CHAOSS Common Metrics Working Group, 3 Oct 2019

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: CHAOSS.Common.October.3.2019

Description

CHAOSS.Common.October.3.2019

A

B

A

B

Welcome people welcome.

A

To our chaos, common metric working with me on August.

A

B

Okay, let's, let's delete and start over no just kidding.

C

B

B

C

The 19th, the geography metrics that Google by yep.

B

C

You go, I, went ahead and just kind of rounded this out and I thought. Maybe we could kind of put some closure to this today. Okay, oh I am I went ahead and I updated the headings and don't mind the molding and the italicizing just because I was just doing that. So we could visually see what's going on here, but I updated it to the way that we've been talking about headings in the context and then a lot of thanks goes to the folks yeah and we're more lab who have spent time.

C

I honestly think the last was it the last two meetings, Georg yeah, trying to think through this metric and they provided some what I believe to be like completely proper and helpful visualizations, and so that's that's there yeah.

B

That looks great yeah.

C

And then tools providing the metric I have not heard anything from auger, so I removed it and so clearly were more lab. Is doing this or it's possible in glue more lab and that's I mean that's how about it and I.

C

So I don't know what people think to me. This is good because I think it's a simpler way of presenting metrics I. Think it's just here's the description, here's the objective and here are some ways of looking at it. That's that's it we're not losing people in these long discussions about.

A

So one question that I have: we have two different options for looking at time and one is local time of the contributor and one is universal time global time. Yep.

C

And that graphs different.

A

Eric, the grass will be different, depending on which one we choose. Okay,.

C

A

So the question is whether, in the description or filters we want to mention when to use which and what they mean, whether you look at the local time or the global time.

C

And I assume in more lab it's fairly easy to set that up one way or the other.

A

B

So in grimore lab, that's something you you pick whether you want to display the time as local time or whether you want to display it as UTC or are they used in different parts of the visualizations.

A

You have to choose which you use for the visualization okay,.

B

A

It's a matter of selecting which data twice, whether you visualize the local time corrected by the time zone or you're looking UTC time of all contributions, and so the difference that I see is if you're looking at the local time.

A

Then you see whether people in their own time zone tend to spend more time on it during the day office hours or evenings three times as a general way to look at it. And then, if we use UTC, then we can see basically we're in the time zone which time zone basically gets most activity. Although that we can also look at which time zones getting to different.

A

B

If you look, if you use the local time zone, let's say I I make a commit at 8 o'clock in London. You make a commit at 8 o'clock in Omaha. Does that show up at the same time, because we both did that at 8 o'clock our local time? That's.

A

B

Then, if we used UTC time, I committed something at 8 o'clock, you know I'm, basically it kind of in UTC time, and then you committed something at 8 o'clock Nebraska time it would. Those would be wildly different times that would be offset by what is it 4 or 5 hours or something okay.

A

So I think that's a differentiation.

B

Yeah, that's a really good, a really good point.

C

With her that's.

B

All she's gonna ask because we have aggregation of time by time zone I, wonder if we want to blow that out just a little bit and talk about how you can either aggregate the data by local time or aggregate the data by UTC time and what that means.

B

Would you agree Garak that that's more of a filter, because it's the time like the metric is stored, the same way either way.

B

If they're a document anywhere that describes how how that's done I'm wondering if we can having, because it's it's not necessarily intuitive unless you're you still looking at this type of data.

C

B

Or or somewhere else, somebody talks about the the differences between the two. It might be useful to link to a doc with a bit more context to this. If we had one, if not, we could probably I, don't know, put some description around it or footnote, or something I guess.

C

I'm thinking that could be a nice pointer just to a reference.

B

A

We included a description of how to generate this metric in criminal so which fields to use.

A

How to build the visualization basically, is that what you're talking about.

A

B

That that could be basically anything that explains the difference between these two, because I think people will look at that, and not without being able to ask questions, wouldn't necessarily know exactly what this is. What this means.

C

We could point them to this recording.

B

Watch this video and.

C

B

Understand how the time works.

C

I'm looking if there's anything online I mean like even just a page that.

B

C

Know that describes what Georg was talking about.

C

It's almost just time zone things.

C

Here, are you suggesting that, in the tools providing metric.

A

Sorry I had a hard time hearing you. Can you repeat that please I.

C

Saw you put that comment about having a description on how to build the metric yeah? Do you think that description would go? You had highlighted tools providing metric? Is that what do you think.

A

C

A

I think that's the best place for it. Okay,.

C

A

Way when you come to this metric- and you see our amazing visualizations- get step-by-step, how to actually build them. Okay,.

C

I asked partly for this, but partly is if we start applying the new structure to say, evolution metrics. How, then they play out.

