GitHub OCTO Speaker Series, 26 Aug 2020

Previous Meeting

⏯

youtube image

►

From YouTube: OCTO Speaker Series Episode 1: Dr. Denae Ford Robinson

Description

Golden Rules of Building Online Communities that Matter: Online communities for programmers, like Stack Overflow and GitHub, have norms that are not obvious nor inclusive to the 50 million programmers visiting monthly.

In this talk, Dr. Denae Ford Robinson, Senior Researcher - Microsoft Research will 1) offer a perspective of understanding what can inhibit participation, 2) describe what mechanisms developers are using to alleviate their own frustrations with communities, and 3) outline what on-platform mechanisms can intervene to better support engagement based on rigorous research.

Learn more about the OCTO Speaker series here: https://octo.github.com/

As always, feel free to leave us a comment below and don't forget to subscribe: http://bit.ly/subgithub

Thanks!

Connect with us.
Facebook: http://fb.com/github
Twitter: http://twitter.com/github
Google+: http://google.com/+github
LinkedIn: http://linkedin.com/company/github

About GitHub
GitHub is the best place to share code with friends, co-workers, classmates, and complete strangers. Millions of people use GitHub to build amazing things together. For more info, go to http://github.com

A

Hi everybody welcome to the very first in the octo speaker series, so glad you can be here today with us or if you're watching it. After the fact thank you for watching I'm jason warner, I'm the cto at github, and earlier this year we launched the office of the cto with the express charter to explore the unknown aspects of technology and specifically the future of software and software development.

A

So we've decided to make that the theme of this speaker series so we'll explore how software will be built or should be or could be built, but not with the time frame of tomorrow, but a year from now or even five years out and the speakers on these topics and in this series will be a mix of academia, industry and open source technologists.

A

We think it's very important that we get a mix of all of those.

A

So I'm looking forward to having you join us for more of these talks in the future, and we set up a bunch of github discussions for you all to correspond with each other on these topics.

A

Can't wait to hear from you so go find all the links on octo.github.com octo.github.com and with that I'm going to pass it along to my colleague uber to kick us off. Thank you.

B

Thanks jason hi everyone uh thanks for joining. My name is uber de moore, I'm the former ceo of summer, and I'm now running. Skunk works for jason in october.

B

A

B

We value our community above everything else. It's you, the people who collaborate on github to advance human progress that make github a force for good.

B

When we thought about creating this series of webinars, it ever seemed natural that the first talk should focus on that community and also explore how we can make the community even better.

B

It gives me immense pleasure to introduce dr dene ford robinson, who is one of the world's foremost authorities on this subject.

B

Deny is presently a senior researcher at microsoft she owns a doctorate in computer science at north carolina state university research identifies cognitive and social barriers to participation in online social, technical ecosystems like github and in particular, her most recent work dismantles such barriers to participation by taking advantage of identity deployment, peer parity and community mentorship.

B

uh I've always liked on our website. She sums up her research at the slogan targeted a few and helped the many so before we uh start the talk itself a few housekeeping remarks. If you have questions for dna, uh please use the q a feature of the go to web webinar app to post these. I will collect the questions throughout the session and at the end of dna, will answer as many of the questions live as possible.

B

So with that today, please take it away.

C

Awesome great, thank you so much for that. Marvelous introduction. Okay, so again, I'm danae ford, robinson, and today I'm going to talk to you all about the golden rules of building online communities. That matter specifically from identifying barriers to participation, all the way to designing interventions in the yahoo answers for programmers and beyond.

C

So any building communities that matter matters really determined by how people meaningfully engage in a community or if they do it all. So much of my research identifies how people engage in these online deaf communities and how they can bring their whole selves to their technical work and to get started. I want to make sure your wheels are turning in the right direction, so I want you to close your eyes for 10 seconds, and I want you to imagine what a software engineer looks like.

C

Okay, let's be honest, you're probably picturing, someone that looks a lot like this, but I would challenge you to shift your thinking to the emerging, diverse groups of people learning to code and thus entering the technical workforce.

C

There's a global emphasis on bringing diverse people in technology, but what about the resources to keep them there? So my work tackles step, one of understanding, the experiences of underrepresented, programmers and step two designing systems to better support. All programmers, essentially targeting the few to help the money and there's definitely quite a few who need help. um But my work focuses on these two large communities for software developers.

C

So that's stack overflow and github much of my work in the past, I started with stack overflow and slowly transitioned to github, so I'll talk to you about both but I'll start with stack the workflow.

