okay. All right, thanks for getting a long one but a good one today. Yeah, thank