CHAOSS Evolution Working Group, 22 Mar 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Evolution Working Group 3/22/22

Description

Links to minutes from this meeting are on https://chaoss.community/participate.

A

All right, um so this is, we did not have a meeting two weeks ago because it was just elizabeth and I, and so this is the march 22nd uh edition of the evolution working group and one of the things that I wanted to talk about.

A

um We had some. uh I just want to talk about metrics models in this group. We don't have to do it first, but I'll put it there first, just in case, and we have some previously reviewed metrics.

A

And bernard, I can't remember if these are ones that are released or not.

B

Change request, review is released and both are released. Patrick.

B

I think like can we open the issues like on the github issues, we'll figure it out? That's what I remember I released both of these yeah. Let me look again.

A

Going over to that, we have yeah. These are these: are the metrics reviews.

A

That so this is the checklist for metrics that need to be reviewed. Yeah.

C

D

I think that um so I am the one that's looking through the evolution metrics to review.

E

Okay- and I have.

D

Not done that yet so, okay, so this is yeah.

A

B

We have three uh release: metrics.

B

Two I release them like uh put the super request. One was done by a doctor german prey, so.

B

And I created a translation request for that yohi contacted me like. I don't see a translation for this so.

A

And- and I think these two met these two metrics that are listed under items three and four yep- these are release.

B

Metrics under review.

A

All right that makes a ton of sense to me. Can you hear me I can?

A

Can you hear me yes, okay, then then we're ready to go, um and maybe we should just quick review um and if we have any issues in prs, uh let's see update funnel image.

A

I guess they made it smaller or bigger. I.

D

Think they changed.

B

Group, a bigger.

D

It's a bigger one.

B

New one is like 1190 by 685, the old one was 865 by 501.. This.

F

Seems reasonable to me any objection.

A

All right and then oh, this is just our add reactions. One that's been hanging there for a while. These are just our release. um I guess we could look and see if we have any comments.

A

Yet, please add, please add issue link. These are all done.

D

um Sean this is the metric that uh this uh google summer of code students are, would be working on to visualize.

A

Right, okay, yeah! Yes, I.

D

Think that's the one mabel who's in here. I think that's the one she was interested in kind of working with, so we were chatting about that quite a bit.

A

Okay, um it is, it is a metric focused on technical contributions, um so.

A

Do we want to, I think, to all that would be required.

A

I agree with this. I don't know how to weave it in.

A

To our metric easily.

A

But this is this: isn't this is an issue not a metric, really.

D

So we need to change the name of the metric to indicate that it's only technical.

D

Or is there room for adding other types of contributions in with this metric.

A

um I'm just going to look at this metric. Well, that's the.

D

There a link in that issue.

G

There there's supposed to be, but uh it doesn't look like people have been using it.

D

F

Let's see if this.

A

I don't see conversion rate.

D

I think it maybe is down under under evolution.

F

Oh is this: is this yeah.

F

A

Conversion rate.

H

Well, the issue.

A

These seem to the filters seem to include non-technical non-technical indicators.

C

A

A

Recopy on maybe just.

G

uh I think the the objectives include non-technical as well. Can you can you scroll up to the description.

G

I don't think there's any there's nothing in here. That says this is only technical, so I'm not sure what his comment is based on. uh It might be the.

A

I have a heart, I'm not certain either, but let me just.

A

D

Would it help if we linked to the types of contributions metric, that's just a point of reference.

A

uh There's a types of in.

D

The metric itself I mean like, would it help clarify.

A

um Yeah, probably.

A

Where did I find that.

A

This types of contribution, that's a that's, a released metric yeah.

D

That's a that's an older metric that just outlines the different types of contributions. Obviously, and so it might make sense to just make a note of that.

A

A

That should I mean, I think, just asking well because maybe he's seeing something in the main description that we're missing.

A

It's hard to it's hard to know when I look at it and I see so many non-technical things.

A

Itemized, I don't know if there were any other.

A

All right these, these look like some suggestions that could be, could be made to the metric uh change request, reviews.

A

Do you wanna kevin, what's the best process for addressing these changes? Is it to change change the metric definition and then let you know.

