 Hi, everyone, welcome to my talk, Astical of Discussive, Keys Masses, Sexuality and Randomness. A quick disclaimer before we get started, this topic will include adult topics, such as discussions of sex, sexuality, and there will be some mathematics, so if that's not something you're interested in, please leave. Also, this is the first time I've ever given a talk at a conference, so I'm probably gonna mess up and I'm a little bit nervous, but my phone just locked, sorry. So yeah, cool, let's get started. Who am I? I am Go Or Me, I am Go On Mastodon, I am u-underscore1f410 on Twitter. I am a Chaos Communication Coven slash Middleways member. Shout out. I'm a DevOps Engineer, I am Queer and Trans, which is my key smashing credentials. I've got a lot of experience, and I'm also termed online, which is why I wrote an entire talk about key smashes. What I am not is a scientist. I am not a mathematician. I am not a linguist, and these things will become very apparent in a minute. Let's get started. What is a key smash? A key smash at its most simple and popular definition is a random string of letters and symbols typed out on a keyboard or touchscreen used to signal intense emotion in written communication. The earliest reference that I could find was about 2009 on some specific Archive Tumblr blogs, referencing key smashing, but I do think that this phenomenon has existed for a lot longer than that online, but I haven't been able to find records of it. If anyone can find earlier records, I would be really interested, tweet at me. Sorry, I lost my place. It gained popularity in the early 2010s, and LGBT and Black Social Circles online, where it's most often seen. In 2015, The New York Times referenced key smashing in a opinion piece about why parents can't understand their kids, so hopefully this talk will help with any parents trying to understand what their kids are talking about when they key smash. In the late 2010s, there was an emergence of key smashing on Stand Twitter, which if you don't know is extreme fans of something, particularly K-pop stands, which eventually led to a form of key smashing entering the mainstream with VSCO girls. If you don't know what a VSCO girl is, it doesn't really matter, it's fine. It's a specific subculture from the late 2010s. They do this key smash, which is SKSKSKSKSKSKSKSKSK back on the keyboard. After this point, the term key smash enters the wider consciousness with a lot of celebrity sign to use it, and that kind of leads us to where we are today. What does that have to do with LG? What does that have to do with being gay? The key smash is often seen inside LGBT communities as a sign of submissive or bottom-like behavior due to its connotations of extreme emotions. On this slide are a few examples of key smashes in situ collected from real Twitter users, just like you. A quick side note, if you don't know what a bottom or a top is, a quick overview is that a bottom receives insects and a top gives. That is a very simple explanation and kind of reductive, but if you don't know what it is and you're here, okay. Often key smashes are used to indicate that a user is flustered or embarrassed, and it's often seen along with the stereotypical bottom emoji, that is the one there with the two pleading fingers like that. It isn't exclusive to bottoms, so in 2020, I made a post, just a random shit post, wondering if bottoms key smash better than tops, because it's important. So I did what any rational person would do, and I set up a survey to gather data to analyze key smashes. To actually do this though, we have to define and figure out two things. What is randomness and how we calculate it, and how do we define bottomness? Let's start with randomness, because that's slightly easier. It's actually pretty hard because I'm very bad at maths, but we're gonna use Shannon entropy. Shannon entropy is the amount of information, surprise, or uncertainty in a set of data. We basically compared the entire possible alphabet of the input, so all possible characters, and then compare how often they are used in the actual string. So basically we can see how often they're used, a bit of distribution. It's not perfect for this use case, but it does work for these intents and purposes. We can also use this to calculate what a perfect entropy score would be, so what the ideal possible random key smash would be, and then compare them to get a percentage, so how random it is. Bottomness and topness. This is harder. Defining bottomness and topness kind of gave me a few false starts. Basically we started with looking at stereotypical traits of bottoms, but that kind of led to this very reductive understanding of what it is, and it kind of trended towards the concept of gay men bottoming, which is, I wanted to be more inclusive than just that. To actually get this data, we decided to allow people to self-identify on a scale of one to 10, how much do they identify as a bottom, and how much do other people consider them a bottom. This probably isn't the best way of gathering data, but data collection's really hard, guys. We also did the same for topness just because it means we can compare and contrast later on. We also got every respondent to give us free key smashes to analyze, to gain an average. Then I set it out. I got 62 respondents from the social media network mastered on. I had to remove a few, because some people left very unhelpful comments instead of actually key-smashing, with one person saying that key-smashing is bad for the keyboard, and that it is a very unclear form of communication, which I would contest. It's very clear. So, data. Firstly, you can see the key-smash randomness as a percentage is mapped against, sorry, bottom quotient. That's the word I used. As you can see, we have a very shallow curve that starts low, picks up in the middle, and drops back down. These results are not conclusive. Clearly, we can see the people who have a low bottom quotient and a high bottom quotient key-smash less randomly, which