CHAOSS Evolution Working Group, 31 Mar 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: CHAOSS Evolution Working Group 3-31-21

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

By the way, uh so this is the evolution meeting for march 31st 2021. Is there anyone do we who uploads the video, the ut recordings to the youtube for those.

B

I do that, okay.

A

B

That and I just uploaded a ton, so I'm.

A

B

A

B

You're gonna get flooded.

A

That's just fine. I was just wondering it occurred to me sometime earlier this week that I wonder if I'm supposed to be uploading, those because I don't I've, never figured out how to upload things to our channel.

C

It's a it's a magic process, it just it happens automatically. I'm pretty sure.

B

Except when the magic breaks the person behind it is out for a bit so yeah, sorry about that they're all up there now or in the process of being their pending because they have they do a copyright check now as well on videos. So.

B

A

So well, let's start with issues. um Let's start at the bottom cycle, analytics.

C

A year ago, today, yeah all year ago, tomorrow, you're going.

C

A

uh All right, we do this.

B

B

These some things, if you just wait long enough, they just take care of themselves.

B

That's my personal strategy.

A

Yeah no, I mean, I think, sometimes that's true so code changes, ambiguity.

A

Yeah changes versus.

D

C

Nice yeah, uh although what is the.

A

uh What does jesus say? uh Jesus always argued for code changes, even though his tools and our tools all say commits. So I'm just going to skip past that at this point, since it's over a year old.

A

Percentage issues resulting in code changes.

C

So yeah I don't know so I don't, I think, connecting issues to code changes is is a bit much like you can connect. uh You can connect an issue to a pull request and you can connect uh and you can say a pull request- was successfully merged or closed, but I think it's a bit much to connect an issue with a code change.

C

C

Although I will say I did, uh I had a paper a year ago, where I looked at uh the merge success of pull requests that were linked to issues and the emerge. Success is higher.

C

A

A

I'm just saying basically yeah, you know you could do it, but it would require practices on the project and even with prs being connected to issues that still requires practices which are going to differ significantly across repositories. So the likelihood of that association being a metric of any consequences, extraordinarily low.

C

Yeah- and I should say the uh the case that I looked at when I did, that was a case where it was part. It was part of their design process to actually link issues to pull requests. So it just like stated that that's what you're supposed to do so line.

C

E

What exactly line impact is.

A

Yeah, I'm looking for.

A

A

It seems like an aspirational metric more than.

E

Anything but I'm I'm still trying to comprehend what what kind of aspiration is that.

A

I think the aspiration is, is the code, that's changed for the good or for the bad of the project.

C

This one yeah: how do you make that? How do you make that quality assessment.

A

Yeah they're very vague in this about how you might do that.

C

So I guess I would try to avoid the vague metrics right now and focus on the concrete.

C

C

Although they so, they tried to put it in.

A

Here's my stab at completing the metric.

C

E

C

The cognitive energy invested into each commitment made.

D

D

B

Also, did he say uh that they were trademarking.

A

Yeah they should write the name, so we would have to.

F

We would have to.

A

Just say we are curious, wb hurting how one really.

F

Accomplishes this major or a conflict.

A

An objective manner.

A

An objective, um consistent manner across thousands, tens of thousands, some tens and thousands ten, I mean really it's a hundred thousand.

D

Plus repositories at this point there's a lot of cognitive energy going into this comment.

A

Well, I mean it's uh and the answer is like.

C

Well, he's asking for chaos teamer users to assist in the effort of experimentally, confirming yeah visualizing.

A

A

uh So I had a powerpoint slide back in september.

G

A

My opinion is something like that would be a summary of several metrics about pull requests and issues. So.

B

Yeah, that seems.

B

A

Is there a way to transfer an issue to another.

A

E

Is there uh I created an issue on the common and it was transferred to the risk for revolution? I guess you did it. How I don't remember exactly my guess is copy and paste? No, it's the.

A

E

Yeah, the full issue was moved to another people. Let's look at these settings.

B

I found the article on the help, but you have to have admin permissions. I.

C

Do so sean does yeah? I think I.