C

So for those of you who may not be familiar, I mentioned stackoverflow being the yahoo answers for programmers. But I went to this drive home. That's essentially a q, a platform where this is one of the 20 million questions that exists on the platform and one of the 30 million answers and, as you can probably guess, from the millions of posts, there are a lot of people who contribute, but there's also quite a few who don't and to really drive home.

C

That example, I want to give you the perspective of this persona, aisha the programmer right. So, let's say aisha is a programmer, a developer at a startup in the bay area and she's working on this code snippet, and she encounters this air that she's not familiar with. So she constantly pays her air into bing because we're on brand um and lo and behold, one of the top hits is stock overflow.

C

So I you should find this question scrolls down to the answer and finds out well, this answer doesn't quite fit for what work for my solution. So she says you know I'm just gonna, I'm gonna write my own question and get a better answer um and she finally messes up enough courage to post and then something happens.

C

uh We want to figure out what happened to aisha and other developers like aisha, who decide to engage in communities, but then don't um one reason could be that she does not see other users who look like her post questions. So on stack overflow, you can see the avatar and the name of the individual who post the question and ask the question likewise on the in the community. You can see that there are modals that pop up that say to prevent thanks and me too, from our new users. um We protected this question.

C

We blocked it so she's not really familiar with the community expectations and what belongs in this community.

C

Also, aisha may think her question is a duplicate because it posts in the when she types her title. The list of other similar questions, pops up, and you see that there are also ones with negative scores, um but also low answers. So this can be quite discouraging taisha.

C

So these are many reasons uh why she may not be posting, but there's also a lot of people who are like her who experience similar challenges because it's like the world is watching them when they're posting and messing up and asking silly questions well, they're they're, not quite wrong uh on stock overflow, 50 million people visit the site monthly. um So you should see people are coming there to build careers and understand programming questions.

C

However, only 12.9 million of those folks have accounts so against a small percentage of people who actually actively engage in this community. Although many lurk, so this breakdown gets a little bit more discouraging when we're referring to uh different demographics.

C

So from the recent stock overflow developer survey from 2020 we're seeing less than nine percent women and non-binary people- and this is reasons that's important because in comparison to large tech companies, there's like our average around 20, so there's a star difference where you can be online in any one. You can be anyone you want to be. However, you're still discouraged, people are still discouraged.

C

Likewise, if the third of the users are engaged in this community, they have less than five years experience.

C

So this goes to show that the surge of interesting programming uh is not over yet there's quite a few people who are interested um and feel comfortable, who are actually quite a few people who feel comfortable, actually engaging technically and those who don't are missing out on a huge economic advantage. So there are job postings that, if you even have the presence on stack overflow, you can get paid 30 an hour thirty to a hundred dollars an hour. If you have twenty thousand points on stock, overflow or even own a popular open source product.

C

So again, there's just these opportunities that people are missing out on um and especially our underrepresented developers in these spaces. So the problem I'm really trying to drive home here is that novices and unrepresented groups encounter challenges when participating in online programming communities.

C

But what makes it really hard to be able to identify these challenges of these groups is that we want to do it without ostracizing these people without stereotyping them. So we want to be mindful of the methodologies we take when studying these groups, and today I'm going to be talking about different ways to empower and identify challenges. People have in these communities um by setting up these golden rules, we should focus on and how we're building communities and how we want to be able to sustain communities.

C

So the first one I talk about is identifying barriers later I'll talk about. You can provide legitimate, peripheral participation through mentorship and I'll. Explain what that means, um and then also we want to be intentional about how we use identity. If that's, how developers are evaluating each other.

C

So first I'll cover these two research questions um for this identifying barriers golden rule. So we wanted to figure out what barriers do women face with engaging on stock overflow and later how these barriers vary by gender? So we started with women because the squeaky wheel gets the oil women are women and underrepresented uh genders. Are the people who are facing challenges where we see that there are low uh representation in this community of these interviews, we did service expanding it to you beyond binary genders as well.

C

So we started off with semi-structured interviews from 22 women. We even had one of our participants identify uh being in the stack overflow top 100 of all time and they had a range of demographics.

C

Some were students, some were professionals, we had a big good bit of lurkers and some active users, and essentially we asked them about their personal issues of the community, how people communicate on the site some incentives that may encourage them to participate and- and what are some that may encourage others to participate I'll spare you the analysis, details but long story short. We conducted a thematic analysis on the transcripts on the interview transcripts identified, 327 barrier statements, which resulted in 14 barriers to save you time.

