1. 5

  2. 10

    To be honest, I’m curious about the demographic breakdown of the applicants for speaking at the conference.With 3%, the demographic could have just not at all applied in the first place, making it extremely difficult to select it.

    1. 15

      I thought I remember hearing that it was a double-blind selection process – which is generally considered a Good Thing (tm). The problem is that when you have only a small number of boxes to fill, it’s easy for the margin to be swept out in a double-blind system. I don’t necessarily think that there’s wrong-doing in the case that they did follow a double-blind selection method. They declared a process, followed it, and got a result they didn’t like; but was ‘true’ to the process in some sense.

      I can see why there was disappointment in that result, but I don’t think the takeaway should be “This process is fundamentally flawed” – the process is very fair (it chose in a way that modeled the distribution, if you have 100’s of submissions, and 10’s of slots, 3% is likely <1 total talk), but in this case your goal isn’t to be fair, it’s to encourage the distribution to go up. You need, in some sense, to take an affirmative action[1] to ensure that the minority gets some minimum of representation.

      What might be better in conference planning, then, is to have a portion (say, 75%) of the talks be chosen blind, and the remainder split absolutely equally based on diversity characteristics (race, gender identity, etc). Do not publicly disclose which talks are which, this allows a conference organizer to get a good sense of actual representation (by examining the 75% after they’ve been picked), and also ensure that some minimum amount of diversity in speakership is achieved. The benefit of this approach is that that minimum amount will always ensure that leaders from the minority community[2] is represented, which encourages other members of that community to engage and generally increase the number of people belonging to that group. As the sample rates change, you can start to assign fewer ‘reserved’ seats to that minority, since they will be accounted for accurately in the blind process (assuming all else is equal).

      I think this is a growing pain, ElectronConf was using a blind selection process, but found it insufficient. That’s a good thing. They postponed the event, that’s an unfortunate thing, but indicates they value diversity more highly than the disruption they caused, that’s also probably a good thing, or at least neutral. I think that as they, and conferences like them, go forward, they’ll find ways to balance the host of concerns around ensuring that underrepresented subgroups get to participate in a way that positively effects the community. I think the thing that will be hardest to realize in these processes is the fact that being diverse and being fair are, in some sense, opposed to one another in this situation. Being 100% biased toward fairness will result in a net zero change to the makeup of the community, it is a male-dominated industry, and when you sample fairly, you’ll get a male-dominated result. If you really want something to be diverse, you have to ‘tip the scale’ towards it, and that’s a really hard thing for people (especially, it seems, people in this industry) to accept.

      [1] I believe I’ve just said some words that might make people mad, I trust lobste.rs will react better than most places, though.

      [2] Because if you’re speaking at a conference, you are a de facto leader of a community, perhaps not the leader, but a leader for sure.

      1. 3

        I agree that you have to selectively bias in favour of the minority/minorities to actually put them in the spotlight, however I appreciate the detailed explanation. My point was that there could have simply been 0% of women applying which would’ve meant 0% of women being selected even if they actively tried to select women.

        I don’t agree that you need to be a leader of any kind to speak. While my previous employment was a 100% white male they did in my opinion get something right in that they encouraged everyone to present. It didn’t matter whether you were the best in the subject matter on the team or not, everyone could teach someone else something new.

        On the back of the last point, it would seem to me (personal opinion) that encouraging non-leaders to speak would help even further with the diversity at conferences. Minorities entering the field are (again, in my opinion) even less likely to be leaders than to have chosen the profession in the first place, so encouraging “fresh” people to speak would in the least get their point of view heard, if not teach people new things or ways of thinking.

        1. 11

          If 0 women applied then they may need to put ore effort into encouraging women to apply if they care about having a diverse lineup.

          1. 4

            If they aren’t actively discriminated against, which I doubt github does, then it is up to these individuals to apply and take the final step. You can lead a horse to water, but can’t force them to drink.

            1. 4

              If Github wants a more diverse lineup, they’ll need to be more proactive about it, since the status quo is clearly not going to get them there.

              1. 4

                doesn’t work that way; women tend to estimate their abilities lower, largely due to the societal mental picture of a good programmer as male, white/asian and young. so if you just put out a generic call for applications, you’re going to get them from people who have spent their lives being told they have stuff worth speaking about.

            2. 3

              My wording was inartful wrt ‘leaders’ – my point was that by speaking, on some level, you become a leader. The way to leadership is by leading, and speaking at a conference is in some way leading people (in the case of a conference talk, it may simply be leading towards a technical end or a social one, but it is leadership of some form).

              I don’t think that you need to go looking for leaders, I think you make them by virtue of giving them a platform.

