CHAOSS Evolution Working Group, 21 Nov 2019

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: CHAOSS.Evolution.November.21.2019

Description

CHAOSS.Evolution.November.21.2019

A

All right welcome everybody to the evolution working group call. It stays December. Fifth, it's not December. Fifth, that's the next meeting today is November 21st, the 2019. Let me go ahead and post the the meeting notes in here for everybody, so I figured today, there's not a whole lot of like pr's and stuff to go through I.

A

Think the main thing that I want to try to work on today is just kind of talking about some metrics working through some stuff together and kind of just turn it as to think through some of what the ones that we want for release are and then trying to work. Those out collaboratively. One thing I did want to bring up really quick, though so in December.

A

We have two meetings scheduled on the fifth and the 19th I'm, assuming the fifth won't be too much of a problem since its kind of so far away from like Christmas and like.

B

Holiday breaks.

A

For most people, but the 19th I know is pretty probably pretty close for a lot of people might be going out of town you doing other stuff, so I just want to get a general feel. Should we just go ahead? Have that meeting I'm totally fine with that I'll be off of school, so I'll be totally free, but if we would rather just cancel that one and continue to work until the metrics release, anybody have any thoughts on that I'm.

B

Available call me around okay, solid week before Christmas, so, okay.

A

Sounds good yeah I just wanted to bring it up something that I thought about. Ok, so last week, Sean Georg and I worked on so I've got a link to the this Google folder, a goof dry folder in the meeting notes, and so what we did last week is we sat down the three of us pardon me and worked on the issue. Resolution duration go ahead and share my screen. If that makes things easier.

B

Coming in alright someday yeah I mean I'll just go right to the okay.

A

Perfect, so what we did is we worked on this issue. Resolution duration, metric and kind of talked through some of the stuff together. I think we do we uh opening to be close, that pull request. Georg. Do we merge it?

A

B

A

I, don't think we did I think we left it open. Oh.

A

Yes, we did not close it, so this one that PR just kind of goes through the Ashura, so it just basically copies it over and deletes the old file. It doesn't delete it so much as it renames it. So it's pulled to 64, which I will put that in the chat.

A

So if we want we can we can talk about this metric if we think it's ready to be merged, we can do that. Yeah.

B

There's just I'm gonna add.

B

Resources at the bottom.

C

B

C

Looking at the.

B

B

Maybe, for each closed issue is to resolution timestamp of issue closed timestamp of issue opened. Maybe it opens.

A

Which part of you under implementation.

B

In the air mutation.

A

Oh yeah I think it opened, makes sense.

B

Okay, yeah I would recommend taking this and putting it into a PR you're already yeah.

A

It's in a PR so and then the adding their resources section. You said as well: yeah.

B

So if you can update the PR for that yep.

A

What should I put under the heading just leave a blank in.

B

Terms of like updates there.

A

Like content under resources, oh yeah.

B

Just leave a blank.

A

A

I'll put a put none for now: okay, so the I've changed that to say opened instead of opens and then how the resources anything else from anybody before I make this commit.

B

Okay, so I'm the spreadsheet, I'm marking issue resolution duration, as basically ready for next release, sounds good to me.

B

A

There are no objections. I said we can I think we're going to go ahead and merge this yeah.

D

A

C

B

Then a hook a link to the spreadsheet there for folks, if they just want to track what I'm doing thanks.

A

D

A

Then also this poor crested gay opens a couple days ago, just about adding the funding die, animal which I'm super in favor of just want to bring it to the attention. Everybody I think it's great to go ahead. Merge for! Okay, with that yeah.

B

A

B

Looks good I think yeah.

E

We already have the metric, so I just went and added.

A

And we appreciate that all right, so we've got those PRS merged. Let me oh sweet I, think your evolution.

B

In that folder that you shared are there other metrics that I.

A

Don't know which ones are closest to being done. I think sean has been working on some of these I know. Contributors has a little bit of stuff written for it and I figured. We could just start at the top and work with that one. So the link is in the the notes, but I'll go ahead and post the link to the contributors document directly and I think just kind of talking through this together and working on what.

B

A

To say would be a fruitful exercise, so what it's got right now is the the question it's trying to answer is how many contributors, I should probably say, does a repository have and the description is I guess an indication of how many contributors and then defines what a contributor is.

A

So is that definition of a contributor.

B

Should we call it a repository contributor, I.

A

Mean well I, think I think they'll be fine. Is there some like? Is there at? Would you imagine there being a separate metric for like organization contributor, perhaps or like file contributor like to a specific set of files or something, or is it just to make it more clear.

