 Nice of you to show up. I'm very glad to be here. I'm Lambda Totoro, and I can't see you anymore. Very good. I'm usually findable at the Acme Labs assembly. And I have come to you to talk about gender. I know. Interesting topic, right? So a couple of things about me. I am many things, among them, non-binary, a master of science in intelligent systems currently. So my background is in computer science. Currently a PhD at the University of Erlangen. And I also do science labs. And basically, this is just another one of those for me. And I've been doing this for roughly, well, I've been pondering this topic for roughly 10 years. That doesn't make me necessarily an expert, but at least it makes me an expert so that I think I can stand up here and tell you some things. OK. What is this talk? This talk is about this exact thing. This is a form that a good friend of mine who wanted to partake in a scientific study had to fill out. And it's in German. I'll translate. It's a question on a scientific survey that asks you for your gender. And it gives four options, male, female, other, and no answer. And while that is not technically the worst thing you can do, it was the motivation for asking myself of how can we do better in this type of situation. So this is a talk about our responsibilities as software craftspeople, what we can do to reflect these things, and what responsibilities we have, and how the tech itself, by itself, just by the virtue of being tech is not neutral. There's some content notes I want to hit on first. This talk will deal with descriptions of transmisia, so hatred against trans people, intermisia, and involuntary surgery. If that's not something you want to deal with, there will be a recording, if you want to see this, at a better time for you. And otherwise, I want to follow roughly this split. I want to do one-part education, one-part form design advice, and one-part fun. Because even though it is the year 2019, and you really think that by now the basics would be down, you still sometimes get forms on the internet like these. I must say at this point, none of these screen grabs that I will show you are made up. I have done exactly zero of them. They were all found somewhere on the internet. OK, a few things this talk is not. This talk is not an invitation to debate the human rights of people, especially trans people, or people in a different culture. So if you want to yell about attack helicopters, and please, there's your fish, do your extremely unfunny jokes somewhere else. All right. So why are forms important? Why is it important that we deal with this at all? Well, if you do sloppy work, it hurts a lot of people. First of all, your data. If your input form for anything, but in this talk, we'll use gender as an example. But if your input form doesn't let people input the correct form or the correct answer to a given question, then that hurts your data. Because you will get incorrect data. It hurts your users. And I can tell you from first, second, and third hand experience that it is really, really, really frustrating if you're supposed to tell somebody something about yourself. And the correct answer is not an option. And it also hurts yourself, because somewhere down the line, these people will complain to you or not use your service when they otherwise would have. So nobody likes this extra frustration. Let's see how we can do better. That brings me to my conclusion slide, which I like to put in front. The central points are as follows. Make the model fit the user. You are modeling your user, if you know it or not. And it's important that you make your model fit the actual users in the world. And don't chuck any model out there and say that the user has to confirm to that. The next one is humans are moving targets. Humans change. Their names change. Their genders sometimes changes. It's really, really, really difficult to just get everything right the first time. And so we change. And your model should reflect that. Of course, the easiest thing to do is just not ask a certain question, because with that you make the fewest mistakes. But sometimes if you do ask something, be at least very precise and upfront about what you're using the data for. Just so people know which is the correct answer to give you. All right. Now that we've got the boring stuff out of the way, we can talk about the fun stuff. The talk is titled, Hello, I'm your edge case. And that is based upon mine and some of my friends' experiences at university. Whenever we wanted to sign up for something like a sports club or a student union or something, whenever there was a little field that says, oh, well, please select your gender male or female, then we would not be able to give the correct answer. And we would have to go talk to an actual person. And the introduction usually went something along the lines of, hello, I'm the person that doesn't fit your system. Or, hello, I'm your edge case. And the format this rest of the talk will take is that we will go through some of the maybe incorrect or maybe a bit biased or maybe a bit, well, let's say updateable functions or preconceptions you might have and then talk a bit about the facts and about the problem and see what we can learn from that. The first one is maybe the most popular one. There's only two genders. Well, some people don't even manage that. And again, these are not made up. Some people apparently still believe that, well, there's just one and it's male, so you're all the same. I mean, while I appreciate generally that the impetus into the direction of everyone being equal, I don't appreciate just everybody being called male instead. But no seriously, there's a lot of non-binary people that are neither male nor