Victoria Machado de Oliveira 4:25 Here is good? 4:27 Thank you. 4:31 Good morning, everyone. 4:33 Good morning. 4:35 Well, let's start this. 4:39 I will speak about how to have more non-English speakers users in general, not only for BlueSky and stuff. 4:46 I will explain a bit how this is going to work. 4:49 I have a little bit of a presentation where I will introduce you to the topic, talk a bit about stuff, and by the end I will have a few minutes so we can discuss things, including people watching from home. 5:11 On screen, please. 5:13 I will read the screenshots if you have any contributions, especially non-speakers. 5:20 Participants will be very welcome to join the discussion at the end. 5:27 Okay. 5:29 Oh, no. 5:32 I don't know. Speaker B 5:36 Help me. Victoria Machado de Oliveira 5:43 Okay, okay, I figured it. 5:46 Okay, let me present, introduce myself a little bit. 5:51 Oh, sorry, I forgot the mic, it's here. 5:55 I will introduce myself a little bit. 5:57 I am Victoria, I'm from Brazil. 6:02 My pronouns are they/she in English. 6:06 My handle is vickwalker.dev.br on Bluesky. 6:12 I am— and my occupation, my real occupation is I'm a QA architect and I also for 8 years already and I also do community moderation a job for Brazilian creators for more around 12 years already. 6:38 Also, a little of my background too. 6:44 I am graduated as game developer and IT technician. 6:52 I also did— I do a little bit of help translating some apps even before the AT protocol. 7:01 I helped some apps like Twidere. 7:05 I don't know if anyone will know this old app for Twitter. 7:10 It was really cool. 7:13 And I helped translate a bit for TalkMekki, the BlueSky client. 7:22 From Japanese to Portuguese. 7:25 I also used it to be a streamer back in some more than 10 years ago, but today I just stream as a hobby. 7:41 I also founded AT Brazil that you can find by the handle AT Proto. 7:49 Www.tutorial.com.br and I'm based in São Paulo, Brazil. 7:58 Also, you might see my— might find me using this icon a lot everywhere, but this is from a little friend of mine, so just to know. 8:15 Okay, let's go to the main topics I will be talking about. 8:20 So, there's 3 big topics I want to talk of a bigger problem, a barrier to keep your users that they don't speak English as a first language. 8:40 First off, the onboarding. 8:44 I want to provoke a little bit of questions for you all. 8:51 The first impression is always important. 8:54 You need to ask, is your onboarding generic enough for people from multiple countries to understand how to get around your app? 9:07 Or not. 9:08 If you are not sure, you can just ask some other— search for some people that are open to give your feedback. 9:20 It's always good. 9:23 Also communication. 9:24 After they are using your app, you can't just take that for granted. 9:32 Having a good open and healthy communication with your user base is really good. 9:36 And communities will always be very happy to be acknowledged and be listened to. 9:46 So please also be very careful because science people speak different languages and not— English is not their first language. 9:57 Words can be understood in different ways. 10:01 So please be very mindful about your wording and stuff. 10:08 It can be very impactful of your success of your app. 10:14 You can scare away many users. 10:19 Also translation. 10:20 I won't go too deep because right after me, my Little friend here will have a full talk about the challenges of translating Blue Sky to French. 10:33 And it's going to go very deeply in this topic. 10:37 But I will give a little bit of tips and resources to help with that. 10:46 Like a little fact check, less than 20% percent of the global population actually speaks English. 10:55 So you can't count with having English is enough for your app. 11:05 Also be very careful because machine translations can't solve everything. 11:14 Even with the advanced AI models, You need to be very careful. 11:19 I will enter into this topic now a little in a few minutes. 11:27 Okay, let's talk about the onboarding. 11:31 Users are having a lot of problems to find what matters to them. 11:39 People that speak their same language, are talking about the same topics, they, they, it's relevant, it's meaningful in their region. 11:54 And so I know it's not easy to solve this problem, to have a good algorithm, but you can also have though, we have a big community of developers here. 12:10 And builders and stuff. 12:12 You can always try to solve this problem together. 12:17 Also, the creators also need to understand if it's worth to use your app. 12:27 Like most— I'm speaking of like content creators in this topic. 12:35 Like There's a lot of creators that are spreading their presence in multiple platforms, and it takes effort to do this. 12:46 So if they don't see this effort paying off, they might not try to keep their presence in your app. 13:05 About the onboarding and stuff, there's a lot of questions. 