is weird. For comparison, key-smash randomness against top scores, a very similar curve, almost, if you kind of squint a bit. What does this tell us? Sorry, that was a surprise. What does this tell us about key-smash randomness and sexuality? It tells us that tops and bottoms have no real relationship to the randomness of key-smashes. It also tells us that those who can't define themselves as either strongly atop or strongly at bottom are more likely to key-smash more randomly, so maybe it switches to key-smash the best. If we circle back to the beginning of the study, I'd said specifically if bottoms key-smash better. And from the definition where I said it is a random spamming of the keyboard, you might think that randomness is best. But I disagree, and that's very important coming up. On the screen, there are two key-smashes. I'm basically gonna ask you to raise your hand if you think the right one is the correct, the better key-smash, and if anyone thinks the one on the left is better. Okay, it's your left. I thought everyone was completely wrong for a moment there. Yeah, exactly, right? That one is far too long. It's got numbers. It just doesn't look right. This one is a little bit more tricky, so on the right again, anyone think that's the correct one? Yours? And the left one? Yeah, same again. It's kind of obvious that that one is the correct one. Obviously, guys, come on. But there are unwritten rules to key-smashing. Gretchen McCulloch, who is a linguist, mentions this in her book because it's net, and I really highly recommend it. It's really interesting. It analyzes how we communicate online and how it's changing, how we talk and net speak. So if you're kind of fan of this kind of stuff, then yeah, obviously you're gonna like that. In her book, she references that there are five-ish rules, kind of vague outlines, so I decided to try and formalize them. So we're gonna start with going through the rules. I meant to show that slide. Rule one, key-smashes start with ASDF. This one is a very contentious rule, and a lot of people disagree with me, but I think it's pretty important to a key-smash. The first character's pry must to prepare to see a key-smash and kind of register it as such, otherwise it's just random. By the way, this is all what you're working with, a QWERTY keyboard. I will come up, that will come up later. Dvorak uses, if you're in the audience, you can't key-smash. Yes. Actually, that's an interesting side note. A lot of people I spoke to who used Dvorak keyboards said that they don't key-smash specifically because it's not socially legible. When they key-smash, it doesn't look right enough, and they just delete it, which is, it's interesting, to me at least, and probably you if you're here, but I hope so. Most people start when they key-smash, whether they're left or right-handed, on the left-hand side of the keyboard, they're left-hand, so let's move on to the next rule. Alternating sides. A key-smash moves back and forth across the keyboard. You move like that. Sticking on one side just doesn't look right, it doesn't look, it just looks... Most key-smashes are written with both hands, and even people on mobile keyboards do this with both thumbs. Rule three, mostly middle row. Most key-smashes are entirely made up of keys on the middle or top row of the keyboard. I think this is because most people type start on the middle row, and when they're excited, they stretch their hands out. Some key-smashed purists do say that it should only be the middle row, and the top row is no-go, but I disagree, I think it's important. Rule four, nope. Rule four, mixed case. No. Rule five, no numbers or punctuation. This one is also a little bit contentious, and some people said, oh, should I not use punctuation? Not really excited, but I don't know. I personally think this is an important rule. So, with these five rules we've defined from careful analysis or key-smashes, we can now go ahead and mathematically apply these rules to key-smashes. This one gives a much better way of scoring how good a key-smash is and what we can actually do with that data. The specific code that I wrote for each rule is not online currently. It's terrible, I need to fix it. If anyone is interested in working on this, hit me up after, I'll let you know where I am. But basically the way it works is each rule, each key-smash is fed through each rule, gives it a value between zero and one, that is all assigned together, to get a total score out of five. So, here are some example key-smashes that I ran through the algorithm, and yeah, you can see it's kind of what we'd expect. The one at the top is clearly a key-smash, and the one at the bottom is kind of, I don't even know what it is, it's just a blur. So, now that we can calculate the perfect key-smash, we need more data. We always need more data with more fields to analyze. The last survey had a lot of issues. For example, there was no way to know if the respondent was using a mobile or desktop keyboard, what keyboard format they were using, and we only can use this with QWERTY keyboards. So, had to basically throw all that data away. So, I sent out another survey asking for more key-smashes. Instead of asking about topness and bottomness, I decided to move away from that and look at sexuality, gender, romantic orientation, as well as asking about device type, keyboard type, keyboard layout, and also most favorite social media platform, because we want to know. It's important research, everyone. I also want to know if this has any effect on the outcome. With the new data cleaned up, 110 records, mostly sourced from master's on Twitter and a few Discord servers. So, before we analyze key-smashes, I just want to have a quick look at the data with everyone. Keyboard information, it's pretty much standard, as we'd expect. Most people using QWERTY, a couple of Zerti users, and I think three Dvorak users. So, you got some friends. One person just decided to key-smash in every single field, so I