GitLab Dev:Manage/Plan/Ecosystem UX Design Reviews, 26 Oct 2020

Previous Meeting

⏯

youtube image

►

From YouTube: Dev:Manage UX Weekly Design Review 2020-10-26

Description

No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).

A

B

C

A

Hey, I appreciate you um dropping some comments in the uh on the agenda beforehand.

A

It's really easy, sometimes to forget. I do all the time and.

B

I just hope to meet him like oh.

A

C

Well, also to help to like um have them documented in case you go back. I always want to go back and look at meeting notes and see what somebody said, because I'd always forget.

B

A

Hello, hey going miller.

B

A

Right, I wonder.

B

Nick's out of office.

A

Yeah he's out office, I don't think mike is joining. I don't know it's earliest time.

B

Oh yeah he's not online.

A

Yeah yeah all right um cool- I was mentioning to daniel earlier I was looks like an emperor, leaving some feedback in the agenda that helps get the notes started. So thank you to you as well amanda for that um I'll I'll share my screen and we can kind of carry through the agenda.

A

I think I still have some outstanding questions. I'd love to get your thoughts on oh gosh, which firefox window do I want to show this one.

A

So one of my just challenges that I run into- and maybe you guys can talk me through how this plays out for your groups is issues will just appear in workflow design, and I think the intention is good. Like hey, I see this ux gap and I think we can make it better.

A

The struggle then becomes for me.

A

Do we need to do this, so I was trying to understand from your perspectives. um I didn't really. You left. You left an idea in terms of working with hayanna and mike on the settings experience.

A

um I did talk to him about it. It wasn't necessarily so much about stuff like this. It was more like the complicated logic behind inheritance across the different levels of get lab, but this, I think this same problem appears elsewhere.

A

Example is we have this amount of inventory for credentials like your ssh, keys and stuff in the admin area?

A

It also exists in your user settings, but they're different experiences in terms of how you find them and interact with them, and I would even argue that the persona of the person dealing in the admin area is different than your general user, which I tried to kind of present back to matt, because, like as an example, I one of my friends was um using gitlab for the first time and like he, he really had like no understanding of like what um a personal access token was.

A

So that's where I kind of run to the some of this. Like crux of, are we actually solving the right problem?

A

um The problem, I think that is addressed in this issue is we have inconsistent ui and I don't know if that's a worthwhile problem to solve all the time over other problems that we could be solving. As a group.

C

Well, I think, maybe for this in terms of the consistent ui, I think I sort of brought up a bit of that. Is that why don't we add all these things to access uh or access section, or something like that, because it's getting these things provide access to some sort of environment the credentials the same thing too, so having it be an access setting would make more sense to me not so much the the group access. I'm just talking about accessing the the concept of getting into a system.

A

Yeah, I'm gonna drop down your notes.

A

Too, your thoughts, amanda.

A

Oh you're, on mute by the way.

B

All right, thank you. um It definitely looks to me that it should be combined um and whatever it's called access credentials, but nonetheless, I think it should be combined in the consistent ui.

C

Okay, was there.

A

Any reason they're not.

C

Or any sort of pushback for you trying to merge them, no.

A

I think if they think I'm the one pushing back.

A

A

So part of the thing is, I I feel that the person that's managing the credential inventory at the admin area is somewhat different than the person that is managing their personal user settings. So I think the person that's going in here and managing their credentials and deploy keys ultimately has a different task than the person that is um going into their user settings and then accessing their different keys and whatnot from here.

A

So to me, it's like why blindly copy one or the other.

A

C

Thinking for the sake of consistency,.

C

If, if you were to go back to the user setting window and then create a screen or something that says credentials or access or whatever, you want to call it, and instead of having ssh and gpg and access tokens all separate, you throw them all in one place.

C

In the other perspective, if you go to the site admin setting.

C

This is the credentials page which is sort of what I had envisioned when I was talking about the personal access tokens or ssh keys.

C

But in this case you could have many or multiple ones, whereas a personal one who don't just have the one theoretically on the user, setting potentially.

A

That's what you're saying.

C

Is that having these two screens be the same, wouldn't make sense.

A

That's what, as I feel, I I'm open for suggestions, but the thing I'm lacking is like: why is this a problem like be? I understand that they're a different experience today, but I don't understand why it's a problem, and I the only problem I can hear articulated is well it's inconsistent ui like well. Not all ui should be consistent.

A

There are times, maybe when deviating is easier and the exam. The very like niche example I brought up is someone that's unfamiliar with personal access tokens. It might be more discoverable for them to be at a level where you can easily find the word access tokens. When we haven't implemented a search capability to help, you find something very specific. They might not understand that it's a credential, they just know they have to find something called a personal access token.