13:09 How to do the best UX to that in many languages, or how to explain the protocol to users that they don't need multiple accounts if they don't want to use multiple apps. 13:27 There's always open questions that we don't have the solid answer yet, but this group of research will be doing, taking these researches. 13:41 So I recommend you scan this QR code and keep them in your notes. 13:51 Also, I forgot to I'm going to show a lot of QR codes. 13:58 Be ready to take the pictures or scan this stuff. 14:02 I will wait a little bit. 14:07 Ready? 14:10 You can always ask me for the links and the presentation after on BlueSky. 14:16 I can share those too. 14:21 So also this wasn't planned originally because when I was here in the Science Day for the conference, I saw this very interesting presentation about they started migration to BlueSky for a scientific community and they have a really Insightful research in data. 14:46 So if you missed it, that it's okay. 14:50 The AT Proto Science will post the recording later. 14:55 So save the— follow AT Proto Science to just check this data because it's not exactly the topic I'm talking about, but this data is kind of related in useful. 15:16 Okay, let's go to communication now. 15:20 If you really want to attract a certain audience, you need to communicate with them, otherwise they might not know about your app and that your app can be useful to them. 15:40 They really appreciate the effort to be included and be listened to. 15:46 You can start doing simple steps like announcing your new features, your features in different languages. 15:55 Simple text can be handled by auto machine translators, but I Still, be careful about translating more complex text. 16:15 I have a little good example. 16:18 Leaflet did a really cool job in the beginning. 16:23 They post like— they ask for feedback and requests for different groups like artists, scientists, fandom people. 16:33 And even though they did it in English, they managed to push new features, very cool features that everyone is benefiting. 16:45 I can see some non-English speakers using Leaflet and enjoying those features. 16:52 So you can do this kind of thing too. 16:57 This QR code will I can take you to this list of posts if you want. 17:06 Also, let's talk about AT Brasil, the work I'm doing. 17:10 Like, the AT Brasil, I'm trying to bridge this gap in communication. 17:17 I try to translate in simple ways, in quick ways, the The biggest updates are around the atmosphere, the blue sky stuff, things I see Brazilians being confused about. 17:32 I try to explain in a simple way for everyone to get and also to provide the technical Brazilian developers the resources to dig deeper in the information that not always is getting out of the English bubble. 17:51 Inside Blue Sky and stuff. 17:55 That's a thing that more people can do and will be very helpful to spread the word. 18:08 A little of reality check. 18:12 That QR code in the top right is a graphic to the most used languages in BlueSky by day, by week, month, and years. 18:23 You can check out, and I will just highlight some examples of the most used languages: Japanese, Portuguese, and Spanish. 18:37 Like, in Japan, I found that around 28% of the population speaks English. 18:48 In Brazil, it's just 5% of people that is comfortable speaking English. 18:55 And Mexico, around 13%. 18:59 And in Spain, 22%. 19:02 So you can see the percentage is not that high, even though you might think about. 19:11 It was higher. 19:12 So you need to think about it. 19:19 Okay. 19:19 Translations. 19:21 It's not as simple. 19:22 I recognize it's not a simple task to have good translations. 19:26 Machine translations can't be up to date with new popular ways to— in terms across the internet. 19:35 The language is moving really fast around the internet. 19:39 So the translation won't always keep up with that. 19:43 I will give a little example in the next slide. 19:48 Be cautious of relying on AI because you don't know if that model is well trained in that language you want. 19:58 There's multiple models that have different training and it can you can maybe have more problems relying on that than not. 20:14 Localizing your app will make a huge difference, not just translating. 20:18 I will give a little bit of examples, but again, the next talk will go even deeper about this. 20:31 A little example is from Twitter. 20:34 In Brazil, tweet became tweet, like that. 20:41 And the users became known as tweeteiros. 20:47 Like this kind of thing, a machine translator won't get for a long time. 20:54 They might know now. 20:57 Because Twitter is around for a couple of years now. 21:02 But for new apps, you won't know by relying on machine translations. 