C

I won't go through all 14 barriers, but I'll cover three um and three of them actually resonate with a lot of us in the audience even outside of programming communities.

C

So one participant mentioned uh that I don't feel like my expertise, is enough for me to actually post an answer on stack overflow, that is, that will be of any help to anyone else and this we call this qualifications barrier. But this sounds like something we've heard in other offline communities as well in process intro right, not feeling like. I have enough to bring to the table to be here.

C

So we see it's interesting to see here that a lot of these challenges that we'll talk about that people face in online communities. They face them offline as well, but I guess, as researchers and scientists we expected, or some folks expected, that being online, having control of your identity can empower people, but it also can have disadvantages as well.

C

Another barrier was this onboarding hoops, so one participant went on to say I feel, like everyone else already knows what it is to post um and I want to stay away from that extra work, to figure out how to use it just figuring out what the etiquette is and all the little social things too are just kind of unspoken and kind of daunting.

C

So this sounds familiar. um We heard aisha's story earlier right. These unfamiliar noise people are facing and it can be quite discouraging. Finally, um this last barrier, which is a sphere of negative feedback participants mentioned. You know it's hard enough to ask for help, but then to ask for help and then get rude, help you're kind of like never mind you kind of gotta. Forget it, um and we uh call this.

C

This conflict fatigue right of just always being disappointed when you're asking for help, and you already had a must-have enough courage to do it and you have these scars right and how do you overcome them? Essentially, so participants really describe scenarios like that as well.

C

So, finally, after we identified these barriers that existed, we wanted to figure out how they varied across genders right. So for the purpose of this work. For our analysis, we included a binary gender of men and women, but I do want to note that these barriers did exist for people across the spectrum. We just use these binary genders for our stat analysis, so we conducted the wilcoxon ranks them, and I just wanted to itemize the barriers here and show you that these barriers are not dimorphisms.

C

It's not that all women said that the fear of negative feedback is a challenge for me. It's not that all men said the stranger discomfort were challenged with them. There is a mix across the spectrum for all these barriers, um and I just wanted to really highlight that these are challenges that exist, although we targeted the few we're finding that these challenges resonate with the many and can disempower that many.

C

um So our observations from this two observations I want to highlight are that there are two types of language that contributors experience, this unwelcoming atmosphere and this uh boys club this idol boys club chatter participants described it as they want to engage, but they don't really feel like they belong.

C

Likewise, people had helpful intentions where participants wanted to to contribute, but didn't feel like they can and based on how challenging it was for them to find opportunities to either post questions that would be valuable or even provide answers that people will find insightful.

C

Finally, I do want to highlight some of the amazing uh why it's important to kind of do this type of work, so jeff atwood, who is a founder of stock, overflow kind of acknowledged the work and how this can really help novices engage in the community. So, at the time they were working on a couple features about uh highlighting new contributors and highlighting the code of conduct.

C

um Likewise, what's been really helpful uh or something that I like to do with most of my research is blog about each of the papers that get published and the research that we're doing, because, let's be real, software developers are not going to read our long research papers. I mean most scientists, don't also read everyone else's features papers either, so we wanted to put it in a digestible format so that people can the developer specifically the community.

C

We're trying to target can give us feedback and tell us what type of interventions they want to see or how they resonated with some of these findings, and you can find the blog on my website there at blog.dnaport.me and you'll, see other developers commenting on content and research as well.

C

Okay, so next step next important rule is providing mentorship as in empowering people to engage. So I already covered a bit of this novice experience where I mentioned that questions are ill-received, where there's negative threes and zeros on question scores, but there's also these questions that don't feel like they follow community etiquette.

C

So from a blog that was shared, someone commented on facebook and said more often than not. My posts get down voted because I didn't know how to follow some arcane etiquette and again people are interested in computing or in contributing, but don't know how so another person commented and said: is there some trick to making it worthwhile to post on stack overflow and what's a way to find value?

C

So this uh begs us to ask this research question, which is: how can users engage novices to ask well receive questions, because we know there are people in the community finding value. We know that there are people trying to find value. Where do we bridge that gap? How do we bridge it so to understand that we're looking at stack overflow as a community of practice? Where learning is peripheral and people are learning by watching others? Do a task so comparable to see uh to have people engage on wikipedia.

