 on this computer and I think as a first part, Andre, go ahead. One second, just waiting for this load up. Right, so we had our front end UX called this week and a designer from another product area raised this issue and I just wanted to make sure if this is an issue for plan. And his point was, as soon as the discussion starts on issues before the cycle, UX designers usually get assigned and they get to start discussing that and there is kind of like a space in time where there are implicit assignments for support engineers or product. So the product manager is not usually assigned to the issue. The front end engineering manager and the back end engineering manager are not usually assigned to the issue but they are supposed to be there for answering any questions that there might be, right? And he was saying that, this was Dimitri, by the way, he was saying that sometimes he feels like that lack of explicit assignment that leaves a little bit of confusion of who to ask about front end issues or back end issues or something like that. And he was wondering if we could find a way around and actually start assigning engineers from the get go. And what I told him is that, I mean, there are certain areas in the product that I mean, if it's portfolio management, I can assign to push all right away so that he can like answer, but other areas are a little bit more loose and unless we get too close to the beginning of the cycle, do we get visibility on who's gonna get the issue? But until then, we the managers are supposed to be there for answering. So those are what we call, what I call implicit assignments, which will be the front, the engineering managers and the product manager are implicit assignments. But my question to everyone, especially to the designers on the call is, do you feel that this is an issue in plan and should we change our behavior? Because up until now we only assign the engineers when they're doing, when we're doing the assignments in the beginning of the cycle, but do you think we should revise this? Do you feel the need to clarify who's the engineering manager answering those questions? So Shane, Peter's not talking because I'm still kind of new to UX so I don't really have an opinion on this. I haven't noticed anything, but... Right, so my feelings that from plan everything is being going smoothly because I've been being on issues and Pedro knows and you know that I'm usually answering those questions. If it's worth whatever, if we don't change anything, at least we have this moment to clarify that any engineering questions should go to the engineering managers before there is an assignment. Everyone agrees with that? Yeah, I think I wrote to my opinion on the doc but I'm not sure if it's synced because I was on the train. Okay, cool. Yeah, so I don't think we need to change anything. I think it's working pretty well. Yeah. So I think it's a moment for us to share the way that we work in plan and if it's working maybe other teams can do the same. So yeah, but thanks for raising it anyway. You're welcome. Pedro, you have the next call, the next point. Yeah, I think it was about some possible questions regarding the design of the filter of issue boards. So right now you cannot filter issue boards, you just have the long dropdown with all of the current boards in the group or project. And yeah, this was motivated by Victor maybe having some questions or I don't know if I think it's also a good time for, in this case, I think it's only you, Andre in front and here. I don't think we have anyone from back end, but whoops, sorry, are you hearing me okay? Yeah, I heard you clearly. So I've already, I had a look, I haven't actually put a weight on it, Victor already pinged me, but I'm gonna have a bit of a backlog on my to-dos trying to clear everything for the batch comments. But we'll be starting looking at particular weights starting next Monday. So we'll be putting a weight, but I can tell you right away that I don't see any issue right there because it's basically just filtering what we have on the list. Eventually we might need to do some rewriting. I don't expect to have like huge refactoring to do, but in terms of UI, it makes a lot of sense and in terms of iteration and I'll be able to weight more precisely next Monday, if that's okay with you. Yeah, for me it's fine, Victor. Yeah, no, that's fine. I think there was a discussion for Pedro and Enable. There was a discussion on what we're trying to achieve here. And I'll give you my frustration with using the board navigation, which is when I go to a board, I wanna look at plan things. And I know that there's a plan board somewhere in that dropdown. And since it's alphabetical, I can scroll pretty well and I know my alphabet pretty well. And I can eventually get to the plan area and then I see plan back in and plan back in next release. And so there's like six there. And so from there, I can immediately go, okay, no, I'm looking for the plan for milestone because I created that previously. I know that's there. Probably somebody didn't delete it. So I'm gonna go and click on that. So that's what I want to do. Or that's what I'm frustrated at when I'm going to gilab.org chevron issue boards. And so that's my particular problem and how I solve it currently. And I think that's a problem that other people share. But maybe since we have other people on the call, do you share the same problem or do you have a different problem when navigating or just in general, when trying to access a board? I share the same problem. I would only add that eventually the star boards will solve it more permanently though. Because I have like four or five boards that I go to frequently. If I have those starred, I would never use the filter anymore. That was only the only thing I wanted to add. Right. And so originally we had a discussion with two features. I didn't want to bias the conversation. So I wanted to state