21:09 A little tip that I see a lot of people struggling around the atmospheres— time zones. 21:16 Please avoid using regional names for time zones. 21:22 US Central doesn't mean anything in Brazil. 21:25 Please. 21:33 A good way to be safe is using UTC. 21:36 UTC is the Universal Time for a reason. 21:40 UTC is more easy to be understood around the globe, so I give you this tip. 21:47 It's easier because if you also want to list a big cities, you will need a really, really huge list to maintain, and you might not be familiar with cities from different countries you never heard about. 22:05 So I think the good old UTC is a safe bet to go for that. 22:17 Now let's talk about tools to help you translate because I know not everyone has a lot of money and multiple people inside their teams to help. 22:30 So you can use some tools to have the community to help you translate. 22:37 This one NPMX is using is called Lunaria. 22:42 You can see in this screenshot, you can see the translation from Portuguese. 22:50 There's 89 strings translated, so there's need some updates and it lists how much, what you need to translate. 23:03 Unfortunately, it's a little bit technical. 23:05 You need to know how the people translating need to know how to commit and push code and find those strings. 23:16 It's not that simple to use. 23:23 There's this tool that is used by TalkMaki. 23:27 It's a little bit easier to use. 23:31 You can use this interface to see different strings from different languages to translate, but it requires a connection with GitHub. 23:46 So having a GitHub account, it's not something that non-devs are very familiar with, but when you have the GitHub account, you can do everything through this interface. 24:01 But that is a point to note that might be a problem to gather help from non-technical people. 24:13 And that is Crowdin that BlueSky is using. 24:17 It's really less technical. 24:20 You don't need to touch code at all. 24:25 You just see the text and translate in an app. 24:29 They have even mobile app, so it's really, really easy to use. 24:35 But also they have some requirements to use this, so check if your project can use this bit first. 24:49 Now it's time for the discussion. 24:53 I will check the live chat, the screenshot. 24:56 And please, I ask for non-English speakers to bring your point of view if you are comfortable about some of the topics I said. 25:09 And English speakers, you are welcome to bring questions and listen. 25:17 And listen. 25:18 So we have about— 8 minutes to do this. 25:25 Let me see the chat. 25:29 Anyone? 25:31 There's a microphone. 25:33 Thank you. 25:34 Yes, so if people can raise hands, I'll bring the mic so the people on the live stream can hear us. Speaker B 25:40 Yeah. 25:48 So when localizing an app user interface, is it enough to just translate the words, or are there certain different paradigms of control and layout that you need to be thinking about based on different expectations from different communities? Victoria Machado de Oliveira 26:08 Yes. 26:08 I can give you some examples that will be more in depth with the next talk too, with real examples from French, but I will give some examples in Portuguese too. 26:25 When translating from English to Portuguese, as an example, there is a big difference. 26:32 Portuguese, French, Spanish, Italian, it's all very gendered languages. 26:40 So imagine if you are translating a word in your interface like followers. 26:50 It's a small word, but in different languages it can be really big. 26:55 And if you want to accommodate different genders, you might need a bigger space for that. 27:02 So that's a really good problem, a really a real problem to accommodate this in your interface. 27:10 How you adapt your user interface to bigger translations, words. 27:18 How do you make these words short or not? 27:22 How do you accommodate the needs? Speaker B 27:30 Yeah, thanks Victoria for that. 27:32 I will add to that because beside French, because I will talk about French for 30 minutes and including about gendered language, and you are not ready for that. 27:42 But also for other languages, I'm not that familiar about that, but please keep in mind that there is right-to-left languages, and this is a big thing about layout and stuff. 27:56 So, but one of the main things that I will also say, but you don't have to do it perfectly at first. 28:03 You just have to care. 28:05 If you care, it's cool. 28:07 Translators will come to you and ask for stuff, and you will have to understand that their demands make sense. 28:15 We don't come to you if we don't think that it is important for us. 28:21 And you just talked about the string length, because the classical example is German. 28:30 German love to agglomerate words into big words, and you may not have that in mind when you are designing your beautiful markup. 28:38 But of course,