A

So that it can be pulled up for the website.

G

uh What do you mean so any any changes made to the markdown will automatically be uh displayed on the website, so we are. We are linking directly to the to this document.

G

So so you don't need to worry about that. The only changes that you would need to, let me know about, are file name changes which would which would break the page on the website.

A

So I have the metric in a google doc here, but presumably we did convert it to markdown right.

B

Yes, uh I was unable to add when I was creating this for loop so because pull request was not accepted at that time. So.

G

Oh, so this one isn't part of the release. This is part of a release.

A

Is it under this change request, reviews.

A

I think that's process quality. What a good guess.

A

And ray's comment is developers to community members.

A

A

Second sentence under objectives.

A

See this is the original.

A

A

Okay, that seems like a reasonable change to me.

A

It's just changing it's changing code, quality of code, the quality of the software and efficiency of development.

A

I think that's that's fair, because it's not just code, that's happening in a process.

A

Quality metric like this, which is called change required. I mean reviews. I do include more general things. I think.

I

A

Yeah, okay: now should the text be bold and I think maybe the link to the data ethics policy isn't right. Data ethics policy we're going to go through and correct at the end right, correct, yeah,.

G

A

G

We're doing so none of the none of the data ethics policy links in any of the documents if they exist, are going to be correct uh and then requiring or and then regarding the uh is it should it be bold.

G

I don't know, I think I think. Currently we have it italicized, it's bold, oh, is it.

A

Yeah, whatever we did there bolded it.

G

uh I think I think, when I added it to the template, it's supposed to be italicized.

A

um And I thought that double underline was italicized as well, but it's coming across as bold and get the mark down. Okay, that's weird! I think a single star is italicized.

A

It's looking that way.

C

D

Yeah that seemed to fix it.

A

It fixed it in the editor, let's see if it's.

A

Okay and john's also asking for a visualization of change.

A

I'm trying to imagine the change request, reviews visualization.

A

Because you would want to visualize whether or not I ice given this somebody's making a lot of noise um given the metric, I think what we want to see in change request reviews is a percent like some sort of percentage of change, requests, accepted and also reviewed versus accepted and not reviewed.

A

G

Yes, however, if we don't have, if we don't have a visualization for that, I think it's okay for us to just say we don't have a visualization for that. Currently.

A

And this is change, request. Reviews visualization, I guess- is the goal there.

A

A

A

Okay, so that's done.

A

We reviewed the issues in the prs.

A

We have two: we have two proposed metrics models which I'm gonna just move up here.

A

um Change requests metrics model, let's find out if.

A

These are in the metrics tracking spreadsheet.

F

That one doesn't seem to exist.

D

I think it would go in that first uh focus area code, development.

A

I think so too. No, I was just checking to see if they existed.

A

All right so code development is in there and.

A

This one was good complexity.

A

I'd say this is a multi-dimensional evaluation of code complexity,.

D

A

It looks like we do: have a google not starting for a one.

A

A

This is, let me see if I can make this bigger.

A

That one wasn't editable, so I'm just gonna change it to an editable version.

A

That didn't work.

A

Why didn't that work.

A

All right that seems to have worked.

A

So it's readable. Can people still see all this I've tried to make it? I guess yeah, it's a very big thing. I know if I make it too big it'll be too big for a lot of screens, and so the metric model.

A

A

Code, development.

F

This just seems.

A

Community development flow metric model.

A

All right so community development, community operation, community government governance.

A

I think this top one is the community development right.

D

I think they were thinking, it would be. The whole thing.

A

This whole thing I.

D

A

D

I could be wrong.

F

But this looks like.

D

Way back in the days when metrics models were first like emerging. So I think, since that we've had those conversations about keeping the metrics models a lot smaller and more um compact yeah simpler, because this is a lot.

G

It is a lot I'm actually I'm not aware of this conversation at all, so I must have missed. I must have missed these meetings. It.

D

Was talking about this yeah? I was asia pacific like months and months ago. um They are the ones who kind of started this. I think metrics model um like actually putting something on paper.

G

Yeah, I remember this model from the uh from the early stages of the of the metrics model working groups. I'm I'm not aware of the of the this model in the context of evolution. I guess I I don't. uh I don't remember discussing it in evolution,.