the problem first, but the two features were pinning the board, starring the board, whatever, and search as Pedro has it here. And then Pedro provided a third feature, which is recent boards. And so for the problem that I described exactly what Andre said, like I would use the pin board because I wouldn't want to every time go there and type plan, the words plan, and then wait for the filter and then click on it. I will want to see the three pin boards that I use often and then I just click on it. I think recent boards is a little slicker in that you don't even have to do the pinning. It's less control obviously, because if you every now and then look at a different board then you may lose it from your recent history. So I'm more partial right now to the viewing the recent one, but I think that searching should be lower priority. And if it's really easy to do, which I hope it is because it's a front end thing, I think it's worth it to do, but I think the higher priority to address at least the problem I have would be to do either recent or pinning. So I'll stop there and let Pedro respond or others respond to that. Yeah, I thank you for not trying to influence Annabelle and Andre on this, but yeah, because we had a long discussion, healthy but long discussion on the priority of these two problems that are two different problems, but related problems. So the two long didn't read is that the filter helps people navigate when they have a lot of boards and they want to access a board that they know exists and the pinning or recent, so the recent takes away a bit of the control, but still has what you could say a list of maybe the most accessed boards, because most people only access, I think a handful of boards that are not managing like 10 boards at the same time. So maybe the recent, as Victor said, is a sleeker way of doing the pinning without any user interaction. And so that's the recent or pinning or favoriting or whatever you want is the problem of accessing or frequent or accessing boards quickly that you frequently want to see, right? So that's two different problems, like for someone that just wants to quickly go to a board that they know exists, the filter is good. And I think it's a good baseline that covers lots of scenarios. The pinning is more specific or pinning or recent is more specific but it solves a related problem. I think another thing that we could do specific now to the filtering that is an issue that I raised regarding the recent is that if you have an issue, a board's list that only has like two, three, four, five boards, the reasons might not make that much sense. And I'm not sure if the search, the filtering also makes sense to have there when you just have five boards. So just a handful of them would, what do you think? Do you think the search makes sense to like occupy that space with the filter box when you just have five boards? It's not a lot of space. I mean, if you have five boards and the vertical dropdown, there's space there, I assume, right? However long it needs to be. So right now it has new board and delete board which we're trying to fix. And then there's maybe five, seven or how many just like the default max number. And so if you add additional search bar, I don't think it's taking a lot of space and it won't be hidden for most cases. So it's a nice optimization but I don't think it's super necessary and it's not like horribly confusing if you add that. So that's what I think about adding the filter bar. And then I think I still disagree with you, Pedro, on frequent and on search. We can dig into it more but let me say that a couple of things. So frequently used boards is what I want, right? So like I said in this scenario, I wanna maybe I'm not thinking of the scenario you are but I want to go to the plan board and that plan board is the board I care about. So it is frequently used. So that is the same, I don't know if that's the same problem but that is if I go to a board I want, I know about it and it's also frequently used. So the recent thing is to me that solves both problems almost. And then the thing I keep saying about the search bar, which I don't think solves the problem as well is because you have to type, right? You have to type four letters for plan if you're like on whatever interface you have to move your hands or whatever you have to change context. So that's why I keep saying that I don't think that's as good. Yeah, no. Yeah, I'll leave one more thing and then I'll let you respond Pedro, which is could we just leave the search there? I don't know if that's good UI but it's like that you type the word plan and hit enter and then that saves it in the search forever. I don't know if that's like just probably a rule that says that's a bad idea because then you never know the full list or something but I'll stop there. Okay, so I think, I don't know, I think having the search separate or something that resets is I think it's a more consistent behavior but the search, so responding to your first point I think it's, I agree with you, it's a problem I also have that I want to access like the plan board or boards that I frequently want to see. So I think that's a very specific thing and so it's different and that's what I'm trying to say is that it's different from searching. Searching does not solve directly the problem of frequently accessing a board. You can bookmark the board, of course you have to click and then type the board name so it's also always a pain if you want to frequently do that. So it doesn't directly solve that, it slightly alleviates the pain of you having to scroll the long list but it does not directly address that problem. The pinning or recents, that's a much better step forward to solving that problem and it's a problem that I also have. So, but at the same time as I said, I don't want us to be very biased by our own specific problem because this problem of accessing a board frequently