CHAOSS CHAOSScon EU 2019, 25 Feb 2019

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Finding the Order in the Chaos(s): Are We There Yet? | Ildiko Vancsa | CHAOSSConEU2019

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

I well, I may need to apologize. I rather have questions in this presentation, as opposed to answers. However, I think I saw in the schedule that there will be another presentation, but mainly answers and not questions, so I believe that we should complement each other pretty well. So let me just dive in and if you have further questions to add to my set feel free to do so, so our community has more things to work with.

A

Well, my name is Neela. Comanche I worked for the OpenStack foundation and, basically that's how much I wanted to spend on the work stuff when it comes to introductions. However, if anyone is interested in what OpenStack is and what the OpenStack foundation is doing in way more details, then I'm happy to chat about that. By the way is there anyone who never heard about OpenStack slash the OpenStack foundation in the room, okay, I think I, think I'm good.

A

So a little bit of background I'm, one of the board members of the chaos community and I'm, partially fully kind of the little bit of the sceptical and pessimistic one I like to combine this as the realistic person in the room. So I always like to ask back that.

A

Okay, if we do this here and if we define this, will they do more harm than good for us, so just basically trying to make sure that we always observe things from different angles, just to make sure that when we are moving, then we are actually really moving forward and we are sure in what we are doing when it comes to like things as pessimism. Is there more hungarians in the room or just me all right, then? Well, it's it's in our blood. I I swear. It's not just me.

A

So most of us are like that. We we are a little bit of on the on the pessimistic side in that little Valley somewhere in Europe still so yes, this is us, so that's that's partially I guess how I got to this I love metrics, because we all do like to know. What's happening around us, we all like to try to make sense of it and have an idea whether it's good or bad, whether we need to take actions or whether things are going well.

A

However, I think in my experience, it's just so easy can I get two quick and not-so-good conclusions from numbers and metrics or use them in a well, not too good or effective or useful way, and that's kind of the hate part in my life, like just saying an example like I love, my mom, but she has like a 4 kilogram or ore, which is I, think translates to 8 pounds or something range which between the oh, my god, you're so thin, and the oh you put on some weight.

A

You know like that's, that's really not a big range right and yeah. If we just get back to metrics like you get, you get to know how stressful my life is, but yeah trying to stay healthy and so forth. A little more background on on this thing on, like metrics and life, and just like how subjective I think data can be because we I think we all just look at it that ok. This is a data set.

A

These are numbers other data, so it must be objective right, but it always depends on where you look at it from what the context is and whether or not you're lazy or not lazy, to actually try to put it in context. So, like I, went to university computer science, I got a job in IT. I travel a lot, so I find a way to wait too much, but still have a reasonable salary know all the amazing people, for example, in this room.

A

So you know that must have been the good choice. Right I mean everything seems to click again. If you ask my mom, she will always tell you that strongly believes that I should have pursued art because I love drawing on paper, not on computers. Computers are great, but not for drawing, not in my opinion, but though I may have just you know, starved to death already or would be starving somewhere, but maybe I will still be happier, I don't know, and the numbers will I think never tell me.

A

The other example is I, have a rather lowish blood pressure, so it's 90 over 60 when I have a lucky day, sometimes go down to 80 over 50. That's not comfortable, but still it doesn't mean that I'm, unhealthy I, don't think so it's out of range. But you know it's context, it's it's. How I am it's? How I am designed yeah? Maybe there are some flaws in that design process. I made my complaints to my parents, but you know this is where we are. So whenever we do this collecting data put it into buckets.

A

We always have to look at the bigger picture than just the data and the bucket itself, in my view, but again feel free to agree or disagree with me and well. I had a corporate life before I joined the open-source side, we still can be called corporate by working for a nonprofit organization at this point, who's working for for-profit organization in the room.

A

Quite some anyone agrees with this or I.

A

Would like to know that everything that meteor said still stands. Let's say that we are talking about another department.

