WEBVTT 00:00.000 --> 00:12.680 Okay, we'll start another one. 00:12.680 --> 00:19.680 We have talked about Ray, done by Ray. 00:19.680 --> 00:21.080 So the stage is here. 00:30.000 --> 00:58.800 So let's start, okay, okay, so good afternoon, welcome everybody. 00:58.800 --> 01:06.600 My name is Ramon Santamalia, Ray, I'm the creator of this small graphics library, Ray 01:06.600 --> 01:07.600 Lip. 01:07.600 --> 01:13.440 Today, I'm going to talk about how it's been, has been my experience, my 12 years experience 01:13.440 --> 01:21.600 working in this library and what have been the main challenges as a solo developer. 01:21.600 --> 01:27.720 So let's start, but before I start, I would like to ask who in the audience knows what 01:27.720 --> 01:28.720 is Ray Lip? 01:28.720 --> 01:29.720 Could you raise your head? 01:29.720 --> 01:39.080 Wow, so nice, what a real, well, for the ones that do not know about this library, I'm going 01:39.080 --> 01:43.440 to start with a brief introduction, what is Ray Lip? 01:43.440 --> 01:52.800 Well, Ray Lip is a graphics library that I created for education and one moment in my life 01:52.800 --> 01:58.200 12 years ago, I had to teach video games programming to artist students with no previous 01:58.200 --> 02:06.320 programming knowledge, yeah, it was a quite difficult task, so I decided to create, to create 02:06.320 --> 02:16.520 a very simple and easy to use a library that was quite enjoyable to use, to put things 02:16.520 --> 02:22.480 on the screen, to put graphics on the screen, well, in the 12 years, the library has 02:22.560 --> 02:29.480 grown a lot, well, actually, grow, not that a lot, but has been ported to many platforms 02:29.480 --> 02:33.280 and has received bindings for many language. 02:33.280 --> 02:39.440 At the moment, Ray Lip can work in many, many platforms, as you can see, and also it has 02:39.440 --> 02:45.440 been ported to, I think, more than 50 or 60 programming languages. 02:45.440 --> 02:53.160 Still, I think that in that as a logon, the key word is enjoy, because it was designed 02:53.160 --> 03:01.080 to be enjoyable to use, and I think that's the key factor of the library. 03:01.080 --> 03:02.080 What can I do with Ray Lip? 03:02.080 --> 03:03.080 What can we do with Ray Lip? 03:03.080 --> 03:11.320 Well, he's a small demo, actually, this presentation has been made in Ray Lip, everything 03:11.320 --> 03:19.160 you can see, it did be made in Ray Lip, Ray Lip is a small code in Ray Lip, actually it's 03:19.160 --> 03:27.040 a functional code in Ray Lip, it has been also made in Ray Lip, with just about 10 or 12 03:27.040 --> 03:33.520 lines of code, you go to a basic window with some text, which is on it, also you can do 03:33.520 --> 03:42.520 to the here, with about 100 lines, you can implement a bunny mark, this is a typical test 03:42.520 --> 03:50.200 to test performance, totally, in Ray Lip internally uses a batching system to draw many, many 03:50.200 --> 03:59.640 many sprites in a very performant way, also it can do 3D, again, with just about 80, 90 lines 03:59.640 --> 04:09.600 of code, you get a full 3D application, you can do shaders, it is a Julia set implemented 04:09.600 --> 04:17.280 with a shader, here you can load the shader, control all the parts, and yeah, it's 04:17.280 --> 04:29.320 actually a functional Julia set running in shader, and ledly Ray Lip also has been used in 04:29.360 --> 04:39.360 some companies for embedded system, it's a company that has recently used Ray Lip to create 04:39.360 --> 04:49.120 all the UI for their autonomous driving devices, they have a device, that you can buy, 04:49.120 --> 04:55.280 all the UI has been made with Ray Lip in the light, latest version of the Komai UI, the 04:55.280 --> 05:11.400 open pilot, open software, yeah that's what you can do with Ray Lip, maybe you are wondering 05:11.400 --> 05:17.960 what happened in 12 years, well actually many, many things happen in 12 years, as I said, 05:17.960 --> 05:26.200 library start as a weekend project, it was a small library for my students to teach them 05:26.200 --> 05:32.360 video games programming or graphics programming, but yeah, along the years has many things 05:32.360 --> 05:40.880 had happened, it's a small timeline with all the events that I have been recording, along 05:40.880 --> 05:46.760 the years of things that happened, everything has happened, asking for some library in 05:46.840 --> 05:54.840 Kronos group forums, releasing versions, keep working on the library, keep working on secondary 05:54.840 --> 06:06.560 library, it's start doing some talks, participating in global game jumps, so, yeah, creating 06:06.640 --> 06:15.200 additional libraries in 217, enjoying the handmade network, that is an network that for 06:15.200 --> 06:21.440 handmade software, and it actually was a boost for the library, so, because many more people 06:21.440 --> 06:30.480 so about it, so, keep working in the library, at some point I start working on separate tools, 06:30.560 --> 06:36.000 a support tools for the library, I could talk a bit more about that later, and then by, yeah, 06:36.000 --> 06:46.640 that's it, by 2018, I got a Google open source peer award, and actually it was a big boost 06:46.640 --> 06:52.800 for the library, because since that moment, many people start noticing it and knowing about the 06:52.880 --> 07:03.680 existence of railing, so, keep working on it, year after year, I have not been working full time 07:03.680 --> 07:11.520 during most of the years, but for some months I've been working full time, also by 2020, I received 07:11.520 --> 07:22.720 an epic mega-gram, and again, the library become more popular, by 2020, I quit my job at the 07:22.720 --> 07:27.440 moment, and decided to work full time for