F

C

F

A

Settings that's for the repository.

A

A

Let's find transfer issue yeah.

B

C

Oh I'm just gonna. I've never seen that.

E

Yep I've seen it previously, but I was not sure where that option is fun.

C

Is that I'm going to transfer that issue to auger I'm going to transfer all my issues to other people.

D

Excellent idea so.

C

I don't want to deal with it, I'm going to send it to common. I meant today so.

A

I I ended up actually on auger when I transferred that issue.

A

That's hilarious.

A

Metric line, it clears from radiometric line impact. We already looked at this.

C

A

Yeah, we did, I think, uh that's the one of the few that I left open branch life cycle. Did we release that branch life cycle.

C

No okay, no, we haven't, I don't think, we've specifically looked at branch. Anything we have we've had some long discussions about branch lifecycle. I mean for our release. Metrics. I don't. I don't think, there's any metric that uh looks at branches.

C

I could be wrong branch life cycle.

B

Yeah we did release it.

A

We spent quite a bit of time.

B

Yeah here I'll link it in the chat.

C

How long ago, oh, was that a is that a recent release.

B

Yeah, I think so.

A

Did we just release.

C

A

Did we release it.

B

Yeah, I put it in the chat yeah.

A

So it's a release metric.

C

Oh, did we just did we just release that this release yeah.

F

We would have okay.

A

That's why I don't remember it because it's new, I have to find, I remember, doing a lot of work on it. That's the only reason.

C

Yeah, I remember it- I don't remember doing any work on this, so I wonder if the was the work out of was the work out of band, or was it done? No.

F

A

Done, oh, it might have been done.

C

During the during my comps, when I was away.

A

Yeah, I think possibly that is the case.

E

Don't mind I'm waiting.

E

uh Yeah it's like every morning, I open the pc. First thing I do is like look at my email is the result in or not.

E

Please don't be.

E

A

This is more of a technical issue.

A

um Basically, the github changed its default branch to um yeah. You know for reasons that are reasonably obvious, so let's transfer that to a tool uh which brings me back to I mean they're really kind of brutal about like. If you transfer an issue you are out of the repo, you were doing issues in add new metrics.

A

One all right, so one pr yeah.

C

If one pr mapping to multiple issues, the pr granularity is too large, if one pr does not have issues, it is abnormal.

B

Yeah, I remember this.

A

Yeah I mean, I think this would be a useful statistic.

C

So this is, this is the one we we mentioned we mentioned earlier when we were talking uh about issues to uh pr success where uh the issue linked to a pr is a value. I I think, that's a, I think. That's a metric, however, taking it to uh to merge success is, uh is a bit too far, but just linking an issue to a pull request like whether or not that link exists or how many different links exist. I think that is definitely a metric.

E

F

C

Okay, it's in the metric spreadsheet too, apparently. Is it still.

A

A

I'm just going to say I think this is still idiosyncratic to each repository.

A

You know it could, because it requires an explicit action on the part of the pull request, at least either the person who creates the pull request or the person who merges the pull request to make that connection. There's no automated connection between a pull request and an.

C

Issue there there is there's, there's no uh there's a tool to do it, but.

A

It doesn't happen.

B

C

You can you can drop a link into a comment and that, and that link is then directly connects that pull request to the the.

B

A

If you do that, yeah.

C

So that would be the metric right. So is there basically the the metric is? Is there a hashtag issue, pull request number linked into a comment right right.

B

But yeah sean's point: if somebody doesn't manually do that, then the data is no good. There's no way to make that connection. So how do we account for that.

A

Right, well, I think we, it would be a metric with a caveat like this is very idiosyncratic right, and this is really this. One really belongs on the web development.

C

What is this one? This is about renaming md documents shouldn't yeah, uh just uh transfer. uh I don't think you need to transfer, because I already have this issue and I'm pretty sure I referenced this. So I just close it and say this is being dealt with in in uh the web in the web content group.

C

So I'm already.

B

F

C

Content group is the group, the team, so so you could say that yeah content group would be.

C

A

uh Standardized working group structure so every.