A

But I have some experience from my working for profit organizations that we kind of just went a little like overboard with this whole metrics thing like your performance report and just trying to figure out how to measure yourself and okay you're contributing to open source. So let's say that if you have I, don't know ten merged comets in the next half year, life will be cool mm-hmm, whatever, like sure we can come up with a number. But what will this be good for I? Don't think we ever actually defined that part or when?

A

Sometimes we wonder my people are doing things better. It's because this is what their bonus is depending on or whether this is something that is actually useful, feel free to blame me for everything. But there are examples to this question popping up and it is unhealthy and it is still data driven. So you know it must be objective. It must be good. We have all that data, so it should be all right, but it's still not because it still matters how you use the data and what are you getting getting out of it?

A

It always depends on on the context and there's some further unhealthy behavior, that I will just get to later so yep. Just from snack from yesterday there you go I love, my colleague and I've loved, all my team, and we all love data. So again what I? What I said before it doesn't mean that I, hate data and I think that we should just live in the dark and never measure anything just go by your instinct and what you feel. That is not the point.

A

The point is always to look around check your environment and and see that which parts you think you would need more data, because it seems to be not working and keep an eye on the base levels and make sure that whenever you're collecting something, you always look into the context and you always analyze it thoroughly, as opposed to just getting into conclusions of things.

A

So yeah we love data, but can we have too much data like what if we get too obsessed with data, and the only thing that we are doing is having you know, lots of it and just dig into trying to see just from the data itself or there's something wrong, whether we should be continuing something or stop doing it is it? Is it always good to like overload ourselves with data?

A

Like an example here, I have my lovely colleague Kendall here and we are running a training like how to contribute to OpenStack some parts of it also relevant. You have to contribute to open source, and sometimes it comes up whether the training is efficient enough I'll be teaching the right things. Is it working? Well so yeah? We can. You know we have the names of people who attended the training. We can look into whether they contributed to any of, for example, the OpenStack Foundation project.

A

What if they didn't? Does that mean that the training it's useless? Does that mean that we teach the wrong things or just? Does that mean that maybe the person or the people in that group? They are just not the developers who will do the direct contribution, and so from which point the training is valuable and from which point we say that.

A

Ok, we really need to change something because, in my view like, if we just have let's say again, just a random number one champion per year or something when I mean champion the person in the company who goes back and says that, oh, my god, I just understood how open-source works, how we should do things and basically change the culture and mindset in their department, poor company.

A

So maybe you get champions out of a training like this. How can you measure it? Can you can you always? Will you always know we'll? Always the really information always get back to you like in a reasonable timeframe to say that oh wow? So this is one outcome of this training. It is extremely useful.

A

But how do you know and what, if you like, stop continuing the training, because you don't see this aspect and the contribution numbers are not really there that you've disfigured to put up there and you stop doing in and you will not get more champions that way. Isn't it bad I think it is but again something that can be just directly objectively, measured just by looking at data.

A

Just let me check where I am bit time before I have 20 minutes, oh okay, that it's all good, so unhealthy, behaviors gaming, we have numbers, we have dashboards out there and people just love to game it for different reasons like when it comes to corporate environments.

A

Sometimes you happen to have a manager and they see like different contribution numbers like reviews, code comments and these sort of things so and they are like. Oh okay, our company should be in the top I, don't know X and then okay, people. We need to achieve this by the end of this month, so go and do reviews I, don't care how you do it, but you need to have 10 reviews per day and what comes from this random plus ones, other random, minus ones?

A

Oh, the the automated testing system failed, minus one like yeah I can see it myself. You don't have to comment on the review for me to see it or just like you know, go and give Plus Ones or like thumbs up to various reviews.

A

What's the value in that, basically, nothing like. Sometimes people say that, yes, we had to practice how the tool works by pushing the same button ten times every single day. I, don't think so, and and then you get to the point of yes, wasn't really our idea and we don't we don't. Even 100 per hundred person know why we are doing this and I'm not saying this to to say that individuals are bad or the manager is a bad person.