B

Well, just because cuz like so, for example, the email that gave word posted today about diversity, access tickets and getting that up and running I mean that's a clear contribution.

C

B

The community, okay, so.

A

B

That would that effort wouldn't be captured. Okay,.

A

Yeah, in that case, I think that's odd.

B

A

I didn't I think repository contributors would be a good shift, cuz I think that's! Then we can still define a separate like all contributors metric like in any shape or form so.

E

They, if they say repository contributors, then we have to remove issues and pull requests, because that is not part of the repository issues. Are discussions and pull requests are code reviews and they have an outside of the git log and I'm, applying a very narrow definition of repository well.

B

Is there a better word than repository and this kind of not loses the other contribution? If that's all,.

E

We sometimes refer to people as Oh.

E

Contributors, but if they are commenting and reviewing on issues and pull requests, that's not code either. Mm-Hmm.

C

A

I, don't know how I mean when I guess maybe like project like the project, consists of issues, pull requests and code, perhaps, but then also you could define the project. I think to also include the case we're talking about like getting the diversity tickets, like that's part of the effort of the project. So with that count, so that might be to j'en. That may be too generic could.

E

Be so let me just leave it contributors and add a note somewhere that.

E

Anyway, slack IRC I mean.

B

That's the other option just to say this is about all contributors, and then we don't have like a I know, though term might not be the best. But like a repository contributor versus a write.

C

B

Contributor verses.

E

And we can have that as a filter.

B

That you could filter on repository activity. Is that what you're, saying or filter on event activity, or whatever it might.

E

Be okay filter by location of engagement which clicking.

B

Review participants.

B

Event, participants.

A

Would we consider things like documentation and marketing to be part of you.

D

B

I'm not sure that's helping the cause of.

A

I'm not sure how to say how.

B

A

B

Just worried yeah the appropriate thing I.

A

Think it I, don't I think it's been a a line, a long time trying to looks at all the different things. I think at this point. If people are reading this, so they would be able to figure out like pretty much anywhere. Somebody contributes.

B

A

I think that's perfect.

E

E

C

Here yep, so am I.

C

A

And so then, we should probably get rid of this section all right and in the description about defining what a contributor is, as somebody who commits code creates an issue creates pull requests.

B

B

Yeah, that's it I mean it doesn't have to be complex.

F

A

Think if I want to know how many contributors yeah.

D

A

Think I mmm I, don't know if like size of the project is a good way to like you know like hum lynnie is I, mean, there's a difference between how many people have ever done anything to the project like you, no matter how small versus like actively maintained and are actively involved so I, don't know if, like sighs but like almost maybe what's the the reach of contributions like that might be something thinking about in terms of like a bus factor which is I, know a totally different metric.

A

But like are there a lot of core contributors? Are there a lot of ephemeral contributors, I.

A

Think I'm trying to at this point its well, we didn't sweetie I, don't think myself onto a circle.

C

It's part in the nine.

E

Days, I was also thinking that we could make this metric more broad by saying who contributes through a project and then in filters or visualizations or somewhere. We say that, like the count.

A

E

The count into a number but I know that many people who use Petrucci dashboards want to know okay, who is actually doing things.

E

Which, then, that makes it into a common metric metric, did.

B

You say that again the our guy was tapering.

E

To two things, one I think contributors is a common metric mm-hmm because it answers a question about who's active and then the second is rather than asking how many contributors doesn't project have to ask who other contributors in a project? And then we can have an aggregator of this some number of contributors as a way to visualize that information.

B

Like that, the question yeah.

F

Yeah one other thing: is we measure contributors or contribution? What, if an takuna artifacts, we deliverable, for example, what a particular release like what New York said. Then at the end, we can still a Grenada okay in this project. This people been contributing, but what's this particular release who are like the active people? So if you want to have, is natural at any given point you'll be able to see those who are actually active, like the current contributors activities that all the time you'll be able to see those who have contributed as well. I.

E

Like that, I added it in filters, I add a release cycle as a filter and then also time of the pivot reading project.

A

This is a pretty small knit kind of an almost a nitpicky. Not really, but should we say who are there for the question? Who are the contributors within a project or to a project kissimmee within is like I, am part of like I've, been in accepted into the project almost in the kids, like a barrier versus to is just like I'm doing something for them like I'm, giving it to them whether I'm inside or not. If that makes sense,.

B

Something is important.

B

Org you had mentioned in terms of visualizations does for moral I'd. Have these types of visualizations.

D

I'll pull in a screenshot in just a moment: okay.

B

B

And I know that, because.

D

B

Talking about your the way that auger now houses facade, that would be similar, I think.

E

Okay, what's the connection between contributors and facade, facade.