it's something, it's a good problem to have because that means people are using boards a lot but it's, I'm assuming that's not a lot of people in the GitLab universe use boards. So searching boards can be good for those people as well and then recents is something that helps them when they are already frequently using them but I think the way you scheduled it and I don't want to need to speak on milestones but I think the way you scheduled it is like first we do recents and then we do the filter and I'm fine with that. I don't want to continue this conversation longer. As long as we do it like in two milestones, I'm happy. Right, right. So yeah, I agree, let's not get, let's not debate further but I think we agree that we don't want to do pin for now and we'll do recents and we'll do search, we'll do both as you've laid it out. Which ones, isn't the search one actually simpler to do, Andre, because there's no back end persistence? Yes, it is simpler. But recent, but the front end for recent would probably be simpler at the same time would it not be? Because there's- In terms of front end effort, let me just think. Let me just think. Because we're front end constrained at the same time, so. That's a good point. Yeah, yeah, so if- Yeah, if it's the recent, yeah, it's just displaying. If it's, that's easier than filtering. Filtering, I just checked, we have that in the view component so it would also be very straightforward but still has a bit more logic and we have to debate the UI as well but we're talking about between a two and a one or between a three and a two. It's very, very low. Okay, then let's just leave it or swap. I don't care. Okay, Jordan. Yeah, at this point in time, I don't care anymore. Okay, I'm just happy that we reached the point of agreement that we agree that both solutions are good. Yeah, that's what I wanted. So one question for you, Andre. Is the filtering or do you think the filtering will, do we already load the whole list of boards and if we filter, are we still having to go to the back ends to do the filtering or as you said, it's only purely front end? Only on the front end. So what we'll do, we get the list of every word already because that's how we're rendering now. Yeah. It's a few components. So basically the text box will just store the filter and as we're rendering, we just render the ones that match the filter. So it will be fast. Yeah. Okay. Oh yeah. I understand your point. Yeah, because the instance searching like the label. Instance, yeah. It will be... Yeah, it's not like the labels that you have to go and search for the labels in the back. How such a computer is not as you look into the internet. Yeah, there's also something that I didn't think about but now I'm thinking of it is when you're searching and you have the results, is it possible to shuffle through the results in the lists with the keyboard and press enter or no? We can discuss the interaction on the issue, I guess, or on the kickoff. Yeah, I think we can just skip to this point but we can do whatever we want as long as it's reasonable. Yeah, yeah, yeah. No, I was just thinking if it was something that we had for free, so to say, or not. I can have a look and I can chat offline off the call. Okay, and lastly, Victor, I think the other thing we should do as soon as possible is persisting the last view of the issue boards. So, and that will also solve that problem even more specifically. So, the last time you loaded up the issue board for gitlab.org was on plan 11.4. When you go there again, it will be 11.4. Yeah, that's an obvious one, yeah. That's like, that is the recent, hmm. That is the recent at the same time, yeah. Then let's just do that and not do recent. Like, we'll leave it, but like, for now, I don't think we even need to. Good point. What do you think? That's even easier. That's like zero front-end work, right? That's just literally. I mean, yeah, I think we should do both, but if it's easier to do the loading up and persisting the states, let's do that first then. I don't know. I don't care if which one gets done first. And then I'll let you obviously think about it to make sure like there's no gotchas there, but let's do that. Okay, thank you. Yeah, like if everything goes well and then we don't think of any weird edge cases, because that is a pretty big change, right? But I can't think of immediately and then like the last three seconds why it would be a bad change, but it is a big change, expectation-wise, right? If you've been using it for so long. And so- Then you get a fault. Yeah, it saves where you were last time. And that's like almost nowhere in GitLab that happens, right, if you think about it. But that view is also pretty unique or that UI is pretty unique in that, I think what like branches has a similar UI where you have a dropdown and you have navigate instead of a list view and there's paginated, right? Like that's sort of like the two paradigms. And so with this dropdown paradigm, I think only boards and branches because I thought about this and asked people are the only two places that have this. So I think it's a good change overall, but it is, let's at least use like the next week to think about it, but I'll put that in 11.5 and then kick out the reasons. Okay, cool, thank you. Awesome. I'll just highlight that Constance posted on the chat saying they can have automatic filtering for free. So that's good. Thank you, Constance, instead of reasons. Next point is Annabelle, I think. Yep. Yeah, okay, cool. So I wanted to let people know or show people the initial thoughts or designs for sub-epics, which is kind of now renamed epic relationships because sub-epics doesn't actually make that much sense. Let me share my screen. Can everyone see my screen? Yes. Cool, so initially we were thinking of just kind of copying what we already have for sub-groups in that you've got to care it and you just expand it and you just keep