A

They are rather maybe a little ignorant and, on the other hand, we're all humans and we tend to be lazy and how easy it is to say that, yes, we are in the top 20 contributing organization based on I, don't know number of proposed changes and like 90 percent, is one character, type of fix, which is you know if, if you are just looking at the dashboard as an external person, just an individual who's interested in it yeah there are numbers you can see like ranking if you're interested in it.

A

I personally was never too interested in, like overall corporate rankings. I, usually always look at the data for myself to see that you know like. Is it visible that I'm doing something or I should do something else, but like so you have this dashboard up, you look at it. It doesn't cost too much time like cost you a few minutes.

A

You go through that data, but just think about the other side, like those people who are caring about the project and maintaining it and has to deal with, like all those patches that really just one character, typo fix.

A

Maybe it's even an automated system that does it you never know, so it actually has a lot of impact on a community and I assume that that it's mainly not necessarily happening within a corporate organization like when you look into internal software development and those metrics, but when it comes to open source and numbers that are basically visible to everyone, when people just jump onto like consequences easily, this can be really harmful because you basically Rob all those maintainer of their time.

A

Because of looking into all these things and I'm, not saying that typo fixes are not valuable contributions but like go through the whole thing, and you know upload one or two patches depending on the size and how many typos you have to fix. That is something that's that's valuable, as opposed to just yep jump on the numbers and and gain them because yeah that must be. That must be the good way of getting recognition in a community. No, that's really!

A

Not, but again we have those numbers out there and can we like come up with metrics that cannot be gained I, don't think so. But again, it is a question.

A

Should we work on metrics, that's harder to game, or should we spend more time on like educating and just making sure that whenever people are looking into whatever numbers that we are visualizing and and giving access to that they that they have an easier time to put it into context or or they have motivation, to put it into context like telling them that? Okay, if this number goes up too quick, then it might actually points to, in some cases rather bad behavior than good.

A

Can we do that rather open questions and again, if we can do it like? Can the chaos community work on this and help with that? Should we try to reach out to organizations and make sure that they are all basically with us on this I hope so, but again open question for for the community to to further further think about and yep metrics and open source ecosystem?

A

So personally, very personally I'm, no idea at this point I mean we talk a lot about metrics but I think most in most cases those metrics are like reviewed and looked at by companies, for example, so I barely see like random individuals. You know in an open-source community and I- don't mean chaos here, because our goal is to figure out this whole metric thing, but in any other life software development community.

A

We are not really chewing on these numbers, like the developers are just sitting there and they are trying to like write the coolest code or the coolest feature, and sometimes the coolest documentation too. We all know that there's a lot of improvement that needs to happen there, but so how does it work in an open-source ecosystem?

A

What's the difference when, when it comes to metrics, like the same thing that we are looking at from more of a corporate perspective, how is it different when we are looking at it from from the open source ecosystems perspective I right now, don't necessarily have an answer to that I'm working for a nonprofit organization.

A

We are trying our best to support open source communities and just help them to like build and sustain a healthy and balanced ecosystem, whatever healthy means and even balanced, but something that seems to be working, and we don't really have the answers at this point to how we can help with metrics to steer things to to the right direction. We are doing analyzes reports looking into numbers, but these are not always the same metrics.

A

These are not always the same questions that we asked, for example, at the end of a release cycle, so we don't have the right answers yet and again, just just the the interesting point of how data can look different when you look at it from different points of view, I don't have any issues with github that start with that. Octocat is also awesome. I love octocat.

A

So the point really is tools versus metrics. Like me, I mentioned github. You have projects they're pretty great, but that's not the only platform, and sometimes you just get like the first quote. So if your project is not on github, it's not even open source man but like OpenStack is not on github I mean to be mirrored there, because we kind of liked the interface to browse code and then and then you end up it. Okay, but you don't have enough stars there.

A

What's up with the issue, handling and poor requests, we don't use any of those and OpenStack itself what 20 30 million lines of code, thousands of people contributing to it very openly, and we are really proud of our for opens open design, open development, open source and open community, and we are really embracing that so I don't know what else we could do to just show and prove that it is in fact open source and it is in fact, is not developed on github. So it's not github fault.