A

So I'm I'm just slightly hesitant because of the fact that, like it would require a lot of change to all these different pages like we basically be tearing down pages, taking away certain aspects um pulling in credentials as they stand here to me is like this seems like a lot of work and we, I don't really know where the problem is stemming. Besides just inconsistent, behavior, so anyways, I think maybe just goes back to that original concept of how consistent should we be?

A

Where do we? Where do we stand as designers and holding or holding true to that or deviating from it?.

C

To me, this feels like the perfect opportunity to just do user testing just to say, okay, I need to stop thinking about this, come up with some ideas and run them by end users and then also grade your um qualifier users to see different buckets like people who know what ssh and pats are people who use them regularly versus somebody who's never heard of it and then offer two different test: scripts around that and test the same environment just to say to get a better perspective or not to keep banging your head against the table right.

C

Because that way you can get actual usable qualitative data.

A

Right because I mean I think the easiest thing would be to say: yeah, let's just make a match, I mean I get that it would probably take more. It would take more effort on, like the research front, to go actually validate this as a problem or put more effort on the product team to go, build it. Okay,.

C

No and that's where I would just do it as a research project and then that's like use, digital comps or mock-ups, and just um go through testing with it.

A

Yeah I mean I would love to do an on unmoderated test. I just don't know how to do it. It's that same tricky problem you run into with access stuff because did.

C

You see a environment for someone.

A

C

Yeah and the problem is if you're gonna do it with yours, um where you want to do it as uh because we don't have this built, then you have to do it as a digital environment. So then you would just use a figma prototype, um but did you look at our dovetail interviews for the cms course?

C

I have not gotten through it now. If you get a chance, go look at the project, I'll find a link here.

C

uh One second.

C

uh So if you look, they were unmoderated and they were a disaster because it was just weird okay, but.

A

C

Very good because they were very helpful in that it explains that nobody knows how to use gitlab or the system, and so that basically makes it um the best case scenario like okay, um then I have free reign to do whatever I need to make it better or to fix it or yeah.

A

I mean user feedback.

C

A

Yeah, like I think I think, I'd want to like try and post is like.

A

Can you create an ssh key, but the the issue might run into someone actually able to do the ssh key and not so much like finding it in the settings, but anyway yeah I'll um I'll loop back to our new ux researcher and see if she has any insights on how we could validate this problem.

C

Well, that too for sure, but I would first start by saying what you want to get out of it like I.

A

C

If this solution proposal makes sense and then say, I want to test this idea and then have users go through the process and validate or invalidate the proposal, because at least you're starting with something, rather than saying, how should I do this or what would be the best solution to start with to go forward?

C

Because the idea being is that if you go- and you say I have this proposal- I want you to solve these tasks, and you have these two different user profiles. The.

A

C

Versus an experience, then you get data and everyone hates it or everyone loves it, and then you can figure out after the fact. If everyone hates it ask for feedback, what would you change? How would you need help to get to the solution or whatnot, because that's what happened for us was that we had some people like the task was to build an organizational structure in gitlab and we would just plop them into the group container and there was a button in the upper right said new project and they didn't even start there.

C

They sometimes did, but they were making projects not subgroups. So that's part of the problem where we had to rethink the whole thing um and try to start with splitting that up. Okay,.

A

Yeah I'll um I'll take that as a launching point thanks for the feedback too I'll keep my next one brief. So I can give amanda some more time.

A

One thing I'm struggling with is this concept of like putting all of the ux work around a specific feature in the app into a giant epic, because I didn't, I think he did a great job starting this, but then it became at some point.

A

It lacked a representation of where we are today and so then it becomes like okay. Now we have all of these phases that are outlined, and maybe these are kind of true. It feels like. Maybe we've deviated a bit and then we have all these issues then, and our sub epics, with a collection of more issues underneath I'm like I don't like. I don't know, if we're going to add security scanning to the compliance dashboard, I've kind of heard that that's not super helpful. In some cases, um that's anyways.

A

I was trying to find a good representation um daniel. I added some links. I had the one.

C

A

Wanted it I just wasn't like: where do we take these things? These things feel like a lot of work to manage, for maybe it's more of a product thing but didn't feel like. It was really helping me all that much.

B

um Import has import love, epic and that actually houses all the ux work, which makes it easy to organize and, like figure out, you know, categorize them.

A

Does that make it too lovable or something or is.

B

It yeah, okay, um yeah and I'll find the import like mega like.

A

Yeah it'd be great if you could share, because I was, I was digging into the one that um daniel shared.

B

Yeah I'll pull it up and link it.

A

C

There's also I'm trying to find.

C

Oh, it was a video I saw on youtube one second.

C

One second: let me go through my youtube. History.

C

Here it is it's not necessarily a dashboard or a epic or planning thing breakdown. It's um nadia made a discussion video with her group for wait.