expanding and you'll keep seeing new sub-issues, sub-epics. Or actually, this is focused on sub-epics, so forget sub-issues. The initial design, I was thinking we were trying to start with like a really, a pyramid basically. So you've got your main overhead epic and then that can be split into two more epics or however many and then each one of those can have unlimited epics within. So you've got kind of, if you pick an epic right in the middle of that pyramid, it can go up and down. So the parents can only ever be one parent at a time. You can't have multiple parents in this design. So that could go on the sidebar, which is right here. What I don't like about this is the title is, oops, the title is duplicated, which we can fix later. But in this example, the immediate parent is this one and then the immediate parent of that is that and that's where it terminates basically. And then to go the opposite way, you're gonna see the epics here where you would normally see the current issues in an epic. Here is the really unattractive design in Balsamic and then we got some feedback from Yov that it took up a little bit too much vertical space. So going from the new design of the related issues, they don't take up as much space as the current version, which is good, I guess, but I was thinking that we could sort of start grouping them by this gray background, which will just switch between the two colors depending on where you are in the tree. The default view will always be collapsed just like a discussion or subgroups. And then you expand one at a time and you'll eventually see the entire tree going down. So these are only the descendants again. My question actually for this is, because this isn't gonna work really if it does act like subgroups. So for example, I don't know how these are implemented, but if you click on a group to see the descendant groups, it takes forever. And sometimes it doesn't load. In this case, it's not even loading. This one I think loads, I tried it earlier. Right, so that's because it only has one subgroup. So Andre, do you know if this is a front end problem or back end or what? Because if that's gonna happen, we can't do this. Now, so one from this perspective that I don't know this code base, so I would have to look into it, but it looks that something's happening whether that request failed, the Mark Cesare one, but the front end seems to be working, but something's wrong for sure. Regardless of what it is, I wouldn't see this as a blocker for what we wanna do there. A cushion block design, we should be able to do better. Absolutely, absolutely. So my message to you is, don't see this as a blocker, continue, because it makes sense. We just have, because you're not, if we're talking about real time updates, we have to check how quickly we could do that, which we're gonna talk about, but this is just listing relationships. We should be able to do this pretty fast. We either, and we can either load this right away, because it's not that much of the data. SNS does a lot of complex queries, but I don't expect, because in the database, you'd have a direct relationship, right? You don't have to go, so one of the issues with the board is that you're gonna have to search issues that have a certain level. So that query is heavier than just like direct relationships. So I don't expect it to be there any bottleneck on the back end, unless they have 1,000 groups and epics and subepics, but that's never usually a use case. Right, and we were thinking, well, I had this idea that we should be restricting them and sort of following a more strict level of, you know, we could only go four deep or something, but for now, I think we're just going unlimited and moving up to the user to use their best judgment and not go crazy. The other question I had was, was dragging a subepic from one level into another. The gray background was supposed to denote, like, here's this grouping, and then you can reorder within this gray background, but I think it makes sense to be able to drag one epic from this level into another level if you messed up or something. So I like the idea that within this, you could just drag this and then it would pop up to the next level as well. That could be in a future iteration. I don't know what the immediate ask is, but that was just another idea. Yeah, I'll just comment on that. That's like an amazing design that I think, like a lot of customers would find super, super compelling. And you know, you're totally right to design it now, but also to comment that like, it's no way we can build that. And maybe even like the first four or five iterations, but just having that aware of that design upfront is precisely what we need in my work exercise. So I appreciate that. Okay, cool. And then we should definitely, I don't know, just in that point, we should definitely double check that they want to move the epic, like have a dialogue or more. Yeah, definitely. You don't want to, I was thinking you could just rearrange and then maybe there'd be like a save button when you're finished. I don't know. I'm not a project manager, so I don't know how much you really rearrange. Victor, you probably know more. Yeah, I would disagree, Andre, with having a dialogue, because if you move, then you can just move, if it's easy, then you can just move back. But yeah, no, I definitely defer to further conversation. That's not a strict, like that's a terrible idea. I don't think it's reasonable. Like right now when I move stuff around in boards and related issues, I created a couple of issues to fix some bugs as well. So I don't know if it's because the bugs are letting me be uncomfortable with that interaction. I find that interaction very intuitive. I like that design, but the one feedback I have with that is that it's almost too easy. So when you have the dialogue, Andre, it makes sense to me, that resonates with me because then it's giving you confirmation