B

Shows who is making commits to the project.

C

B

Do a screenshot of that Carter, just yeah.

C

B

B

So that would be under a list of contributor names.

C

B

C

All right get there so we're the.

C

It's being weird.

A

A

Okay, that's, but eventually, here we go.

A

While you guys were doing that, I also I added an aggregator and a parameter, just the aggregator being account and a parameter being the period of time for which we are considering up under implementation. Just wanted to mention that okay.

B

I'm gonna pull the Zephyr Oh.

C

Bernie sorry.

C

D

So I'm thinking.

B

Well, I had it that ready it's up atop. This was a little stall. You did. You do. I didn't even see that Levi.

A

B

Fur, let's see.

B

C

Here's this event, I.

B

Just pulled that from DNI.

A

Looks good to me I like it oh well,.

B

We have to pull them to rewrite these questions.

A

You say we need I assume we need to rewrite these questions, though you said man.

B

I'm just pulling it from diversity just so we can.

E

So if we ask the question, who are the contributors to a project.

E

That's what we are trying to get it right, yeah.

B

I guess, if we're asking community member yeah.

A

What is the Likert scale? I got.

B

Rate 1/2 7, how you feel: okay,.

A

C

Didn't know they had a name.

B

Does that surprise, you.

A

B

We need to change these questions to.

B

The formatting didn't.

E

There was a table mixed in I know.

B

I, don't I just it was weird I, just pulled it up from the DNI PR. You know how I believed that to be a table, but whatever we can.

E

Just removing the table okay model so now it should be fixed.

B

So I don't know what a survey would ask here.

A

Like maybe do you feel your contribution, it's important.

F

mmm If it's not what done it, why did they do it? That's.

A

Like that's a good one, I.

B

Wouldn't that wouldn't really be a Likert scale, question.

E

A

C

E

E

D

A

So is the is the first one just like I contribute jet a scale ones um Jeff you I'm contributing I like that and like to them.

D

Yeah, it could.

B

Be okay like if I asked you on a scale one to seven? How much do you feel like you're, contributing to the chaos project? Okay,.

E

Yeah, the is absolutely has an item here like.

B

E

I'll have to look it up. What it exactly is: yes, Senora.

D

E

They call it how often do engage in the following activities in the project or in the days.

B

Yeah, why not just reuse that you know.

B

Okay, so I like the idea of this actually becoming a common metric, I, agree yeah, because I think it's a way to like. If we're looking at time zones or the types of contributors, this um we have like organizational affiliation, that these are all things in common I think this would be a useful metric there I.

A

B

C

B

This is done, I'm gonna, make a note here. Contributor, so I'm just gonna say: moved common.

A

In the really sheet, yeah.

B

So in evolution has moved to common.

B

I'm gonna mark it actually is red.

B

And then in common.

B

Definitely so they have a contributor, they actually already have a contributor metric there do they.

B

How is this better?

B

They don't have one they. It was just on their list of the who's. So point being they were thinking. The common working group was thinking the same way, good.

F

E

F

Can also add that might make our work useful, for both the academic and industrial is try to find out how much the language of the programming language affects contributions, because a lot of work is not yet done in this area.

F

You know, like we whip some people like criteria, people use to choose the project that is joined, to contribute. What choice or the language itself is also something that interests me. They don't know much. What in the academics as not really Colin's area.

B

So would that be it's not really a location issue but implement filters so Armstrong? Do you think that that so I thought the language would actually be kind of a top-level filter as to how we think about contributions? Is that right, yeah.

F

Because you see like most ecosystems, that I viewed on Piton these days or, let's see something like genetics, kind of language level tends to attract a lot of contributors right.

B

So your contention is is that if it's in a fairly obscure programming language that might inhibit contribution, some.

F

Something like that, we might now have little information to make meaningful conclusion. Okay,.

B

A

They probably would be applicable to every project, but maybe also dealing with like certain languages within a project like there, if there's, maybe a back-end and front-end Department in the backends and JavaScript and the or the front ends and JavaScript in the backends and Python, maybe also seeing you know which parts of the if it's particularly a large project, maybe the back end, has far more contributions than the front end, because maybe more people know Python in that ecosystem whatever. But just thought.

C

A

Yeah I really like this well.

B

I didn't move evolution forward, I'm gonna work here, but.

E

We did amazing work. Look at us. We.

A

Did and you know what I feel like if it's common it affects everybody, so so.

B

I'm gonna somebody actually issue. This is a PR in common yeah. Oh.

A

The oh, the images are gonna, come through funny, aren't they yeah how.

E

Do you want to share the images with you, because I can give you the raw screenshots that I have.