C

We can check the edit history and see what people have edited and there's also features like this wikipedia tea house, which allows you to guide novices through that first experience and through how to contribute so for here we're looking through mutual engagement where newcomers can watch experts engage and learn from that to understand these design principles. More concretely, we designed this pilot on slack and again. This was in collaboration with stack overflow, so we worked with christina lustig jeffrey banks.

C

At the time, christine was one of the first researchers that stacked overflow and jeffrey banks was our developer who's, helping us build out these systems. um We conducted a pilot and we learned a lot.

C

A lot of things went right, but also some things went wrong, um but what we did wanted to know moving forward was that we found out how valuable it was to allow mistakes in a private space, so giving newcomers opportunity to ask this elite questions figure out the norms of how people engage in a space a safe space to make them like it's. Okay to mess up and to not get it right on the first try.

C

um Also in this space we found that it was important to provide formative and timely feedback. So again we didn't helpful for novices to get help to post a question and get an answer like two days later. No I've been posting the question. I need the answer or not right, um so we went to make sure they were getting feedback, similarly, how they get the feedback on stock exchange.

C

And, finally, we wanted to make sure you were adding value to have people already use the community so stock overflow there's already you post a question and you get an answer, but so we didn't want to do exactly that. We wanted to teach people how to do that, so teach them the fish, not fish for them. So we didn't want to answer the questions but help them to ask better questions.

C

So from this pilot we decided that our next implement this feature uh called collaborative editing on stack overflows platform and it had two core components. First, was this collaborative question draft? Where, here you can see that there is the same type of markdown text, you use to enter your title into the list of tags in the body of your question. Well, we just recreated that in one of the staff overflow chat rooms right, we just had a dedicated space. Well then, we call it a help room to discuss and edit those types of drafts.

C

um So it's kind of what I really want important. uh What I want to point out here is that we are using existing infrastructure right and building on top of that and explicitly calling it the safe space. um This help room a place to empower- and I think, that's important as we think about the new communities that we build in the future like how do we take advantage of existing infrastructure but reshape it for something intentional?

C

So to test out this intervention of how these collaborative drafts work? We conducted this just-in-time mentorship study where we had over 70 000 eligible novices.

C

um We had 63 mentors, the 70 000 ended up going down to 520 who actually entered the room and engaged with mentors, um but we learned a lot and we saw some interesting results from that help room. So over 30 days we saw that the mentored questions had a 50 score increase on average, so 50 higher score than the questions that were not mentored, so using stock overflows very complex, analytics uh combinations that, where negative scores are bad, anything that's negative is bad and anything that's positive scores or you can upvote.

C

The question uh is good. uh We saw less bad questions and more good questions, um so it was just really interesting again to see how the community perceives you. So the reason upvotes are important is and down votes are important because that's how the community receives your question, how they find value in your question, so we wanted to measure based on the metrics that currently exist.

C

uh One of the also interesting findings from this work is that we conducted a lot thematic analysis on the types of dialogue that mentors have with novices and we saw that mentor suggested some pretty high fidelity for improvements. um I won't go over all of them.

C

I just highlight this culture of asking, um so what mentors did suggest other things like formatting community triage, such as where the community, uh where this post belongs, which community this post belongs in and how to frame your question, how to add the right structure to it, but this culture of asking was quite interesting, so I'll highlight one quote from a participant or from a couple participants.

C

um So in the help room, one mentor actually says the novice. You know you also may want to edit out. Thank you at the end of your question.

C

I know it seems polite, but people on stack overflow object to it, and this is just a uh an image of another question that has thank you very much and also has a negative score. I'm not to say that the thank you very much is the reason it had a negative score.

C

But it's interesting to see that you do see questions like this on the platform um and then the mentor followed up and mentioned to the novice, um while it's good to be social, it's kind of just fluff on a q, a site and what was really powerful quote this powerful quote from the novice said. You know, okay, thanks for your help, and I hope this time people won't attack me.

C

um So it's interesting to see first, that you know people have these dialogues, these banters that happen in the comments that can be discouraging and sometimes we brush them off. But what we notice here is that this novice mentioned that you know they felt attacked. They felt targeted.

C

It felt a little bit more personal than it needed to be, um and it was interesting to have that document and figuring out like okay, we're adding value somewhere and now this is uh less than three questions are finding it valuable so looking forward to where that study goes next, so um one thing last final thing I want to highlight from the study is that mentors use identity as a signal in some way to empower each other.