that you changed something. So I wouldn't necessarily make it at friction, but maybe more feedback. Like I thought in my brain on a weekend, but I didn't create the issue. Like maybe you have like a little hover thing that pops up and says like change a save or something like that. So I don't know if that's important or not, but that's definitely the future. And then an undo link. Oh yeah, like, well no, that's the thing with undo. Like these types of things, like why would you need undo if you can just as easily drag it back? Well, because one thing I forgot to put in here were the X buttons. So if you combine those two things, if you start deleting things and then you realize- Yep, I agree with that. So it could just be combined like Xing or removing them and shifting them around could be one sort of, if you made some changes, is this what you want to do? Or I don't know. That's true, that's true. So, go ahead Payton. So my question is not thinking specifically about issue relationships and how this relates to epic relationships. What kind of issue relationships would be supporting? Just child parents or anything else? I think in this view, only child parents could be used. Like if we show blocking, well, the issue relationships that I know of that at least that are in my brain are blocking child, parent and related issues. Right. Which you have now. And so like I can't think of a way to express blocking and related here. Yeah, so what I was thinking is do we need to show, so like issues will always be at the same level. So for example, if I'm putting two issues and saying all these two issues are related or this issue is blocking that other issue or they are still at the system level, they are at the same level, right? But if I say that an issue is actually a parent of another issue or if an epic is a parent of another epic, they are still at the same level, but at the system level, but conceptually they are now child parents. So one is beneath the other. So what I'm thinking is wouldn't be easier for us to start with just one level of child's parents and not think about further children inside of the parent or further levels of the tree. And my second question is, could we have a way for you to quickly switch? So if I add a related epic or a related issue, can I quickly switch the type of relationship? And then I have a drop down there that you change. Oh, this one is actually not only not a related issue. This is, well, this is a related issue, but it's a parent or this is a related issue, but it's a child or you know what I mean? Yeah, no, that's an epic. That's a pretty interesting one. And I think a little bit of scope of here, but pretty super relevant because to me that's, yeah, no, that's like you have at least three relationships and they're like essentially arrows in my brain, directed arrows, but most people probably don't think about them like that. So how do you present that on it? I would assume you would do that manipulation on an issue screen itself and not on this screen. Yeah, exactly, yeah. I think that's related here and something good to think about and maybe even design out as part of this exercise, but at least for the purpose of this, whatever Annabelle has designed already, I think like if once we have parent child issues, the natural design that Annabelle had is that you just keep going deeper. And then if you're looking at any parent, whether that is an epic object or an issue object, you get the same UI, which is you just get a tree rooted at your certain object and computer science trees grow upside down. As my computer science teacher taught me many years ago, which I never thought about, but it's true. Also, family trees grow upside down, but then you just see them going downward and then that's the consistent design. I think that's great. But yeah, whatever you think related and then push them in. Yeah, I was definitely thinking vertical. I didn't actually consider going horizontal, like in this diagram. If you're looking at like on the far left, you got flexible work breakdown structure with sub-epics. I was thinking just going up and down because polished sub-epics is it's sibling, but it's not, you can always click to the parent to see the children, all of the children was what I was thinking. Eventually, maybe we could do some sort of like see the whole tree and it would show this explosion, all of them, but that would be not fun to build. But it would be really useful probably. Well, but we do have this, right? Like if you wanted to see what you show on the screen, you just have to keep going to your parent nodes. But you can't see siblings, right? You can't see siblings, you're right. Of yourself, but if you want to see your sibling of yourself, you just go to your parent. You go to your parent, yeah. Which you have, you didn't talk about yet, but it's like the sidebar thing that you design, which is awesome, you would be able to do that. So maybe there's some extension in design to indicate that you have a sibling, but I would think that's overkill. Like I would make the user like, if you want to know if you have a sibling or not, just click on your parent and then you see it from there. Because that seems crazy to design that complexity into it. I must say that from the front end perspective, vertical, that you were saying Annabel really works great for adapting for smaller screens. So that's good. Yeah, I'm just thinking if we have to, if we have to cross, if it's now it's time for us to cross this bridge of showing even the issues inside of a sub-epic. So basically have such a comprehensive view of all of the children and all of the parents and all of the issues inside of the parents and children. So can't we just make it a bit dumb in the beginning and you go, if you want to see all of the parents, you go up the parent tree, click in the parent and then another parent, do we necessarily have to and even the