C

um What they're doing for the three year vision, I think, is what it is, and that might be kind of a good way to see how you want to do yours because it sounds more or less what you're trying to do um and what they're doing is they're just talking about how they want to do the pm or ux discussion around what will be done over the next three years, and I don't.

B

C

It I think, as a as the info on the about.getlab.com for their stage and so using that, as kind of like a holding point for all this, um the idea that you would use an epic for it.

C

I would do it kind of like a high level. Just say these are the concepts or things you want to do as an epic or at some point in time and then break down everything. That's can be granular, broken down into smaller parts into sub epics and then sub issues.

C

And that way, you kind of have this long-term project, but it'll also help you because then you'll be doing the long-term planning you'll be doing the research associated with that and then you'll be doing. The minor products feature work and any bug stuff that gets dropped in like um who was it? Yes, I think it was the issue that nick posted about um changing the way that uh the groups or the group work is done.

C

uh Hyanna had made a good uh responsibility, yeah yeah, so that one was the one that I was kind of thinking about in relation to that and the video also okay,.

A

Sweet amanda. Sorry, I only left you like seven minutes.

B

Oh, no, that's fine, um so I need help with my button migration issue and it seems, like you, guys, have already kind of helped me.

B

Yeah I've been working with tori a little bit, but I'm just having an issue just like pushing um or merging my code.

A

Changes: okay, yeah.

B

So I can continue to work with her um on doing that. It's it's man! It's a.

C

A

Man there were a couple times I totally broke like my commit history and like that really jacked up pushing.

A

I don't know if you've taken this as an opportunity yet, but it's a great chance to meet with some of your front-end engineers and just get their insights and help, um but as a way just to like talk with them, but also they are pretty good about what the basics are just getting to push to master and yeah.

B

Being done with it, yeah okay.

C

Yeah yeah, that's the only way I would do it. I tried looking at the stuff by myself, was like I'm not gonna. Be able to do this without a developer like and.

B

Reading the instructions yeah.

C

Yeah not even that, but it's like saying: okay now type this and here I'll, send you this copy paste this and it's like, because I'm not a developer, and so I don't know anything of what's going on here, um yeah, so yeah, it's like I wouldn't be able to do it without somebody sitting there specifically telling me what to do and.

B

C

Then I had uh tried one with peter uh our front end and we couldn't find where the button existed in the entire environment. It was like we spent a whole hour on that. It's like okay,.

A

C

Good, uh and so I asked jerick and he's like some of these are just really problematic and they should probably be.

B

Yeah and uh yeah.

C

Some of them are just real simple. The simple ones have already been done. Apparently,.

B

Yeah, they were probably first first picked up.

A

Yeah I picked chances.

B

A

B

A

Yeah I picked some that were like buried deep within pipelines.

A

That was what a journey also I found one that was like for creating a merger quest, but it was like a very weird fringe case in which, like there were no commits and nothing had been created, it was very hard to get the things set up, but yeah working with the dev really really helped.

C

Yeah definitely schedule some time with somebody dev related.

B

Also, he's really.

C

B

There's um there's someone I'm working with on the merge requests the front end, um so I have a little bit of help.

A

Yeah, if you can't get anybody else, I'd be happy to try and look at it with you, like. I don't know what um issue you're running into right now, um but yeah you, you have a couple of failing pipelines so can always dig into that and try and figure out. What's going on.

B

Yeah- and I wanted to do this, like as a learning process but.

A

Oh yeah, it's so it's a learning process yeah. It does feel good, though, when that thing gets emerging. Yes, I actually changed git lab for the better.

B

C

No, I I had that problem where it just if I could not get beyond any sort of progress to have it uh to like, actually show work right. It's just like. I can't even find the thing yeah so because of that I was like okay. This is ridiculous. Great.

B

A

Don't forget to add your screenshots, because you don't want to have to reset your environment to look for the old one.

B

I already have it, I already have it ready to go um all right. Let's move on to the next wait. Let me share my screen. Okay,.

B

Okay where's, my screen can y'all see my screen: yeah yeah, okay, so just real, quick.

B

Well, I know austin, you had a comment, but um what I'm looking for is feedback on are the source links. So these link out to um like the github project, for instance the destination links within gitlab.

B

um Do they need separate visual treatment, or is that not necessary.

A

You mean like because the pathway goes to different websites.

B

Yeah, like this launches in a new window, sorry, the source link launches in a new window and destination um launches within gitlab.

A

um I would say from a ui point based off the comment that I dropped in there. I think it's a good mentality links go buttons do so. I don't know if that has to necessarily be different, but I like your idea of trying to make it clear that the destination is within gitlab.

A

um Do we have like a tanuki icon that we can put in there as if it were um just a part of the link itself.

C

I don't know like a badge or something because we have badges on links already and that.

A

Will work yeah, perhaps a badge or something.

C

Yeah they have. um I know I saw one. I.