A

So this is the this is a metrics model template.

A

So if we wanted to.

A

A

This is almost like. This thing is almost like context.

A

And then this is where we might begin the development of a code development metrics model. um I don't.

A

I don't have the ability to log into the chaos user to create what I don't did elizabeth. Did you see my note about giving me access to the chaos folders.

D

I did not, but you should have access to everything in there.

A

Okay, yeah, I thought I did maybe.

D

I sent it to the wrong email like the wrong sean goggins, because I have like four emails for you. So.

F

Yeah, um maybe.

D

Or maybe I just sent it to some random sean somewhere who knows.

A

What's the root of the folder, is it chaos.

D

uh Yeah uh all chaos docs. I think.

A

Okay, okay attacks.

A

F

Here I have pasted the link for that.

D

Oh there you go thanks, so nod.

A

Sorry about that had a question.

D

But I posted the link in the chat.

A

A

All right, well, it certainly looks like I have access to it. So.

A

Add shortcut to drive- maybe that's my maybe that was my shortcoming.

A

So in evolution.

A

We would create a new document.

A

Oops, I guess that was wrong.

A

And we were, this would be called point. This would be if code development.

A

Right, that's the one we would be working on. I.

D

Think code, complexity or we.

A

Had two um one was oops, one was code complexity and the other was code development, okay, I've say, code complexity seems easier to grok than code development, which seems unwieldy.

A

Agreed so maybe maybe we start with code development or code complexity. Excuse me.

A

I'll put that into the spreadsheet.

A

And I can share that in the chat.

A

There- and we could just we- could take some time.

A

To work through this metrics model, if we want.

A

It might give us a sense of what it's like to work through a metrics model. If does anybody object or not want to do this?.

D

Now I'm good with working on this.

A

All right um so.

D

I'm going to pause the recording for a second yeah.

I

That's I can make a short summary. I can make a short summary and I just explain in few lines how we could do that at the individual contributors level, the fire level and the project level. It's just it's just.

E

I

Normalized form of the shannon's entropy.

A

Okay and we're talking about complexity, we were looking at auger's value worker, which calculates a version of kokomo and indicates the total lines. The code lines comments blanks and does a kokomo based estimate of the complexity of the code and you're discussing entropy, I mean entropy.

A

My interpretation of entropy is just sort of a nascent decay through lack of maintenance. Is that conceptually what you mean armstrong or? Is it a more complex.

I

It's more complex than that.

A

I

A

Okay, so I think maybe if you could bring a definition of that to the next meeting, we we could discuss code and thank you, okay um and you know help us develop the common understanding of it. I think mine when we talk about things like code complexity and entropy,.

A

These are, I don't want to say, static because they evolve, but they are properties of the project as a whole, which I would distinguish between development process, issue resolution development, efficiency, development activity, community growth.

A

Do you agree yeah so so like do, I think, do we need another? I guess the question I'm putting on the table is: do we need another focus area for these types? I don't think so. I don't think so yeah. Where would they where so? Where, for example, would code complexity go in our existing.

I

Well, I told you how you had something uh already fitting good.

H

Complexity in that category, so I put yeah, I put code complexity under metrics models. Yeah, that's a model.

A

Yes, oh okay, and- and we could just say that you know code complexity is a metrics model, but usually metrics models rely on metrics and we don't have a metric for different ways of evaluating code complexity. So we sort of end up in this gordian. Okay, yeah. I see.

D

So if we take that model that you were using to calculate that code complexity number whatever those whatever that those columns were, we should have a metric for each one of those in.

E

D

Right but it's simple stuff: it's like lines of code lines up, like you know, oh.

J

Yeah, that's true yeah, a number of files.

D

Yeah we should. We should have metrics for all that.

A

Yeah, I tend to agree because they're a way of statically looking at I don't know if this is the way to do this.

D

I don't know- maybe maybe it belongs in common too- that I mean there's something to be said for putting it in common also, but.

H

A

I don't, I don't have a, I suppose.

D

I mean we can do it here too. Like comment already has enough to do. Let's.