children, do you, I'm thinking about the use cases of you being in a child epic and then having to go like two or three levels really fast to the parent epic, like to the top most and the same thing, right, down. So Pedro, when I look at what Annabel is showing on the screen there, to me that's like, that's more than enough, that's gonna be like 95% of the use cases and then there's gonna be like strange 5% of users who are gonna be asking, maybe more actually, users asking really ridiculous things like show my siblings and then we would like not do it unless it's like a really compelling reason. So are you saying that what you see on the screen is too much, is that what you're saying? Or are you- Yeah, I'm asking if we need this right now. I think we need it eventually. I think like this is like, this is super compelling. It's a really good use case to be able to see the entire tree wherever you are, assuming you're the root. Definitely we can't build it all in one shot, but I think it should be a clear goal and something that we build toward is my answer. Especially in implementing this, it would be very useful for us to have that visibility so we can eventually plan to load each level synchronously which will have a sort of a loading thing but might be worth thinking about right from the start and personally I can tell you that when I saw it, when Anibal was presenting it earlier today, I felt really great and useful to see the issues down to the issues. Yeah, to the leaf nodes essentially. Yeah, so I mean there's a bazillion use cases, Pedro, but like product people want epics and subepics so that you can track stuff like initiatives like the original or not the original, the graphic that Anibal had on the other page showing the taxonomy or whatever you call it and then engineer like Andre and Sean want sub-issues for, well, I don't think they said they want but I think it solves well the one product issue and then it's subbed by two implementation issues front end and back end. So that's yet another use case and then like a product manager, a designer can see that from the epic level, from the epic they see five product issues and then immediately you click or you see right away those are further leaf node by front end issues and back end issues. So I think just myself, there's like at least two or three use cases there and then customers have been asking for these crazy roll-ups as a thing, spanning like four or five levels, essentially reflecting their org structure as well if you think about it. Yeah, so all of this looks really good to me. Yeah, so if there's no more comments, I think we can definitely start building this, I mean, depending on capacity, but if there's anything, if there's nothing else from this design, Enable that you had further thoughts on, to me the next area there that would be relevant would be the roadmap view which is gonna be really crazy. Is that in the epic or did you even think about it at all? It's probably not in that epic, right? I don't know if it's in the epic, but now I haven't thought about it. Yeah, so it's totally fair that you didn't think about it at all, but do you agree that that's something to, oh no, it is in the last issue there in the epic. So do you agree that it's something to look at next or did you have plans to, you wanted to detail more additional things on that tree view or stuff like that? I'm gonna have to look at it. I'll get back to you on that. Okay, okay. But if anyone has any more comments or anything, I linked to the issue in the doc, so please comment. Okay, thanks. Yeah, as far as I'm concerned, the design is amazing and we can build the very first thing. And the very first thing would be what Pedro might have been alluding to earlier, which is the way I see building the very first thing is that essentially the design would not change with the existing thing. Right now you have inside an epic, you have like a window with issues and you have five of them. You'll just be able to add epics there, right? So to me, that's the first iteration. You just see your immediate children and that's it. And then that's, but right when we do that, when we do that, we will say like in our blog posts, we invented like sub-epics or whatever we call it, which like product marketing cringe at us. Oh, can we talk about this later? We never do, we talk about what we have right away. We'll figure that out later, but my point being is like the very like the quote unquote, well, like the first hard part I would see is the backend making that possible. And then the rest of the hard part is the crazy front end logic to make more, like seeing more stuff on one page. But like backend, I presume you do it once and then you just keep reusing that relationship. Yeah, so the first part will be backend heavy, but then the rest of like this five, six iterations will all be hard core front end. Yeah, but I agree with that. The first step would be important to be like the first stone of a very important revolution. Yeah. So which is why we need the entire design up front, which Perfect. Yeah, I agree. All right. Pedro, any other points before we give 13 minutes to see Andre? No, no, I'll comment on the issue. Awesome. Okay, can I take my next point? Yes, please. Right. So charts, I don't know if everyone is aware, most of you are, I think, that the opportunity of reusing Meltono's chart library came up. And we had a call with Jacob and he showed what he had. He passed me the repo that he had already started working on a few charts. And then I went back to my little cave with my front end managers and we took it a little bit apart and we talked a little bit more. And there are several questions here. One is the kind of graphs that we are going to need to build for the GitLab product as a whole are not entirely visible. We are, we have good visibility on the plan side. So thanks, Victor. I passed that along to them already. So