A

If you just send me an email and I can just drop them into the PR I think that should work right, yeah.

E

Covered Landis at gmail.com, yep.

A

See Carolinas at gmail.com, that's one.

B

You know sent you the one I used. Thank you much.

C

A

So, where should I issue this WP.

B

C

A

D

A

Have been looking at the folk.

A

Under the the common repo WG common right, mm-hmm.

F

A

So I'm I'm just trying to figure out which focus area. It's probably the who focus area no.

B

Yes, in the who, okay and.

A

Then I just need to create.

E

One of the suggestions I have for the names on the screenshots we have is to blur them out.

A

Good idea, that's.

E

What I've been doing.

C

A

A

Yeah I can I can do that after the McCall I don't want to.

A

Okay, so you're welcome common group. We love you guys.

B

So we're all good there perfect! You.

E

Know you share more with common than you realize. We share the same timeslot, just yeah.

A

A

Do a merger, okay, yeah I, will totally get that finished up after college should be no problem. I know: we've we got about 20 minutes left, so we probably could work a little bit more on another one. We might not get all the way through it and so which one we pick and how fast we go, but sounds good, so I know can so we just did contributor, there's issue resolution duration, that we are we've already done.

A

We could do issue contributors, not sure if that would probably be another I mean this one is specifically related to issues, so it might be more in line with evolution, but again it is measuring primarily contributors. So it's probably also a common metric.

B

So maybe, okay, if I, actually, if I look at issue age.

A

The open issue H.

B

Yeah, it's just called issue age.

B

Maybe that one started a little bit yeah.

A

I think I think that one's good that.

B

One shouldn't be too hard: nah.

A

All right, let's go for it.

B

So the metric is an indication of how long issues that are currently open have been open. So the objective is to identify.

A

So to me that the question in the description there's a bit of a maybe it's just my understanding of the language to can see but the being what is the average age of issues open? Does that mean across all time, like on average? How long is any one issue likely to be open or more along the lines of the ones that are open right now, out of all like? What's the average of those currently open.

C

Issues issues, age, issue, age,.

B

Well, we I mean, obviously those are two.

B

Two things that we could think about because they are.

B

How long issues that are currently open have been open, so I mean this description is the ones that are just currently open. So maybe we'll just stick with that. Okay, what is the average age.

B

Which is was the average time online issues that are currently open have been open. All right, so I do identify issues that identify open issues.

B

What does does augur do this? Maybe you could go pull the API yeah I'm.

A

Pretty sure that it does.

A

Look at the exact implementation details.

E

B

Run issued age and so issue age was.

B

Yeah, so the one of the questions that we talked about was this the age of issues that are currently open or is this like the age of all? You know issues.

E

If you, if we look at the meaning behind the two options, yeah.

D

E

The age of all issues just means: how long has the project existed, because if we continue to count the age of course issues, then they just keep increasing the number every day. But the only useful piece of information is to look at the open issues. That's.

B

What we opted, that's where we learned yeah.

E

B

You can see the question: where does the average time that an open issue has been open.

E

E

B

It was half and I. Would it as.

B

The metric is an indication of how long issues that are currently open have been open. The objective you know I mean it could be a couple things to identify oldest open issues in a project.

B

Why else do people want to know this Georg.

E

They want to know how well the maintainer keep on top of the incoming requests.

E

E

Features or books or song are addressed.

F

Yeah that's important.

E

Yeah they're the way I'm describing it it's very similar to the time to first response metric that we have so I'm. Maybe you can help me think through what the difference is between looking at the age of issues and time to first responds I.

B

Think I mean, of course, they're like if you have a high time. The first response you would expect your open issue age to be low, but I mean even time. The first response, I think, can even be like thanks for your issue, I.

E

Changed it to resolving issues, okay,.

F

Time to first respond is very good, but now what happens if the time to first respond is short and then the iteration continues and becomes delete?

F

That's why we may have now an issue because you can post something like an issue in less than no time somebody responds to eat, but then does it. They take that forever to resolve the issue, so they don't have consistency over like follow-up. That's when, where another matrix I what you can come in now to like.

A

Some sort of we've identified a fix it's coming in the next release and then it's really system Philip month. That kind of thing yeah.

F

A

It just sits, there looks like it's, not nothing's happening, but.

F

A

E

Think that's why we have a host of different metrics.

F

E

Get to those edge cases.

D

What is this objective kind of hit? What.

B

We're looking for.

E

Yeah I'm not I'm, not super aligned with the first sentence and the question so identify the oldest open issue.