C

So, at the end of one of our chats, we saw that a mentor after they helped the knox. With a question had said: hey, you know, I noticed your name. Are you from egypt by chance and it's funny, because I think christine and I were both in the help room at this time when it happened, and then the novice said you know yeah. Actually I am. I thought I picked up on your name too.

C

We thought it was going to be some type of code of content uh violation of privacy, but it was kind of interesting to see here how the mentor and an office the manager may have been selecting that novice to help based on their location, so stock exchange. Exactly you can see your country, your name as users describe them to be on the website. Self-Reported information so interesting.

C

It makes us think about the ways that identity can be an advantage and how people are using that as a signal to engage or a signal to choose who to help and who to empower so that they are not the only ones engaging.

C

I do want to highlight some of the impact from the just-in-time mentorship and how valuable some folks from the community saw so jay helen who, at the time at sac overflow, was the executive vice president of culture and experience really acknowledged how this mentorship, or this coaching experiment forged a new experience for novices and kind of resulted in this question. Asking template for novices um and figuring out testing out new interventions right. It really encouraged the spirit of um let's do a study, let's test out and see what works.

C

So julia social at the time was also a data scientist that stock overflow um really open up this program for other people to conduct other studies on stack overflows platform. So at the bottom of my slide, it's kind of hard to see there but I'll be sure to post these slides publicly, and you can see that there is a link as a call for papers, essentially a call for academic partnerships.

C

So I want to encourage all of your scientists listening as well to um reach out to these organizations and their studies, and I found that it was really helpful for us for the blogs that we shared um to connect with uh industry at that time, and I just want to make sure that everyone takes advantage of those opportunities to do impactful work and finally, we're currently talking to github discussions team to support newcomers.

C

So those who were tuning in into the the what was that satellite that just passed there were representations talking a lot about the new opportunities to empower newcomers and support maintainers and we're trying to figure out how some of these concepts of legitimate peripheral participation, which is just watching others and coaching others, do that. First task can help people feel more engaged and find value in their contributions.

C

So one of the last golden rules I wanted to touch on for the main meet of this talk uh was being intentional about identity, and this work was mostly done um on github. So I think a lot of people may be interested in this findings, so simply put pull request.

C

Excuse me all the decoupled process of submitting code to repository and having it reviewed, where the main points of the process we're going to focus on we're. Really. This idea of when you submit your pull request like you, submit your code for review and then when it's merged.

C

um However, there's quite a bit of my clarity in the decision making process that happens before is accepted um before people merge or close or before they accept or reject your pull requests oftentimes. The decision can be made behind the scenes and it's not explicit to the contributor or the project owner. um So I'm really trying to figure out explicitly what types of signals people are valuable or finding valuable and which ones we should be amplifying or which ones people are already amplifying and focusing on.

C

um Should we divert their attention so specifically using the context of github, where your profile image is very visible, um so on the github about your profile page in 2019, it highlighted that your profile is about your repositories, you're interested in the contributions you've made and the conversations you've had, but is that enough and is that all developers are using to interpret a story about a pulver request.

C

So this is an example of a profile page on on github. Are there refined details that people are using are the indicators of a popular repository, uh the contributions you made in the projects to cut a project that you've contributed to? Are those actually interesting to reviewers?

C

um Or is it the language of the repositories, the density of your contribution, heat map, or is it the context of the code that makes more sense so to understand more of this process of signals? I'm really trying to understand how programmers review and trust technical signals from each other, so we're in the study uh where, if, if for pull requests in the single commit of a pull request, if it was truly a meritocracy, then the only thing that should matter is the single commit the couple lines of code from the pull request.

C

But it may be valuable that the peripheral technical information is interesting too, which is what you see in blue here, but also social signals which are in green, um which are things that tell you the identity of the user, where they're from their popularity based on their repositories and followers and following may also be valuable.

C

But in fact we even see it may be valuable. Based on this experiment, we conducted where, before accepting or merging this pull request or deciding to accept their merge. This developer went through and looked at how to look back at the profile page to make a decision so I'll run that back one more time.

C

So there's the single commit the pull request. That's me: that's my my hair. In the background, they click on the profile page to go back through and look at the avatar image. The contribution heat map and the recent contribution activity before making a decision.

C

So I'm really interested in in the initial question, for this study really are: how do programmers review pull requests? Where do programmers think they look versus where they actually look um and what strategies they use? But I won't cover too much of the strategies they use. Yet I'm going to talk about that for some of the vision work, but I'm going to try to zoom through the technical details, but I think it's important that I show you um so from our sample.