A

Yeah, if we look at the metric that I make to you.

B

A

Has it under development limitations, which is yep.

C

A

C

This thing: okay, okay, cool.

A

And I'm torn a little bit between putting it into the metric versus just linking to CBO so another place for one, because it's we're just describing the metric here and then we can describe how to implement it elsewhere. But at the same time, if you come to this metric, you just want to know how to use it, having some step by steps right here- they're not long so anyway, I'm trying a little bit and I can go either way.

C

If the, if you pointed them out to some tooling Ikram or lab I mean, is it just pointing to the tooling kind of broadly or is it some way.

A

Right because we don't currently describe anywhere how to build this visualization, okay,.

C

Obviously, I'm always I'm not a huge fan of putting steps into documents, because if anything changes.

B

In grimore lab, like.

C

A new release, the steps might change, and then we have to backfill that so I have expressed this stuff in the evolution meetings a long time ago, but I mean I. Guess if we reversing the metrics and part of versioning would just say: hey. The new version includes a new set of steps. So maybe it is fair to include it here, particularly if it's not that long, yeah.

A

B

We need steps for creating the visualization I thought the visualizations are more just so. People can see how it's used more of an example or did I miss something yeah.

C

So I think it's about the steps that would be undertaken to differentiate between local time and Utah.

B

Oh I see I, see, I, see.

C

B

C

Of address your concern I'm just giving some people a little bit more understanding of what what we're talking about here, yeah.

C

I'd say maybe go ahead and provide this time scammer. If it's not too much trouble.

A

B

Other option would be to show two different visualizations one for UTC and one for local for a relatively small project that makes it kind of obvious with the differences I don't know if you would have an example that would that would be like that.

A

Yeah I'm waiting for.

A

For the latest update on the dashboard and the chaos dashboard, because in the next update people have support for this metric out of the box, but it's right now, it's not out of the box so describing the steps right now is more difficult. When.

C

Will that be done.

A

What's the next update whenever that will be that'll.

C

Be before FOSDEM I'm guessing.

A

Yeah I'm thinking sometime October.

B

A

B

To do this before we release it, no stress about doing it right now, yeah.

C

So how about that? Because then we can just point out to it this metric in practice, and particularly that's part of release, does that work. Okay,.

B

Anything else on this metric I think.

C

It's mostly done, maybe this issue.

B

Garrick, the other thing I would like to add to the agenda, or either for October 17th or for a future meeting, is kind of similar to what I've added for for Sean, because I know that grimoire labs, like with this metric, has a bunch of stuff. That's common metrics that are already implemented that are in the software that we've never bothered to define and.

A

B

I'm I'm wondering if it would be possible for you to kind of take this back to somebody at the church, eeeh or or just do this yourself and come up with, even if it's just a list like a subset like a list of a few that we could work on, that, you could kind of describe. You know, show us how they're implemented in the software and that we can define for the next release.

A

Yeah I understand.

A

What the evolution working group I think two weeks ago was to.

A

Go through good data that you're currently housed and I'm looking at them right now, the schema they're gonna post a link here in the chat.

A

So the room or lab data.

A

Mostly around the different data sources, and then some of them are combined between data sources and so from there we can define metrics that use the data as they are here.

B

B

B

Wonder if we could maybe pick a few, maybe even it was just like, like maybe three to five- that fit within common metrics and maybe are ones that you know are used a lot by people who use scrum or lab so, like maybe popular ones, and we could try to define those and see just see how this works kind of validate. My my thinking around this.

A

B

Not necessarily right now, this is something we can think about. I.

C

Mean if I just look at I get CSV yeah I mean when I started. Thinking about like.

C

Names of people like author that seems like it would be.

C

You know, author name.

B

C

Ten yeah, that seems like a very likely candidate.

A

And we already doing all that one name which is the next one yeah.

C

I mean that would be another we have organizational affiliation. Is that that mm-hmm.

C

So um github repo seems like a good one. Another candidate me to know where the work is being done. Mm-Hmm.

B

C

A

We can, we can have a nice heat map where the activities happening, which recalls actually.

C

Those two right out of the gate, some sort of author, the name of a person and then a location of activity. Oh yeah, that was make a ton of sense from a common perspective. Mm-Hmm.

B

A

What are labels so in the Rogue 41 we have a repository labeled, and this is something that we use for grouping repositories and defining projects. So if you go to the chaos work and the chaos dashboard, the working groups have their own label, which, for example, allows to have the many lists and repo for the DNI working group to easily filter by that.

A

So it's I, don't know if it's a metric or just a tool for filtering to be able to tag and label data.

B

And you said that was the repo label that you were talking about on 41, okay, I.

A

B

Just adding these to the agenda for next time.