A

Mean we must because they're in the button right. So if you go to the previous, the screenshot to the left, there are icons next to all of those.

B

C

There's even like an a base: black and white tanooki icon, yeah.

B

A

I don't know if, like uh uh pending, that as a prefix or um maybe putting it in the table, header for the column, even maybe that would help draw a distinction that, like this is gitlab, and this is.

B

C

The there's badges that have icons.

C

And there's icon badges. So if you go to look at the badge, that might be something like what you're interested in.

B

Like well, I have a an issue that we're gonna update the icons that I've been working.

A

B

On and we're using a badge for gosh hold on.

B

For the status column.

B

B

um I did have originally the external link icon next to the source link and mike suggested, to take them out, because it looks really redundant to have all.

A

Of those icons.

B

Yeah, I'm imagining that um if we have a tanuki in front of the destination like that, might look a little redundant is.

A

That yeah always get led, though.

B

A

Is the git lab sorry? Is the destination always get lab yeah? We could say the column header would be get lab destination. Then.

A

That might be clarifying enough, I mean each of the link is well. I know that doesn't make sense, because technically they could call there or whatever they want. It doesn't happen, so I say gitlab in it so yeah. I I stand by that. I think you could change the column header, maybe to get lab destination.

B

A

B

I external source: what do you think about that, or is that not necessary? I.

A

Mean I think, that's a bad idea. I think you're trying to issues where, like you might have imported a source from another gitlab project, and then it's not really external. It's kind of internal potentially, so I think maybe just keeping it source makes sense. I think it might. I think it's gonna work just fine, because, assuming that these links are representing the domain, they came from they're kind of self-explanatory for github bitbucket yeah.

C

Whatever they might be, that's what my question would have been is that what's the purpose of seeing whether it's uh an external versus a gitlab.

B

I mean that's my question, so even in the url we have github github bit background. So I don't know if that's enough.

C

Well, how big are the urls too, what if they go longer than that.

B

They'll have an ellipse.

A

B

um Okay, cool. Thank you for the feedback. Austin. Do you want to vocalize your question.

A

Yeah I was curious, so the history takes you to a new page from my understanding why I interpreted your job to be done statement that history would help the user know. Do I need to import more projects from github, because if I can see that all of my seven projects have been imported, then I'm good? I don't need to do any more necessarily, so I was wondering.

A

Would it be reasonable, instead of going to new page, maybe having a collapsible section or something below the top of your page below the import project from so the user could quickly identify that, as opposed to having to go back and forth.

B

Well, this could be quite a long page, so I don't know because there's so much information I don't know if it would make sense to put it behind an accordion.

B

A

Yeah, I I think, inevitably, the user's gonna have to probably sort through a long list. Anyways um probably you have to cap it. I don't know 25, I'm making up a number here of projects that can get listed at one time. The intent I thought behind the accordion is they wouldn't have to load it when the page gets built, they can maybe load in that data when the user hits expand. That way, you're not like making them wait on pulling every single project or something in before the page is ready to render.

A

I think my question was just: does taking the user to a new page, take them off the task of importing a project.

B

B

Well, the task.

A

Is for thought, you don't have to yeah.

B

A

No, it's just yeah.

B

Something to consider okay, that makes sense.

A

But yeah I was thinking for myself like when I was carrying over all of my stuff from github um and I was doing an import, I'm like well yeah. It would have been nice to have seen. I have these four projects. These three are done. I know we're just the last one to do and I'm just going to hit github again.

B

A

Whereas here I'd have to go, look and then hit the back button on my browser, hopefully go back to the right page and then hit github.

B

This is also a place to view details that have failed, which will take you to an additional, a new page right.

A

Right yeah now that makes sense because then you'd be going down a different workflow. In my case, I was thinking I've clicked import for three projects. I know in my mind, there's a fourth one that lives out on my github page and I just need to go grab that one too.

B

B

Interesting, I.

A

Mean it's not a problem to load them then, like maybe no accordion and just like this table, as you show now just sits below, but I just was assuming there'd be some sort of um constraint there.

C

These are loading um all in tandem or they're, just all coming up at the same time, or is it all kind of.

B

The progressive like are you talking about, would it be infinite scroll versus pagination.

C

Well, the pagination they make us, um I think, is a limit. I don't know any.

B

C

Of infinite scroll, um but otherwise yeah I mean how often because it says the status done versus failed, and I think you mentioned that earlier, like what happens if you walk away, if you close the window or whatever, and you come back to the screen.

B

You're saying, would it load the same.

C

uh Would it continue showing new or updated data.

B

B

Yeah and all these would take you to a log, essentially that's interesting.

C

I think I should be.

A

Fine all right well quick time check for two minutes over. Do we have any uh final thoughts.

B

B

I don't have any have a good week.

A

Yeah, okay, so good week too,.

C

Cool. Thank you.

B

C