C

A

Group has this, I believe, yeah and.

C

A

C

E

Are we are we supposed to do it or is any intern doing it or like google somewhere for the student? I left a little. It felt.

C

A little ambiguous at the end of the meeting yesterday to be candid, so there is a poll there is a pull request by an intern in the pr, uh and it does mention this. I don't know what they did.

A

All right, I'm going to say.

B

A

Is currently being worked on at the chaos community.

A

B

Yeah we had a volunteer that was going through all of the working groups and doing that, um but.

D

B

Know about actually merging it in if that's something we were supposed to do or not. So that's a point of clarification.

C

Yeah ah yeah, so there's that there's the comment so mentioned this issue 11 hours ago, different folder for completed and underdeveloped under development metrics, and then I know, we've talked about in the past that the the only metrics that should be in this repo are completed. Metrics under development metrics should only exist in github or in google documents.

E

C

So so there will not be separate folders for completed metrics. We will be. We will be removing any metrics that haven't been released from the repository, so the the the evolution. The working group repository should only contain metrics that have been released, yep or are in the process of being released. Can someone paste a link to the spreadsheet?

C

A

C

Yep, I'm there now actually so.

A

That way, I don't have to leave this screen and go back to the spreadsheet.

A

Of course, I still managed to open the spreadsheet in a new window, so.

A

Okay, how did I end up back here.

A

Release notes, I think we're done with release notes right or do we leave these open forever? This is just.

A

Does this stay open forever? This is the new one. Oh okay,.

C

So this would stay yep. I created it to capture any new new notes for the next release.

C

Yeah yeah, that's probably that seems like it's a.

C

C

A

So interesting.

F

E

A

F

A

Is this that this is issue closed? um So I would say that uh this would be a tool-specific implementation problem, because we've defined closed issues fairly clearly right and it's up to the tool makers to ensure that that status is reflected. Can we look.

C

At the document, first of this release.

A

C

So the error is that that, wherever it says active issues, there should say closed issues.

B

A

Are these? Oh, is this yeah? It's just uh here's what I found excuse me. So it's just referring to a metro. Metric issue closed! It's just it's typos! Let's, uh let's just go! Look at.

A

Metrics closed, it was under closed issue issues closed that there was this one.

A

All right yeah, this is just.

F

A

A

It's closed across the.

A

A

Yeah, um what was the issue number.

A

A

A

All right so created a pull request.

A

There any objections to just merging it.

B

C

I'm confident in the merging, because it's linked to an issue.

A

Yeah, let's see linked issues. Interestingly,.

C

It wasn't uh oh so there I think there are. There are two ways you can link it. So if you you can link it in the comment and then you can also link it over here on the side, but but this is going to be github specific right, yeah.

F

C

A

I don't know so I.

F

A

Think I'd have if I'd have linked the issue before I merged the pull request.

A

um Fixed in full request number.

D

B

I thought it would automatically close that issue.

C

Only if it's linked only if it's linked uh on the on the right side, where you have linked pull requests, if you just mention it in the comment, it creates a link to a lesser degree where it gets.

A

C

In the pull request, but it won't close it automatically. Actually, it says link to pull requests that will.

A

Close this issue.

E

Twenty seconds.

A

C

E

Let me refresh this yeah no, so there is a link like milestone under my one thing,.

C

E

A

uh It's showing one link to.

E

A

All right, I'm just gonna, do yeah.

C

If for data collection, I think you you'd have to you'd actually actually grab that explicit link, but then you probably also have to scrape the comments yeah to pull out the hashtag number, and that would be the that's, probably the two-way for github. That's how you would that's how you would figure out if it's linked.

A

Yeah and uh yeah.

A

So um I'm going to say the chaos community decided that we're going to remove metrics and development and from the metrics folder and make them part of the spreadsheet.

C

uh Yeah metrics under development yeah, they should only exist in the spreadsheet and in google docs. Only yeah only released metrics or metrics that are going through the review process for release should exist in the repository.

B

Sorry to interject, um but I just did a quick google search if you type in fixes, then hashtag the issue that will automatically close it.

D

A

D

A

To know on github.