female. Non-binary people have been part of almost every culture for almost as long as we have records going back. This is not a new phenomenon. This is not just a trend. And it's gotten so far to the point that even some countries, you can see here, have started adding legal recognition to non-binary genders as well, including Germany. In some sense, we'll get to that a bit later on. But the list is there. It's not very short and it's only ever getting longer. So maybe don't store gender in a boolean anymore. I beg of you, please don't. It is just not enough. It can only hold two things when there's clearly more than two things. OK, now that we've done that, the next step that somebody might take is, OK, OK, OK, OK, I get it. Not two. It's three, right? Three. That's what you said. It's like male, female, and all of the others, or the other. So I'll just do three instead, that I can do. And of course, first of all, I would like to ask of you, if you're designing a computer system and you want to give a third option in gender, please, please make sure it's actually a gender. This thing was given there where you have female, male, and LGBT. LGBT, of course, stands for lesbian, gay, bisexual, and trans. And none of those are genders. Your trans, if your gender doesn't match the gender you were assigned at birth, otherwise you're cis. But it's not a gender in itself. It's not the third option. Please, please never do male, female trans. It just shows how ignorant you are. OK, but if you don't do that, even some corporations have caught on to that there might be more than just those or just three options of male, female, and other. Tinder has 37 by now. And of course, 37 is a lot, but the thing is, is that now the exact number? And of course, it isn't. Whenever you get to a finite number, there will be somebody on Earth and probably somebody within your user base that doesn't fit into any of those boxes you've laid out. I mean, just from the 37 here, we also have gender fluid. And I think this is gender queer, sorry, gender queer, and gender fluid, and agender, and neutra. And we have bi-gender, tri-gender. We have poly-gender, and pan-gender, and a range of other non-binary identities. And the idea is just whenever you get to a certain finite point, it is always somebody who is also beyond that. Don't just put out a limited number of boxes and say everybody fit into that. OK. And the next step, of course, would be, OK, you said 37 and then you said some others. I'll just build all of those. And here's the interactive part of the talk. I don't know if we can get the room lights a bit up. If not, that's also not probably. Whoever here has heard of something called a demi-boy. That's a gender identity? Who has? That's around half a dozen people that I can see. And there's at least 100 people here. So that's a perfect illustration of why you would need a text box, because you don't know all of the possible answers that you can give. And for that, this slide gives a reasonable approximation and is also a little bit of a humorous take on something extremely serious. Humans still change. Humans are still a moving target. Even if you know about all the genders that exist today, you might want to add another one tomorrow because you get a new user that has that gender. And for that, it would be ideal if you can have a text box where people can just add their own gender. It's not a statistical outlier if enough people do it. It's, of course, a little bit of a funny take, but it's also an invitation. If you are non-binary yourself and you feel safe in a certain space, please, if somebody gives you a text field, do enter your true gender. Because if more people do it, it will create more visibility. All right. So from that, we've seen that maybe don't store gender in an enum, either. OK, next up. This is also a bit of a contentious subject, but a lot of people come to me and say, well, I learned in school that there's only two biological sexes, right? Well, you might have learned that in school. I won't dispute that. But I can tell you from my side of things that I once learned in school that there's no gravity in space. Well, that is obviously wrong, even though it's a simplification, something that my younger self was or needed to be told so I couldn't understand or have my understanding move forward a bit so I can then later understand the more complexities that, of course, there's gravity in space, but it behaves differently than the one you know. And it in same things goes with biological configurations and as that relates to gender. There's something called intersex people, and that is if you are born with any configuration of your biology that might be your gonads, that might be your chromosomes, that might be a lot of things, your hormone levels a bunch. And any of these can fluctuate so that you don't fit into the two classical categories. And if you are, if you are born such, then you might find yourself into the label intersex. And intersex people don't necessarily become non-binary automatically. They are not non-binary automatically, but some of them might claim that label. It's up to them, of course. Gender is something you know best what you are, so please just let people talk for themselves. While I'm on the topic, also there's a very serious thing that I want to mention where a lot of intersex people when they were born and people notice that maybe their genitals don't conform to one of the two classical categories, then people are still being subjected to unnecessary surgeries. Like there's no medical need for them, but just to make it easier for the person, a functioning set of testes is removed or the clitoris is reduced. And that is still happening today in Germany too