had to delete that respondent, don't know what that means. There's a really even split between mobile and desktop users, which was pretty useful, actually, and also a very even split between touch typers and non-touch typers on desktop. This is where I learned a very important lesson. Don't let people put in whatever the hell they want in your forms. Use fields. God, because I spent so much time cleaning up data and making it look right and sorting out sexualities and defining it. If someone put pan slash by or by slash pan, I had to kind of reduct it down to one, which isn't the best. I'm not, like I said, I'm not a scientist. Sorry, my phone locked again. Sexuality and romantic orientation were pretty standard. There was a large amount of asexual representation. The romantic orientation is pretty evenly split as well. Like I just said, all those points about sorting out data, well, they're about to get a lot worse. Gender. Gender. I spent a couple hours trying to sort out the gender pie, but I could not. There were, at the end of the day, 62 discreet different gender entries. And the ones on screen are just the largest. It is basically impossible to do anything of this data because it's collected in a way that makes no sense because if someone put male, I don't know if that's cis or trans or non-binary, just useless. So that was fun. Ideally, I do want to do this study again with all the things that I've learned. And if any universities or research institutions would like to fund me, again, get in touch afterwards. Finally, social media usage. Pretty standard. The majority of people mostly use Discord or Twitter as their primary form of social media. When asking for primary form of social media, most people gave multiples. So I just picked the one at the front, the one they put in first because I did ask for one and they gave me multiple, which is wrong. It's not me who's wrong, it's the respondent. Okay, with this out of the way, it is time to rank data. Using the average key smash score from the free key smashes provided, we can now look at the key smashes, look at the data, and see if we can draw any interesting results. So, nope. By the way, I've also included key smash randomness on the other side, just as an interesting comparison. The key smash score is the calculated thing. The highest scoring key smashes come from Masteredon. Good job, everyone, on the FETIverse. Second comes from Slack. Free people said that Slack was their primary form of social media. I don't know if the kids these days are really into Slack or if it's a joke, but they only key smashed a little bit worse than Masteredon users. Falling on from that is Discord and then Twitter and surprisingly Tumblr. I don't know about a lot of people in this room, but I used Tumblr and a lot of people on their keys smash a lot. So I was kind of expecting to have a higher score. Sexuality, bisexuals in the audience. Mathematically, you are the best. I've proved it. Behind that is asexuals, by a large margin for this system, lesbians, quite a big gap. And a surprise to everyone, especially me coming into this research, heterosexuals above homosexuals, men loving men, I had to split what that meant because a lot of people put in, I'm just gay and I'm just, okay, this is not very specific, it's fine, it's fine, it's my fault. That's interesting. Key smashes are an LGBT thing, right? Well if we look at this data, it says otherwise, but if we do use romanticness as well, so heterosexual, heteroromantic, you know, normal, we can see that they actually have a lower key smash score than homosexuals, so everything is right and it's in its place. So let's look at some issues with this because God, are there a lot. Firstly, there is no algorithmic waiting. Each rule is worth exactly the same number of points. That gives us a bit of a problem because some rules are more important than others. Which ones? I don't know. Everyone I spoke to about this has different opinions. Secondly, what even is a perfect key smash? Well I did it, I did the maths, but a lot of people said that these key smashes look formulaic and boring and very clinical and I kind of agree. I think there's a little bit more to it than just the rules that we have defined. There is a breaking of these rules that gives it a real certain something, you know? A little secret sauce. Also it is only, quotey only, the data collection design issues that I mentioned. Conclusions, defining language or key smashing is really hard. Linguists have it really rough, apparently. Language is malleable and changes all the time. It is very difficult to define something which is almost exclusively online. And why do we even care how good a key smash is anyway? When we are locked inside for two years, no way to communicate face to face like we used to with all these little micro expressions and gestures. This is how we express ourselves. This is how we show emotion in a way that is beyond language. It is like emojis and, you know, lol. It shows emotion without real words and I think that's really kind of beautiful. I think it's beautiful, you don't have to agree. Key smashing isn't just for bottoms. It isn't, everyone does it, probably. I'm kind of rushing for time because I realize I'm about to run over. Take your time. If you're gonna do something like this, if you're gonna take a shit post, a joke idea and turn it into an entire talk like I did, don't. No, do, seriously. It was a great learning experience. I learned a lot about science and key smashes and linguistics and now I'm here. Just get a second pair of eyes on things, try and refine your search methods, your data collection methods before you actually do it or you'll end up like me. That's it, really. Before I finish, I'd just like to thank my partner, Kim, my friend, Colin and Ziya for helping me and letting me yell at them about key smashes. If you're interested in any more, I will be at Millerways or the Coven or around that area for most of EMF. Thank you for listening.