A

It's more of a question of we have a tool out there, which is really really popular, which is great. So it is really easy to look at those statistics that github can show you like the github stars or just when you talk about contributions, they always say pull requests and issues, because github is a veil known tool, but it doesn't mean that that all these terms and and metrics that github and data that github can give.

A

You is the only thing to look at and to measure the whole ecosystem based on what one particular tool gives to you. Even if there is a really famous and popular tool, it is a tool and it is a tool to design which was designed for collaboration and not to define like the the base set of metrics, of how you measure open-source contribution. So again, what can we do because we can't just walk around with like signs that github stats are just one way of looking at things and they are not the only ones.

A

So education is still one one part, but but how can we like work together on maybe change some perceptions educate people about how this whole ecosystem is built up, and maybe just saying that tools are not everything. Tools are giving us a lot of help with collaborating doing doing the things developing the things that we we believe in. We love work together, and this is what we use tools like github, for so how we integrate data, that's coming from github interview that is able to incorporate all other tools and processes and ways of working.

A

How do we do that? Do we do that already today? Maybe I'm the only one who's experiencing this like this little obsession with github stats themselves- I, don't know, do you have concerns like this, or maybe our concerns.

A

Questions like like this that how to how to deal with this yeah I can see people nodding around so another thing that we can think about how to how to address as a community who's trying to help the ecosystem to come up with ways to figure out how we are just moving to the right direction or we need to steer things around.

A

So we have this amazing community and I figure that I will put up the mission that what we are trying to do here. So there are a lot of questions out there. I. Don't personally, think that we would have all the answers.

A

But again there are a couple of ways to tackle this and the communities working on both from metrics perspective on what to look into how to look into how to put it into context, but also looking into how to show examples to the ecosystem, how to provide some tooling around how to provide ways to exchange data. So the mission is basically a little high-level on where we are trying to get to just to figure out how to get a better view on what we might consider healthy.

A

What we might consider an active and balanced environment when it comes to open source projects and the whole open source ecosystem and what ways we can probably do that. Just a quick note on the mission as well as the goals. So we are a community continuously and constantly evaluating ourselves as well.

A

So I believe it was like a couple months ago when we had a discussion on the mailing list about the mission and the goals and whether it's still like, covering and in line with what we identified that we would like to do and achieve so. It is always something that everyone is free to look into and raise a hand that hey I, think we are doing a few things.

A

That's not captured here, and it would be important to do so to make sure that it's visible for everybody on what we are trying to achieve and also that we can find partners for that work. This is how the goals look like basically again, work on metrics work on software work on things that that's reproducible and create like an internal loop, so work on metrics and software all together, so getting there.

A

So I will not spend a lot of time here, but, as you can see, we are trying to make sure that whatever we do, you can you can try and test for yourself and see how the data is getting collected, visualized and all together, trying to make sense of it and all this open source available and for collaboration.

A

So metrics part, this slide is more of a history. I believe I actually left it in here from my presentation a year ago. So this is where we started from defined a few groups of metrics that we thought that would be valuable to start with, just to again get to the point of what is healthy. What are the stages of our communities and projects life cycle? How we can measure what's happening, how we can again make sense of it identify dependencies, something that is really important, looking into risks and diversity and inclusion. I, don't think.

A

There's a way to leave that one out so where the community has got since then is we are organizing our work into working groups, basically looking into areas with focused groups and defining a process with which we are discussing and designing and developing those metrics and again making sure that it all boils down to the software part. So, there's a little of your organization going on within the community and the two currently operating working groups are growth, maturity and decline is one of them and the process is on this slide.

A

Basically, defining focus areas see what our goals are: identify, questions to reach those goals and and define metrics, to answer our questions and what I repeated at least I don't know 10 15 times in the past, 25 minutes use cases put it into context. Give an example. So this is the metric. This is the data from this data source. So how you interpret this data and how the metric is changing, give a use case.

A

The use case will tell you that why you're looking into that particular set of data- and it will also guide you how to evaluate based on the context. So this is really important, and so these two parts are kind of going hand-in-hand and you can find the github repo for for the working group on the slide and I believe Georg already uploaded the slide or will upload it to the website.