E

B

I'm, just I'm trying to get a some issues are open for a long time, but there's a reason why they're open for a long time. So it's just trying to get insight like it may not be a bad thing, but something's open for a long time. But if something's open for 90 days, it's probably worth yeah, we.

D

Might know about it this.

B

Is why I was open again.

E

Here so I added in here when the issue ages identify the oldest open issues that might be tied directly to the question back. That's.

D

Good okay, can we stop here? I have.

B

Another meeting at 11:00 that I actually have to go to yeah.

F

B

About that you yeah yeah.

A

No, of course, we appreciate you hopping on with us Matt thanks for your help, sure.

B

So this is issue age, I'm, gonna mark this as working on it.

C

B

Could we I mean, would it be possible? Could we do an action item that we actually try to pull this one out, yeah yeah.

A

I'll have a I'll be on Thanksgiving break all next week and interspersed with all of my setting and stuff I'm gonna try to hit on some more of these metrics. Keep it moving through for a release. At the end.

B

Of the year start, the next meeting in two weeks with like here's, two of.

A

Them that are yeah in.

C

B

Concluded former.

B

If you could do that Carver, that would be great yeah.

A

Of course, a problem.

A

D

C

D

C

F

A

Right I know: you've got a couple minutes left.

E

So I put in implementation is that clear to you is the description of clear.

A

For all of initiatives get date, the issue is opening calculate the number of days to the current date. Aggregate using the average function. Yeah, it's good at me.

A

I'm gonna put it in the do the aggregators we've been doing them like in this style.

A

So average age all currently open issues.

A

And then I think parameters be a period of time.

A

A

A

Not maybe maybe this is because.

E

Filtering modules we can filter by working groups, we can filter by.

E

B

E

Don't know by tags issue tags are.

A

A

There's one of the other one in there you said language.

A

A

Maybe issues that have been reopened like they've closed and now they're open again, maybe that time then maybe that's more of a parameter but like including like if I have an issue that opens and then it gets closed and then it gets opened again and it's open in the time period I'm looking at like do we count just the second time. That's open. Do we count from the first time it was open all the way up until now, including the time when it was closed.

A

That may be more of like a parameter like that behavior, but that's something that I think could could be captured. Yeah.

E

A

Not sure the best way to go about.

C

C

A

Don't know how to best put this parameter like I, don't want to say, like definition of an opening issue, how about like include reopened issues.

C

Whether or not shoes that have been reopened.

A

A

Feel like if it's been reopened, we should include it I. Think because I mean it's an open issue, you know but I'm just trying to think about like do. We include it from the first open. Maybe that's the better. You know I'm saying.

E

Sorry, the visualization for this metric you're.

A

Good, so if I've got.

A

Do we want to worry about the behavior capturing the behavior of having a reopened issue like? Should we include it just from the reopen or from the first time, I was opened all the way to the president.

E

But my opinion is to do it from the beginning. That's.

A

What I was thinking.

A

So maybe that goes in echo in description that it also includes issues that have been reopened.

C

I'm going to save this.

E

There are both the visualization Frost.

A

C

Working group tags.

A

Or labels- oh there, it goes I.

A

A

B

A

Are any other filters? We're gonna worry about, I think those are probably good for now.

A

Preferences I can't think of any data collection strategies. I feel like that one's pretty.

A

However, we've been continuing to define I know in other places in evolution. We've got examples of how we collect issues I'm just.

B

A

To copy pasted it from from there.

E

Or we can say the same as issue open, yeah.

B

E

Issue just see me refer to it because we already defined it elsewhere. Yeah.

A

I think that's actually I think we did that for a couple of the metrics and wasn't that one wasn't that one I think it was closed. District resolution duration or just issue resolution duration, Oh.

A

Yep I'm just going to do that.

C

A

And that all dude that'll be a relative link. I.

A

Mean I think this is a pretty good start.

A

For this same thing else, you can think of GERD.

E

No I'm mulling over the visualization that I built because it doesn't make sense to me.

E

So if I take open issues that were open 200 days ago, then the average age of all issues on that day is 200. That's.

B

E

150 days ago, it's 150 open five days ago. It's five, so I'm expecting a slope here, but the graph goes up and down and it doesn't just go down to zero because there are no issues. It goes down to a value that it should not exist.

E

Does not make sense well.

A

I'll hold off on making a PR for this one. Then.

E

Remove this graph because I don't understand what I've built here: okay,.

A

I think we've got one for Margaery as well. I know it's a little bit after 11:00, so we can go ahead and finish up but offer one from Agra in there as well. Okay,.

E

If you have a graph go for it, yep.

A

Will do alright, I gotta head out yeah.

D

Thank you, Eric.