and in the US and in a lot of other places. And if you're maybe social justice adjacent and looking for a good cause to spend some time in, maybe read up on this and learn about that and see how you can affect that. That's definitely something extremely troubling to me. Okay, so biology, maybe bimodal, that is a distribution that maybe has two peaks with the most common outcomes, but it's definitely not binary either. Okay, also you probably don't need to know. Unless you're a doctor and making some input form for very specific medical procedures, then you probably don't need to know. If you're just a social media site or something, you do not need to know your user's biology. Yeah, well, this I also hear a lot. There's so many options now. You said there's at least 37 or might be more. That sounds like work and sure, but I don't buy that. There's a, here's a form from the German military interest union or something. And they give you two options for gender and they give you 71 options for academic title. I could read them all to you, but I literally do not have enough time left in the talk. And with that, I just don't believe you that it's too much work if you do something like this. So if you don't want to deal with the whole thing, please be advised that not asking is always an option. You can always just not ask about it. The information is either needed or not, act accordingly. If it's needed, get the right one. If it's not needed, maybe just don't ask. Some people also say, oh, well, I'll just ask for preferred salutation and infer gender from that. Yeah, spoiler alert, guessing gender doesn't work. Not from name, not from face, not from their tweets. From, there's nothing that you can guess gender from except a field where the user can input their gender. And so you will get this wrong. Even if there's some online services who tell you that they can, they only give you a best guess. And that's usually not enough. If you want to do salutations, sure, more power to you, but at least include like a neutral option. This is one from Lobby Control, I like that a very lot. They have two options and they have a third one that just says hello in German. And the English language also has mix for not mister, not miss is not miss, but mix that's gender neutral is in the dictionary. But whatever language you're using, hello just works everywhere. Okay, quick case study in the case and you're working in Germany in particular, we have four gender markers in our passports. We have numerous options in biology and we have countless options in gender. And I wanted to see, okay, political parties, how do they do this? Who of them shapes up the best? And well, to nobody's surprise, the social democrats do it the worst. They just ask for binary gender, nothing to do about it. And since Germany has a third entry, I'm pretty sure that's illegal or at least legally questionable by now. So next up, also to nobody's surprise, the two conservative parties just ask for binary salutation. That's not very good. We have the free democrats that ask for male, female or diverse. At least they didn't mention all of them. The green party has male, female, inter or diverse or no answer, which is basically okay. And the left party, the socialists, they don't ask at all, which I think is really the best way. The empty model makes the fewest mistakes. Yes, thank you. Okay, my time is running to an end, but you kind of also came for this next slide. No, seriously, how do you ask people for their gender? So here's my best take. Before asking, make sure you need the information. Make sure this is the information you need and you're not really asking something else. Make sure it's mutable in your model. Make sure you can change it. And then just ask. Either tell them, definitely tell them what you need the answer for and definitely tell them that they can just enter a free form answer in a text field, but then you can give either no options and have everybody use the text field or give as many as reasonable possible, that's more than two, and then also have some text field for the ones that are not in the boxes listed above. This is very popular with the target demographic. I also would very much like that, more seen, even there's no text field. Of course, this applies to other questions too. I don't know if you could be Danish, but apparently you can. But generally, this is not just gender, it's not just sexual orientation, it's also names, addresses, all of these things. Not everybody has one, even if you think they do, and it can be really, really complicated once you get down to it. My conclusion slide is this again, if you wanted to talk to me again, you can come up after or meet me at the Acme Labs assembly. Here's some recommendations if you want to read more. That's all I have. Thank you for your attention. Applause, applause, applause, applause. And on the left and on the right are the usual microphones. If there are any questions from the audience, it's now the time or perhaps from the internet. So, somebody is running, it's a good sign. Yes? Hi, is there, you said that in the beginning, there were booleans to store gender, which is not recommended. Do you have any data on whether there was a preference to give the true and the false to male or female or vice versa? Well, the thing is both are false. To Jay. I don't know what you are. The thing is there's not enough answers in two. Both are false. It was just idle curiosity as to whether there is some preference in the data as to in which direction they got it wrong more. The preference is to not give the wrong answer. Good. Okay, a second question or anyone from the internet. No questions, it means thanks a lot. Another applause for Lambda Totoro.