that was part of the discussion. The kind of maps that, the kind of graphs that we are foreseeing and having in the plan side in portfolio management as well. But we don't have much visibility on other product areas, specifically monitoring and that kind of thing might be very useful. So our initial strategy is I'm preparing a document that I will be passing along to all product managers of all areas to fill in with forecasts of graphs that they're thinking of getting, not high res, but just like thinking of mock-ups concepts. And that will give us a hint. And then we'll also get it through UX to have more of an interaction information, whether they'll be very interactive, not very interactive, which will then give us more of a notion how to build and which library to use. Right now we are leaning towards D3 just because there's a lot more complexity, sorry, there's a lot more knowledge in the team. But also it allows for more customization from what we've seen than charts. Chart might be easier to use, it might be simpler, but D3 has a little bit more power in customization. So we are leaning towards that end. The other part that we want is we wanna have, and Mac also highlighted this point, we wanna have this from the get go with very high quality standards. So testing will play a very big part, eventually including visual diffing, which all of these points, including documentation and spreading the word and everything is the GitLab UI mission. So there's a very big alignment in that part. So what we're thinking that we could basically get your budget with one stone is that charts could be a category inside GitLab UI. We'll piggyback on all the effort that we've been doing in that side and we'll move everything along with wrapped components that will provide kinds of graphs that they only need to be fed data. And this will be documented, this will be tested visually as well using the GitLab infrastructure. We consider splitting it all up, but for now we're just going all in with the GitLab UI. But all of this is depending on that exploration part that we're gonna trigger. So I have the document already ready with the plan issues. I might have forgotten a few things, so I'm gonna link it after this. I'm gonna link it here and I'm gonna ping you, Victor, just to make sure that I didn't forget any important charts. And after that I'm gonna prepare sort of a call for comments for all product managers. I might put it on a company call and then spread it around on the product channel so that we get like within a week or so, we get feedback from all product areas. We forecast the charts and after that it will be in a position to start making decisions in how we're gonna build this and which line we're gonna use. Now, having said that, I don't think we should stop product development of any charts. That's why when you ping me earlier on Slack, I do think we can get that into 11.5 just because that's a regular chart as we've been doing so far. Once we get the new components charts in GitLab UI, all we have to do is go back to this, to all of the charts and re-update them to use the new components. So I don't think we should stop product development but we are gonna make an effort of pushing this forward in parallel. If we have like the burn up charts which are significantly harder to implement, we might consider those for the GitLab UI charts but for this kind of charts that we already have like the bars charts that we already have, we can just go ahead and start working on them as we usually did until now. And there won't be too much of a wasted effort in setting things up just because everything is set up already. We did the D3. So I think I covered all of the grounds. Any questions, any comments on that? Yeah, so you said burn up charts. You said it was weird. I just wanted to catch those. Wondering why. Yeah, I remember that was, and I'm not sure what you call it in particular but you had an example mock up in Boston or something of a burn down but it was going up. I'm not entirely sure if that is. It's like a stacked thing, right? Yeah, so I don't know. It's essentially a stacked chart but like I just shared that because it was interesting as a feature but it's also interesting. Yeah. With a stacked, it's both stacked and not stacked. So that's why I found that interesting. I wanted to share it and just like if it were, I don't think it should be difficult but I don't know. But basically with a stacked chart and these type of stacked bar chart or, so yeah, it's a stacked, there's bar charts which are individual bars and then there's a lot, so this is a stacked line chart, sorry, not a stacked bar chart. And with stacked line charts, what you have is you have a line and then the diff is the next value so it stacks on top of it and then so on and so forth, right? And so the value of each series, you're adding it as a diff on top. And so that's one way to do quote unquote stacked lines. Another way to visualize it is you have a line which is this and then you have another value but it's not the diff on top of it, it's the actual value. So then you can have two parallel lines but it's not that. And then so the visualization that I was showing you or the feature had, it's combining both at the same time. So like n minus one of them are the diffs and then the last one is the total. And so to me, that's like just an interesting feature. I'm like, oh, that's like a weird edge case where like maybe some libraries where you say, oh, I want this particular chart and here's my series of data. Like one of them is like weird. So that's all I mentioned. So, and that's a very good example of stuff that we need heavy customization on that. I will admit that my experience on D3 is limited but I passed it on to Tim and I showed him that graph and everything's invisible. I just posted the link on the document of today's agenda. So if you want to have a look of the graphs that