C

We actually had the large samples over 42 participants and they all reviewed two uh profiles and two pull requests, but for the sample for this paper we actually used 20 participants and where we had uh over 11, I think we had about 11 uh pull requests that were reasonable to accept uh or unreasonable to accept and nine that were so, neither they should have been accepted and 11 that should have been declined.

C

However, we only saw 12 correct decisions so like 12, that were true accepts or true right, um and we also seen that there was also, though we have our code signals, which are the primary contribution developers should be looking at. There's. Also a comparison between how often they're, looking at code signals and some of the social signals, even though they make up less of the content on the page.

C

So I'll just show you this, but I don't need you to read this figure it's hard to read. uh What I do want to highlight again here is that the code signals which are ideally the primary technical contribution developers spent majority of their time looking there. Secondly, the technical signals, the peripheral, the peripheral information, but then, finally, the social, which is the smallest amount of entity, but a screen real estate, the smallest amount of area of interest, um but should also be the least important if things are in meritocracy.

C

But I did want to highlight this idea that, or at this point, that we saw a 5 out of 20 of our participants spent 48 to 62 percent of their time, fixating, as in their eye movements focusing on technical signals and above average time on social signals.

C

It's important to note this, because social signals again make up the least amount of areas of interest on the page and to figure out how developers actually found this valuable. I wanted to compare. I asked them, you know what signals did you find valuable and which ones did you actually look at from what I had from our eye movement data?

C

So this is the ranged uh stacked chart here, a dot plot here. um The orange here highlights the code signals, so we saw at 73 reported using the co-snippet to make a decision which, ideally should have been 100 right. um 45 percent uh noticed using the uh previous contributions, the other peripheral, technical information and honestly, only one person, only one participant mentioned using the avatar image, although all all participants did so.

C

The way to read this chart again is this: is the post survey, where participants list how many uh contributions and uh avatar or areas of interest they focus on, and then what we have here is the eye movements of how much people focus on. So, if there's a the way you read this is that, if there's a large distance between both items, that means the big discrepancy between what people reviewed and what they said.

C

They reviewed and we see, there's a big difference between how people are honest about the social signals they're reviewing versus what they actually reviewed, and this is not the call developers are liars. That's not the point here. The point was that there are social signals that are being used so does viewing. This content necessarily mean that this information influence the decision. Well we're using the I my hypotheses that, if you're looking at something that means it's you're tending to it now, we need to figure out to determine how much value it is.

C

But the major takeaway here is that people are using their social and human capital explicitly people are using the technical values that people have and bring to the platform, but also using the social ones as in what they look like their name, um the avatar image. So we want to. If people are using this, maybe we want to be strategic about how we deploy identity and how we give people control over that identity.

C

So there's a couple open questions here: um how do signals change based on the code attributes peripheral code attributes like if I change what country I'm from or the names of the projects? If the project I'm doing I submitted my pull request on is about games? Does my resume? Does my profile page show that I've already done a lot of other games? Does it matter?

C

Does that previous experience matter, um but we should really think about how we can build tools to better reflect contributions or or should we should we be directing developers attentions one way or the other, um but at the very least we know we should be able to be intentional about how we design systems to facilitate this process in open source communities.

C

This really business to ask the question um again. What does asking for help in the future look like, um and I really like to draw on this quote from bogdan vaselescu who's. Another researcher studying this space at carnegie mellon- and this quote he says, is the software- is being developed by increasingly distributed and increasingly diverse groups of individuals.

C

I would posit that how our programmers identity is presented will become increasingly influential. um I mean we have not evolved to be able to control our identity automatically our fact that communities have been emphasizing it. So I really want us to like really give ourselves some far out there, thoughts and space to be creative and think about different ways. We can evolve as a community, so as online communities evolve, we see you just having more variety and levels of privacy and control and how they share content.

C

But this is also evident when we're asking about questions or asking questions in a private team to help communities. So sometimes I'll. Ask private questions to my colleagues, because I feel more comfortable. um Think about people post on social media like on instagram and facebook. You have the ability to post to a close group of friends, um but with that being said, can we now enhance the control of how we ask questions or who we show our pull requests to to specialize audiences?

C

Can we take advantage of this anonymity right this segmented anonymity of like we're only going to show different parts of our profile page to certain people, because we feel comfortable with them? uh Should we do that? How? How can we do that? That's a better question. I think we move past should I think we know that having control over how you share yourself is important, um but further as our traditionally technical spaces, like github, become more social. So here you see that there's the uh you set your status and you can have like an icon.