              Definitely there is a possibility that 0% of women (in this case) applied, I think that you’d have to seek out women directly for that subset of talks positively biased (i.e., pull in speakers, rather than having speakers pushed toward you.

              Generally though, I think we’re in agreement, your last paragraph better says what I intended wrt the ‘leadership’ stuff, bringing in people who are not beforehand leaders is definitely a good thing.

        2. 6

          The headline is kinda click-bait-ish. It’s not an all-male conference, it’s a conference open to everybody that happens to have all male speakers.

          In any case, they’ve put themselves in a really awkward position. I think it would have been better to admit that women make up a small percentage of their group, and that it just happened that no females speakers were accepted. With only 3% of the group being women that’s easy enough to believe. Statistically wouldn’t approximately 1 in 33 talks have a female presenter? If the conference isn’t that big, it’s hard to claim they’re discriminating.

          But now how do they change the speaker line up without sending the message, “We don’t value what you have to say, but we need a female speaker”? Maybe others feel differently, but if all the other papers were selected on merit, then being a “token minority” is actually quite demeaning.

          1. 7

            The headline is kinda click-bait-ish. It’s not an all-male conference, it’s a conference open to everybody that happens to have all male speakers.

            “all-male conference line up”

            1. 4

              In any case, they’ve put themselves in a really awkward position.

              They certainly have, and this is why committing to a diverse outcome needs to come from the top.

              In any program context like this one, you build a lineup with goals in mind. You need to choose between big-draw celebrities who might not have anything new to offer and upcoming nobodies with promising ideas. Or between your company’s employees to justify a promotional goal for the conference and outsiders/competitors to show a broad base of industry support. Or between reliable self-promoters who answer every call for participation and self-effacing diamonds in the rough whose future careers might owe your event a debt of gratitude. Any event planner will have these varied, often-conflicting goals in mind as they build a balanced program.

              Passively surrendering the 1/33 status quo is how Github leadership communicates that they don’t care.

            2. 3

              To see what a remarkably diverse speaker line-up looks like for a very technical conference, check out Syntaxcon in Charleston, SC. And 100% of these talks were top-notch.

              1. [Comment from banned user removed]

                1. 8

                  What makes humans so special?

                  The fact that we can decide to control these things.

                  1. [Comment from banned user removed]

                    1. 6

                      and thanks to our big brains we figured out that selecting candidates based solely on merit is superior to selecting them based on gender or race or any other irrelevant trait.

                      A few years ago I’d probably have agreed with this. But now I believe that sometimes you have to artificially create the distribution one wants to become a reality. If done right, it will create back-pressure on the lower levels to make the merit in the desired distribution. There is quite a many good reasons to believe this “it’s just merit, silly” world view is neither productive nor valid.

                      1. 3

                        There’s an important distinction between a quota system (at least N women) and an outreach system (pro-actively ask women to submit proposals so that the inputs to the double-blind selection process match your desired distribution).

                        1. 1

                          Agreed, but having a quota system is the only real mechanism you have available after the CFP is over. Until you reach a point where you get fairly equal distributions you probably need to do both

                          1. 1

                            Stop, go back, and try again. Learn from the mistake. Don’t push forward with a bad cohort of talks; there’s nothing to be salvaged.

                        2. [Comment from banned user removed]

                          1. 5

                            What happened? Who got into your head and convinced you to leave rational methods based on centuries of positive results behind in favor of insane what ifs and cultural experimentation?

                            If your response to someone disagreeing with you on the internet is to get as upset as this post makes you sound, it’s possible that claiming rationality is not your finest move.

                            1. 3

                              What happened? Who got into your head and convinced you to leave rational methods based on centuries of positive results behind in favor of insane what ifs and cultural experimentation?

                              You mean centuries of white-men running the show? Programmers used to be all women up until the white-males realized there was more money there than hardware. So there is no reason to think the current distribution of merit in IT is a reflection of the true underlying distribution of merit. It seems more that it’s a cultural artifact. If something is a cultural artifact you have to manage it on those grounds. Doing everything based on merit will just give you what the culture produces.

                          2. 5

                            Do realize that when you say “we” you don’t speak on behalf of the whole humanity, though. There’s a growing part of hunamity — us, who decided that meritocracy alone is not efficient in the long run and we won’t to change it. So “we” are now competing with “you” on that. Please also appreciate that your point of view is not more natural than ours simply because it was in place for some time. We are just trying something new. This is how change works.

                            1. 2

                              Per my reply down-thread; given a set of candidates, blind selection on merit is the best way to choose from them.

                              How do you get the set of candidates?

                              If your answer is ‘post a CFP and wait for responses’, what makes you think you can attract the best candidates that way?

                              If your answer is ‘proactively reach out to speakers you like’, you’re applying a blind selection method to a biased sampling.