I've identified, there could be some missing on the plan side. I just did a quick overview. And this is the document I'll prepare to soak around product managers. This is awesome. Does that sound good? Do you mind if I add to this? No, that's fine. I may edit abilities so that you can edit and that's gonna go around with all the product managers so you have a bit of a head start from then. But thank you for having such a, I have to give you credit for having such a long running forecast of stuff because it really helped grasp the needs and you can make a more educated guess on the library. Yeah, please go ahead and ping all the PMs. I will definitely put this on the product meeting agenda we have on Tuesdays as well so that will give them two points of annoying them to do this. I plan to do this at latest tomorrow. We'll be starting the message tomorrow. And then so looping back to your original points, so does this mean we are not using Meltano? That has already been decided. I chatted with Jacob earlier today and the idea is that we're not using Meltano at this point. And just because there's a difference in goals and also I think it might be even better because we won't hinder their evolution. They can evolve in their own way. If there was a lot of similar points, it would be beneficial. So that was definitely a good point of raising that possibility. We just didn't see that in the code as it is now in our own assumptions of the graphs that we're gonna be building. So to be absolutely sure, that's why we wanna do this round of requests for comments. Make sure that we're aligned with the product needs and then we'll be free to build whatever we want. But we didn't feel like that would be a lot of benefit of reusing that and they can do the charts that they want. Eventually they might even, if it gets to that, they might reuse our part or we might even reuse theirs if you get to that. But at this point, I don't see that happening. Okay. Because we want, oh, and also one other thing is that we will be looking to put people assigned to this part that I have experience on D3. That hasn't been completely decided if we're gonna get exclusive people assigned to this. But it will be part of a team to be responsible for this part, but that needs to be decided on how exclusive that will be. But it makes everything easier to just like could gravitate everyone towards this effort of putting GitLab UI as a center of reusable components from GitLab. Perfect, awesome. Okay, yeah, I look forward to it. So you said we, I think there's like one chart in 11.5 you want to take that out right now, right? I think that, I don't want to take that out. I think that we can still carry that out. Okay, so you- The bar charts, we already have that in other parts of the product. It's basically to be light on front end. That's going to be basically just preparing the data and putting the data where we already have. Okay, so you mean like this exercise, you don't want this to drag on you, you think like you're- Oh no, no. Okay. We definitely want to start this as soon as possible. What I don't think we should do is stop development of product unless it's complicated graphs. If it's complicated graphs. So for instance, the value stream management might be worth being one of the first ones on this GitLab UI thing, but this simple graph charts that we already have, I don't think it's worth stopping because later when we add to the GitLab UI, we just go back and change it and we start using the new one. Perfect. Yeah, from a backend perspective, I definitely think the quality team one where it just sucks in issues and counts them. That's probably one of the easiest. When the VSM one is probably a little bit harder. So I'll leave that one in right now, which is the one that the quality team wants. Yeah. So I'll leave that issue in five and then we'll go. Okay, so two minutes left. Real-time borderless. So I just wanted to quickly say that for the charts, I'm planning to have the, and I think, yeah, Sarah is also on board with this to have Amelia work on the pattern library for charts and everything that we're going to do with charts in GitLab, where they have a lot of the groundwork and that's essentially what probably you have already linked in that document and the other program I was just going to link to. But we have an issue that I'll put in this agenda when I'm at the computer. It's an issue for our pattern library and I don't think it's already assigned to Amelia but essentially what she's going to do is exactly what you're saying you're going to do for the front-end side. She's going to do that for the design side of things. And I mean, they're all, they're married, right? So everything must make sense. You're going to look at if it's going to have interactivity, she's going to look into that as well so that we have the guidelines from the design side of things. So user experience wise, she'll be the go-to person. And I mean, it sounds like you'll be the go-to person for the front-end side of things. So you have to be, yeah. What I'm trying to say is that you have to be aligned on this so that there are, the expectations are. I will sync up with her before sending the message because I want her to be on board with this and not duplicate effort because you might be thinking of being a product manager as well. Yes, yes, yes exactly. So yeah, I told her that not necessarily to ping product managers, but to tackle that issue, we needed a comprehensive view of what we already have, what we're planning to do, exactly what you're saying you're going to do for the front-end side, that knowledge is cross. Okay, perfect. I'll sync up with her. Okay, so time's up. Andre, I will ping you on that Epic. I wanted to get your take on it, but it's not urgent at all. So if you don't mind it. I can go through the Epic and comment there. It's fine. Yeah, yeah, okay. All right, thanks everyone. Okay, bye.