C

I kind of mocked up that close friends thing, but I want us to think about how we share different parts of ourselves with others, but as our traditionally technical spaces, like github, become more social and as our traditionally social spaces like facebook become more technical. Here's someone asking for java sdk is on facebook right. How do we manage this context? Collapse across platforms does this allow us to bring more of our authentic selves to our technical work through the image of people pinning cute photos of their children to their profile page on github?

C

So I'm interested in studying how.

C

C

C

There was a study I did in 2018.

C

C

Hope you obviously hear me there's a study I did in 2018 about understanding, developer signals um and their identity, and we found that connecting individuals through an identifiable, marginalized identity.

C

Awesome, marginalized identity can enrich the experience of being the only one and thus encouraging empowerment. But how can we create that for marginalized identities, that shared only oneness was what people find valuable right, um but simply put if marginalized people are not engaging in the popular deep programming communities that we are engaging in then? Where are they seeking help?

C

Where are they going for help? More importantly, can we have our communities look more like that, so there are communities like ask questions.tech, which brands itself as an inclusive space, to ask technical questions.

C

Likewise, our studio has been intentional about encouraging developers to always be civil in their community and those python tutor tool which kind of highlights different countries and regions people are coming from.

C

um Can we make our communities can be intentional about the rules explicit beyond just the code of conduct uh and even empower developers beyond those communities? So another question here is: how can we better empower global south programmers, so programmers from these emerging regions to contribute to open source so to be specific, on grid hub they're, seeing a growth of developers from emerging contributing countries?

C

So this is from the october survey from a couple of october report from a couple years ago, um and what we're seeing is that the emergence, some of the places that are on this list, the regions is common for people in these regions to have mobile devices rather than a laptop, so with the emergence of github mobile and with opportunity to create even more devs for these from these regions.

C

Further in the stock overflow developer survey. From the past couple of years, visual studio code has been ranked as one of the top popular environment tools, with over 50 percent of respondents claiming to use it if majority of developers are using visual studio.

C

Does this create a unique opportunity to grow and encourage developers to from these emerging countries to be builders of these systems um to be contributors? Builders of platforms?

C

You know other unique opportunities to empower marginalized developers to build these extensions. Basically, but can we even be a little bit more explicit about that?

C

Can we intentionally craft opportunities to invite diverse perspectives to the table, be equitable, with the resources to build and then inclusively sustain participation or give the resources to the same participation as a contributor, I'm interested in seeing what this looks like then?

C

It's important to note that it's not the job of the marginalized populations to build these systems, but it is important to us as allies and as researchers, to advocate to foster spaces that can make this happen um and that as I'm starting to close here, I wanted to really highlight and when we think about how do we it's our jobs, our responsibility to treat others the way we want to be treated, which is the original golden rule right, but there's more there's more voter rules that matter and I'm actually kind of curious to hear your golden rules that matter if the communities that you've built, whether or not you're a developer or your online facebook group or your online technical community, of how you talk about your professional work.

C

um There are some rules like be welcoming, which I know the code newbies community that saron has developed is really intentional about this. um There's rules like making governance explicit so amy zhang who's at the university of washington and jay nathan. Matthias who's at cornell are looking at different governance models of how to empower developers um develop forms of legitimate peripheral participation.

C

I talked about mentorship, so if you're steinbacher who's at northern arizona, university and amita sarma who's at oregon state are looking at a lot of amazing ways of mentorship for newcomers and how to get newcomers started and their pathways.

C

So there's a lot of ways and rules that we should be considering for how we build our communities and some apply to our specific communities more than others. One of my golden rules, as a researcher and as a ally to figure out how to support developers in this space is to have amazing collaborators.

C

So here I want to give a big shout out to all my collaborators and even my more amazing interns and visitors, virtual visitors. We have this summer, who are working on amazing work to empower developers through open source software through social impact projects, social good, supporting developer communities for blind and low vision, programmers also for supporting developer identity through youtube, the machine, learning models and onboarding developers even in the offline space.

C

So to close, I want to do a short recap that online communities there are a lot of unfamiliar norms that developers and newcomers are not really familiar with, but there's interventions and we can design systems and we can build rules in place, have rules in place to empower people to join um such as just-in-time mentorship. What I talked about earlier, but as we think about the rules that we may already have set in place, what are the new ones? We can build up where?