C

I wonder how many how many things I don't know about github. I bet you. I bet you it's 99 of uh what's possible.

A

So all right and so then we're just kind of waiting on the the stan. I mean, I guess we're. We have to go through and clean up the repository at some point and then add new metrics label, sages tags associated spr.

C

Open 12 minutes ago.

A

um This, let's see I mean we do this in augur.

A

So we could create a metric for that.

A

uh And we may have a metric for that- let's not even kid ourselves.

A

Tag, no so pull request. Tags are basically we've done. Some analysis of this in the risk working group. There's every there's a lot of tags and each repository has its own nomenclature for what they take things as, although there are some ones that are pretty common and.

C

That's a big, that's a big problem, though, because you can't you can't really compare because they all have their own tags well in the in the.

A

Case of um this bag of words, it's it's like a bag of words like in auger, like kate stewart, and I collected a set of tags for 50 projects. um There were there were common tags across projects.

C

A

Feature bug yeah, so it's it's really a question of like we could probably get some amazon turkers to identify, synonyms and classify the tags um as as being in the same category. So um auger does this already and you know creating a metric around it is I mean it's. I think it's a worthwhile.

C

Endeavor, this would be kind of tagged, it'd, be kind of a thematic analysis of tags right. So you have these.

A

C

Tags schematic in.

A

Classification, I think so some you're right some tags are thematic in nature, uh but, like bug, fixes features, front, end features, back-end features, api features.

C

Others that might be an interesting paper.

E

Yes, I have written an uh intro on that. I was thinking of proposing it uh sean like because I, when I was joining the risk working group, I saw you and kept doing this and I was like this is a you are collecting entire data you're doing everything is just we can like make a paper out of it, so I've added, I have already written an intro. I might send it to you guys um once a look at it.

A

Labels, so labels are basically tags. I can label an issue. I can label a pull request. This is an interesting one because it applies to both issues and pull requests right um code development process. That way we keep it out of issues and pull requests.

F

A

A

Hello process quality: it's not quality efficiency.

A

It's property efficiency.

A

I think it's process quality, because it's creating uh metadata.

A

It so and it's creating metadata in the pipeline from issues to pull requests. So it's not directly so quality or efficiency. It could.

E

Be efficiency too, it could be efficiency. You can directly point out by looking at the label and go to that particular thing.

C

I think I do I prefer process quality, though I think you, I think you were right the first time all right. So this one.

F

All right, I gotta fix this.

A

Because it doesn't.

A

A

A

Sorry, it was like everything was gonna break, and so I think uh another so that you said, if you say, quality or efficiency at the end, kevin.

C

I think for that right or the one we were talking about, not that.

A

Yeah yeah so for for this issue, no, not that issue.

A

This issue new matrices.

A

uh This one we would, you want under quality or.

C

So it's process right, so it's development process, mm-hmm.

A

C

Is so process quality? Yes yeah? I think so.

A

So duplicate row copy.

F

I don't think we have a name for it. Change, request, issue.

A

And then there's an issue that is directly connected to it. Is it change request and issue tagging? I think so. Let me see what it says: issue yeah literally they say issues in prs and I think I think that's valid because you can take both of them.

A

um And this is not in progress. I would call this.

C

A

C

Do you do you want to put the the and in there explicitly uh what did I do, change request and issue tagging change, requests.

F

And issues can both be tagged. Oh in the name.

A

Oh okay, what do I have in the name? Change request? Oh yeah. Yes, that's that would be good to have that and in there.

A

That's a good one to consider.

B

A

Gayward all right: we've, uh okay,.

F

A

Recorded this as an issue being considered.

A

A

A

A

A

A

All right pull requests. We are only five minutes, uh georg.

B

Has a comment too in the chat about the.

A

B

G

um Yeah, whenever we're ready.

A

Yeah, I think I don't think we're going to get uh we've made some we cleared out our issue queue and yeah.

C

Next meeting we can clear out our pull request: queue.

A

Yeah so yeah. Let's talk about that georg.