A

So you will find this and and the focus areas if this is something that you would like to join and participate in I just put up a diagram that only shows like lines of code- and you know guess, activities based on that again. Nice set of numbers, I I, don't think that it tells anything about about the relationship of like, for example, Facebook and Android, nothing to do with each other different numbers. So I, don't think that we would like to rank that Facebook is way cooler, because it has way more lines of code.

A

Then Android, no Android is a small thing. It has to fit on your mobile phone, so I really hope that they will not get up to like 61 million lines of code, so just an example again always put it into context, while you're looking into it, why it might look like how it does and look into wilcos context background big picture diversity and inclusion.

A

I, don't have a lot of time to to where to talk about this, so but it's the interesting part is that I was looking for like one infographic or or something to to put up on the slide just to basically follow my concept of making it a bit more colorful than how my slides usually look like and then I found that puzzle like image and I, think it gives a really good view that, while I think three four years ago, when I got to my point of open source involvement, that I started to see other things than code and documentation contributions, so we were really really focused on gender diversity.

A

But that's you know. That's like this small piece of the puzzle. There are way more items here to consider and there are three dots, because I'm sure that there are aspects that we are not necessarily considering today. Maybe we need to. Maybe we don't I, don't know.

A

Maybe, like a few years later, we will have aliens on this planet and then we will have like a way bigger diagram to to consider I, don't know, but again, just a lot of aspects to to look into as we progress with all this activity and I mean just jump to the next I, just to make sure that I finish in time. So the diversity and inclusion working group is basically operating with the same process and principle as the growth maturity decline. Working group.

A

You can see the focus areas here and I just wanted to highlight a few things like recognition of good work that doesn't necessarily have to do anything with the diversity part so like, for example, I grew up in an environment. I've, never really learned to like give positive feedback. I also have no idea how to receive it, because I didn't like as I said: I love my mom, but she always tells me when she thinks something is off and because this is this is just our like way of.

A

We all grew up kind of a tradition that, if everything is fine, then I'm not saying anything. If something is wrong, then I will flag it to you that you need to. You know, look into that and everything else is fine, because I didn't say anything. But again, if you act like this in a like multicultural environment, that's not good, and just in general, like telling people that hey I, think what you just did is pretty amazing. That is great. It shows it feels great for the person who you told it to.

A

They will know that that that what they did was valuable to at least somebody, and it will also feel good for you. But this is something that some of us need to need to learn and well. I put off the put up the photo of the dude there, because when, when we talk about diversity, sometimes I remember hearing a heated discussion where someone just like was attacking the middle-aged white dude in the room when he tried to be an ally to women in the in the room. It's like.

A

Oh, you can't be a good community manager because you're the middle-aged white guy. What does it have to do with it like? Really it's? It's not how it works. So again, when we come when we talk about diversity and inclusion, it really is about all of us like not putting anybody into a bucket, not even the middle-aged white dudes, which kind of became a stereotype in I.t yep.

A

We have a lot of them great they're, great people, and they are just people like everybody else and saying that we are talking about diversity and inclusion and just figuring out how open we are and how we can help just people in general to join activities and be recognizable and so forth. That is really really important, and the the other diagram is just to remind myself to highlight that when we talk about diversity, that's usually sensitive data.

A

So whenever you are trying to figure these things out, just always think why you're trying to get data and what you're trying to do with it. And if you have no idea just don't do it. It is sensitive data. It is not a game so again think twice, because it is a really important area. It just helps kind of humanity in general.

A

So let's try to do it the right way and we have software as I mentioned I, believe I'm running out of time and I'm also not super involved in the software side of things but again linked here, and there are a few more links on this slide and the very important care of that community link. You will find pointers from there to all the other amazing work. That's ongoing good thing is that the there will be a presentation, the software parts in the afternoon, as well as two tutorial sessions for the two working groups.

A

So you will be able to learn more details and how they operate and kind of try it out yourself how they work. So please stay and visit them and be active, and thank you.