C

How can we be intentional about identity, for example, as a rule and attention about editing, look different for every community? So I'm really curious to hear what you all have already started. um What are you thinking about, and I empower you all to do the work to make it happen so with that I'll close, and I thank you all for your.

C

B

Thanks very much today I'll say our tremendously thought-provoking uh talk um for all the participants. Are you uh do you have an opportunity to ask questions? uh As I said previously, please use the.

B

C

uh Oh, I think I lost your audio there. Oh good.

B

Decide to go to the rural sales community.

C

I'm sorry, I kind of lost your audio there. I just got you back.

B

Okay, wonderful! Thank you all right, so uh you mentioned these golden rules. uh How did you decide these were the rules that the community should follow? uh Where did the rules come yeah.

C

So yeah, so the rules really came from talking to people in the community and looking at literature right. So we know that in offline communities, uh offline technical communities is important to have empowerment for folks, so think about how a large tech organization or any tech company that there are these affinity groups to empower women or to power lgbtqa, plus people right and empower black developers, latinx developers, and we want to figure out how to build systems that support them.

C

So looking at literature from the offline spaces and talking to developers like having conversations um face-to-face hard right now, but zoom and go-to meeting are great opportunities and teams, because we're on brand uh are very helpful resources for that. So I would encourage community builders to talk to their users and receive that feedback.

B

Great, um so you also mentioned the mentorship mentorship program that you did with stack overflow.

B

What do you think could make that such a big success? Is it something we can replicate elsewhere.

C

Yeah, so I think what really made that really awesome. Our work with the stack overflow program is that we were explicit about who we wanted to help and we figured out which challenge people had in the beginning and figured out that from that work we were able to outline what were the specific things participants mentioned that were valuable in their experience and when they tried to post for the ones that didn't quite make it over the hump.

C

We also find it valuable to talk to those who did make it over the hump, so we were able to have perspective of the lurkers and the active users and be explicit and extract their responses and use their suggestions to figure out or interpret their suggestions when they weren't able to describe them in full context to build those systems so the mentorship program. It was an amazing collaboration with stack overflow and it really came from writing the blog right.

C

So wrote the paper jeff atwood tweeted about the work, um but when they tweet about the work- and they talk about it- they're not talking about the long paper right they're talking about the blog, which was helpful and digestible.

C

So I think it was helpful that developers were also able to receive and give feedback on the program because we asked them in the meta stack overflow community. We were like we're building the system. What should it look like? Getting feedback at all? Those stages was very helpful. I.

B

B

For a researcher to use media can.

C

Okay, I think you were saying the motivations for writing the papers or writing the writing. The blogs to supplement the papers correct.

B

Okay, yes, so why? Why are you using blogs? Lastly, uh as a way of communicating.

C

Yeah, so the the blogs, the intentions behind the blogs again, um they started off as a selfish reason for myself to figure out how I need to digest the work. How can I communicate the motivations of the work and make it impactful?

C

um Likewise, I wanted it to be an opportunity again to reach the community that I was trying to help and I think, as researchers and allies in these spaces we're trying to build systems for people, but if the people don't want them, if they don't find it valuable, if they don't see that your work is happening, then what's the point, um so that's just my perspective of how I like to do this type of work.

C

So, although all the developers in the world are not reading my blogs, um if they wanted to they're there, you know I can at least take an attempt to reach them, because I'm trying to help.

B

C

I'm sorry, I couldn't hear that last one is anyway, you can type it in.

C

B

You talk about the golden rules as they are today.

B

Do you think the rules will change and if so,.

C

Yes, so I think I think rules do change um and I think it's important that the rules also thing I did not know is that the rules change with the community size right if your developer community, if it's just your project, your open source project and there's only five of your buddies working on it, then the rules may be a little more implicit and if you just got started on the project into such a small group, then they're going to grow as your community grows as you attract more developers as you bring new perspectives from different regions.

C

um But I think it's important to at least have some type of rules set up. So when you're inviting people in they know that they're structured here, they know that they're they have mechanisms in place to make them feel safe. If something ever happens, and they don't feel supported, um because the goal again is to make everyone feel supported, um and even if it's just your buddies at the first five group. um Eventually, you might want to grow that project and you want to already have an environment there. That's welcoming for that.

C

So I definitely see rules changing over time for each project for each target. Demographic of people you're trying to bring to your project.

B

C

Yes, thank you for having me.