G

So, as we discussed in the weekly meeting, we have a proposal for standardizing what our working group repositories look like in terms of where to find certain documents, and we have seen google some of code students starting to submit pull requests to modify our working group repositories in line with the guidelines that we discussed some months ago.

G

And I just wanted to give us all as a working group here, a heads up that this might be coming and we might be seeing a pull request and to reflect on what that would look like and what we expect from from this change.

G

And so I think the biggest change for the evolution working group is going to be moving.

B

G

From the metrics folder into the focus areas, folder.

C

And also deleting metrics that haven't been released.

G

A

So when we put in the focus areas folder um are we putting them? I guess they'll be just move the metrics folder here and then unreleased metrics will be removed either moved to the spreadsheet or killed. I think most of the unreleased metrics in the evolution working group are early versions of the metrics that were renamed and they know they're they're very similar to things that we've released and they're that I think I think carter did a good job, reducing that clutter, but I don't think the reduction's complete.

A

I know we were reticent to delete old versions of commits and things like that, just because there'd been so much discussion around it. So some of these deletions won't move to a spreadsheet that will just go away.

E

So maybe create a folder and say old discussion or like something in that and move it over.

C

Here, no, I think the I think the point is to to just get rid of it completely, so that we remove confusion.

A

So what I'll do is what I suggest is we create a branch off of our current master, called repo structure, standardization, and that that way it preserves the history of all these metrics somewhere and then working in that repository. We delete things that are not released.

A

G

So create an archive of the current version in a branch that we preserve for the future. And then, when you.

A

Preserve really preserve for the historical record, because yeah not for the future, the the evolution working group went through. It was a very active group and there was a good deal of debate at the very beginning, intense debate.

B

A

And I don't- and I think some of that intense debate is reflected in the metrics- that we ended up not moving out and I don't want to lose that like we've moved on from it, but I think I think, having that history available uh is helpful.

C

A

Our process for defining.

C

Or coming up with metrics now.

G

C

Not include going through this list of metrics and saying hey, we should do this. One.

A

It also shows her. I think it also shows respect for the work that jesus did early on to to try to standardize it in in the way that that we envisioned initially- and I don't want to just wipe that out- like this- isn't code where wiping out old code evaporates. I think this is- I mean for me, it's more like an act of respect.

A

um Then it is a um like a need to preserve like to need to be a digital pack rat, um because there is a strong history here.

G

Yes and I I recognize the history and when we go through the through the gitlock, all of that history is there and will never go away and that's.

A

G

A branch is one way of preserving this another that we have is through the release tags. So, if you click on the link, I shared in the chat for every release.

B

G

um The um a release tag- and I see that back at the in 2020 we or you sean, actually created a tag for hey before folders clean out. We can just create another tag to tag hey.

G

We are now changing the structure. Here is what it looked like before, and we can always refer.

A

A

That's a good point um that we've done okay, so we've done it this way before, let's.

A

What did I call that.

G

Branch- and I I think, keeping it in the same gitlock history is cleaner than branching it off into a different branch, and then we might confuse where work is happening or whatnot.

A

Yep, I just deleted the branch.

A

I think I think that's a that's a that's a good idea, um so yeah we'll just do a release after cleanup.

G

So, okay, any questions about the new proposed structure that the working group would be getting.

A

I think, if there's a document it will help us like. If there's a document, I can go through and clean it up between now and two weeks and documents just posted.

G

This is what we've been discussing.

C

E

The issue we closed just for I don't think we have closed it. We have left uh kept it open right.

C

Oh yeah, we closed that one yeah sure yep yeah. Should we just.

A

Reopen that.

C

One just for um I well, it vaguely alluded to um standardized working group repository structure.

A

This is a good. I think this, I think, putting a link to the document that york shared okay as a new issue.

A

D

G

A

G

If the one of the students had already made a list of here are the specific changes that would need to happen, I don't know where that is currently, but I expect once we get the pr we'll know anyway,.

A

Yep, that sounds good. um I will sit. uh We are five minutes past. So, if that student wants to look, we have an issue for it and uh thanks for a productive meeting cleaning up some some of our old issue work and I will catch you all later.