A

B

So we said author name repository name, what was the? What's the other one? We were just talking about label.

C

B

Label, no, the one before that, the one you mentioned. First, oh.

C

B

That's right, it was repository cool.

A

Guitar repo yeah.

B

Okay, I, don't.

A

Understand the difference, that's.

C

B

C

Start those would be great because then combined with the one we just did.

C

B

Happened where he had been working on the organizational affiliation but I think we.

C

B

Is working on one? He was working on time to response. Okay, our responsiveness, metrics I, think in general, of which several are in here as well.

C

B

If you look at the so.

B

C

Time, stamping there's.

B

Time to commit hours, but not looking at that's that's the time from when the that was originally created to when the commit was made to the repository, is that the difference between the time like I, commit it to my local fork and it gets merged into the project. Is that the difference there.

B

Dude, do you know Derek I.

A

Don't I think what we want to go after is pull the clips at that point where we can get time to create corner questions I'm too nervous. Oh I,.

B

See that's a different CSV file got it.

A

You did I, don't think github that differentiation between authored and permitted yeah.

B

I didn't think so either, which is why I was kind of curious. What that meant in this github CSV. It.

A

Comes into play when you have patches that are distributed via email and then merge manually I, don't think it I.

B

See I see, select the Linux kernel model yeah.

A

B

But this pull request. One does have several of the responsiveness metrics that we were talking about with Daniel, also time open days time to close days time to merge, request response.

C

A

B

Pull requests CSV.

A

B

At row, 54 I.

C

A

A

C

Those all seem like anything around this month, yeah.

B

Yeah exactly and we.

A

Kind of start with pull requests and then generalize it from there because responsiveness, also existing email, yeah.

B

Issues there's all kinds of responsiveness: metrics. We could use cool.

B

Okay, so maybe we will tackle those next time, I.

C

Will for next time too, what I'll do is I'm gonna I'll go back through the route metrics that were released last time and try to rework them under the new headings yep. So those are done. Okay,.

C

B

I'm gonna move these out of the agenda for next week and put them in the minutes just because I think the we, since we continue to talk about them.

B

A

More interesting.

A

Feature requests that will probably be implemented in the next couple months is to track progress of issues that are then github projects, user interface, but you can have a Kanban and all the issues are moving through the different stages of the Kanban. Oh.

B

That would be interesting. This is the github project for it's quite a lot in kubernetes.

A

Mm-Hmm so yeah, that's a feature request from a potential customer that still needs to sign on, but they signed on the loads of all of that feature. Nice.

C

Never seen these.

A

C

Project boards, oh yeah,.

B

C

Long have they been around I.

B

I think we've been using them in kubernetes for close to a year, whoa.

A

B

Don't remember not using them as part of the project.

B

Yeah I mean it's a really nice way to keep track of things when you're working on something big. So we use them for for the conferences for sure. So the commuter summits when you organize those because those are organized by the kubernetes community, and so we talk a lots of stuff in the and the boards for those and I know.

C

B

Use them for other stuff within them.

C

You can you can leave issues in them.

B

Yeah the issues are actually just github issues, so you basically put that issue in the board and then anytime, the issues updated it automatically. Like changes, show up. Okay,.

C

B

All it's all sort of integrated, so basically you tie issues to the board and get up knows where those issues are, and it knows you know if things are updated.

B

We could we tend not to create issues for all the stuff we need to do. Oh, the other thing we need to do is we do need to gather the list of people who've agreed to organize and submit a pull request for that yeah.

A

B

I noticed that that was the one thing that's on the same page, just the CSV, but that we haven't updated and I wasn't sure if we had a list yet- and it was all kind of an email and I sort of lost track of it.

C

We did good support from gitlab.

B

Support meaning like a sponsorship, yeah, sweet.

A

B

Well, the nice thing about the that you miss is it's it's way less expensive and it's pretty easy to I think we're allowed to just like bring stuff, or at least they have in the past. They've just brought snacks, so we just bought stuff and brought it in with us. It.

C

Was kind of like the Vancouver remember out of the University.

C

B

Double double checking with the hotel, so don't get in trouble. We should see how much it costs for the hotel to provide coffee cuz that would be easier than trying to lug it. One.

A

For the full day, it's 15 euros per person. Okay, let me fill up the room with 70 people we're talking about 1050. Oh.

B

It's still not bad. Oh that's.

C

Pretty good yeah.

B

Given what hotel hotels which ours it's like thousands of dollars for coffee, like you, gotta, be kidding me, it's crappy coffee, I'd pay thousands dollars if we're getting cappuccinos, but not not coffee. Out of like earns it. You have to like ya, know.

C

I will maybe you could stop recording here. We seem like we're done with oh.

B

Yeah I think where we've diverged.