A

uh Let's, let's start it here, we can always. We can always move it. What.

H

G

What are we adding.

H

A

Project level characteristics I would propose project level, code, project level, code characteristics, so they're the s, the evolving state of a project not looking at all of the actions performed against it, but at x point in time. The project contains this many lines of code in this language. This many lines of code in this other language there's a percentage of the project that is in each language. So.

E

J

A

Complex code complexity in each language, are you you're, proposing models or you're, proposing a new focus.

H

Area we're proposing.

A

A new focus area because we need the metrics for the code complexity model to create that metric smoke. For that metrics model, we don't have met the metrics that we.

G

Need, okay, so and the and the metrics that we would need for that model, don't fit within our other focus groups. I think.

A

To elizabeth's point they could fit in common; they. I don't see them fitting anywhere else which, which ones do we need um the ones I thought were um like.

D

If you look at sean's spreadsheet thing that not spreadsheet but uh code thing with the complex code, complexity numbers in it.

A

Yeah, uh so it was languages in a project lines of code by language. Well,.

D

It would just be lines of code right and then language.

J

The filter would be the language.

G

Okay lines of code is a that's a metric that, if that doesn't fit in our existing focus areas, then we should re-evaluate what our existing focus areas are because lines of code is uh that should be.

A

So this is so we have lines of code in a code, change or a commit or a pull request. We don't have any metrics that say in this project at this moment in time. It has this many lines of code.

D

Number of files would be another one, sean.

H

D

And like community growth, that's where we're changing that right, because we just copied that.

A

All right, yeah, okay,.

A

D

Like would we need a metric for blank lines of code that was in your.

C

A

Yeah we had project level files, project file, complexity,.

A

Project blank lines.

A

Lines and then I suppose, project uh comment lines.

D

These would be pretty straightforward to yeah. They would not be hard. I don't think.

A

Yeah and then like, I would see like all of the bottom ones like languages like we could do lines of code by language files by language complexity by language. Blank lines can comment lines by language and I think I've got project level files twice, but there was probably another metric in here.

A

So these so the distinction between these metrics and other metrics is simply that they're. Just this is the project at a moment in time. We're not looking at any of the change activity. That's happening around it.

A

And that's what we would need to evaluate complexity.

A

And we're at time.

G

I think we probably need to re-look at our focus areas.

G

So the way, the way that we're organizing our metrics.

G

um And that's maybe that's something to look at in the future. Here tell me what you mean uh so the the the need to create a new focus area here. uh I think we're adding the way that we, the.

E

G

Have this organized is not uh uh it's no longer coherent right, so we're we're organizing the focus areas in in different ways.

G

So when we go to the uh when you scroll to the top, where we've created focus areas around activity, development, activity, efficiency and quality, uh an issue resolution, uh so the these these focus areas as a whole, they don't really, they don't really go together right. So it's I feel like we're. We're almost at this point we're almost randomly creating focus area uh because.

A

G

Existing buckets, don't work.

A

So the diff, well the buckets work for the ones that we have in the buckets. I think the distinction for this one is that these these are static properties. They are, they don't reflect change activity at all and that's the big difference and we can discuss this. I agree next time, and so I will put.

G

So when you start, when you start talking about project level, I'm just saying maybe it makes more sense to organize the focus areas by by spaces within the project rather than activities. For example, maybe change requests becomes the focus area issues becomes the focus area. Project level becomes the focus area uh rather than uh because, because we also have one that's inter inter dependencies, which is so it's the the we're handling the focus areas in two different ways.

G

So it's there's inconsistency in the way that we've created them uh and in the way that we're uh putting metrics in.

A

I see what you're saying, there's definitely inconsistency, and I think the question to debate next time is whether that inconsistency is organic to the problem or if it's something that we need to do something with.

G

So I think too many too many focus areas would be a bad thing for us. I that I believe that so I think it's yeah.

A

It makes work harder.

G

A

No, this would be some and we'd have to explain this differently in any sort of knowledge base, because these are different in very fundamental ways, but I'm gonna cut this meeting off. So we have a minute between this and the next meeting, and thank you all and talk with you soon, like in five minutes.