some time in the library, so, in the following, we 07:27.440 --> 07:36.480 engaged, many things happens, and other Google open source, no, peer bonus award, keep developing tools, 07:36.480 --> 07:43.760 keep giving talks, keep developing tools, some more tools, some more tools, some different versions 07:43.760 --> 07:50.640 of different tools, and support libraries, well, I will talk a bit more about that later, but, yeah, 07:50.640 --> 07:56.080 I keep working in the library very, very actively, and in the last year, actually, I was doing a 07:56.080 --> 08:06.400 lot of talks around developing, supporting, yeah, and that's important, and recently, but, 08:06.400 --> 08:15.840 meet last year, I received an NL at Fund, actually, I don't know if they are in the audience, 08:15.840 --> 08:22.960 but I want to thank you them, because for the first time in 12 years, I was able to work full time 08:22.960 --> 08:31.680 paid on the library, and it was a real boost for it, because I could manage my time better, 08:31.680 --> 08:38.640 I'm plan new feature and improvements for the library, and since then, I've been working full time, 08:38.800 --> 08:49.200 developing, oh, yeah, also in by the end of last year, a puffer AI, that it's a reinforced learning 08:49.760 --> 08:57.760 company that uses really to set up training environments to train, age, age, and that play video games, 08:59.600 --> 09:06.320 has been a sponsoring the library, it's the first kind of corporate sponsor I had ever, 09:06.960 --> 09:17.920 and, yeah, and also, like AI, became the second corporate sponsor in 12 years, and yeah, 09:17.920 --> 09:25.120 and that has been everything that has happened, many, many things happened, I know that it went very fast, 09:25.120 --> 09:40.320 but actually, it was quite an adventure, 12 years, so. So, now, let's talk about what happened 09:40.320 --> 09:47.200 for me, the main challenges, working in the library as a solid developer, as a solid maintainer, 09:48.080 --> 09:58.800 I define those challenges, the three amps for me that were maintenance itself, it's a big 09:58.800 --> 10:04.880 library, there is a lot of things to do all the time, issues, it supports many platforms, so maintenance itself 10:04.880 --> 10:14.240 is a really hard work motivation, because it's quite difficult to stay motivated all the time 10:14.240 --> 10:22.800 along 12 years, so, well, it was difficult, so I will explain what I might techniques, 10:22.800 --> 10:30.800 so they motivated, and money, okay, I might have done money, so I need to open some software. 10:30.800 --> 10:39.040 So, let's start with maintenance, that's a question that I would, I would have liked to ask 10:39.040 --> 10:44.000 me myself when I just started this weekend project, however I'm willing to work on my project, 10:44.400 --> 10:49.360 because I actually, it's a thing, I think it's a question that everybody should ask themselves 10:49.360 --> 10:53.760 when working in an open software project, because depending on the answer to that question, 10:53.760 --> 11:00.640 you can't plan for the future way better, because if you know that you are working in a project 11:00.640 --> 11:07.120 for five years, ten years, 12 years, probably you will take some decisions from the beginning, 11:07.120 --> 11:12.480 some better decisions from the beginning, but it's difficult to ask that question, 11:12.800 --> 11:19.120 it's not common to ask that question, so considering that you probably, or probably not many people 11:19.120 --> 11:23.200 ask that question when it's starting an open software project, my recommendation, 11:23.840 --> 11:29.360 keep it simple, keep it simple, it's something that work really, really, really well with 11:29.360 --> 11:35.680 really, along the years, when I just started the library, the library was intended for 11:35.680 --> 11:42.080 the youngest students, and I decided to keep it as much simple as possible, and yeah, and it 11:42.080 --> 11:50.480 worked, what does it mean to keep it simple? Well, in the case of Riley lip, it means simple and clear 11:50.480 --> 11:57.680 code ways, the code base, it's relatively small, so it's about 50,000 miles of code that 11:57.680 --> 12:01.680 really be self, then there are some support libraries that they are usually single file header 12:01.680 --> 12:09.360 only libraries, that are way bigger than 30,000 lines, but Riley lip itself, it's only 30,000 12:09.360 --> 12:17.840 lines of code. I try to minimize API additions, at the moment Riley has 600 functions, approximately 12:18.880 --> 12:25.600 first version is start only with 80, 80 functions, so it has grown a lot in the years, it's 12:25.600 --> 12:31.520 difficult to keep a project simple along the years, because sooner or later you start adding features, 12:31.680 --> 12:41.520 other functionality, and it's very easy to go out of control, so be careful, Riley has a very 12:41.520 --> 12:50.880 simple build system, I don't like build systems, so I try to keep things super simple to build, 12:50.880 --> 13:00.160 I tend to work usually with a single C file, very long C file, but I like to only have one file, 13:00.160 --> 13:05.760 and be able to build everything in the command line with just one line, so that's something that 13:05.760 --> 13:12.320 I also apply to Riley, because back in the past working in some companies, there were always 13:12.320 --> 13:18.480 problems with the build systems, super complicated build systems, so I try to keep it simple in 13:18.480 --> 13:26.160 that in that words, and finally something that really helped Riley has been automation, at some point 13:26.640 --> 13:33.520 Riley can start supporting many platforms, and yeah, it was, and many architecture, 13:34.640 --> 13:39.040 some kind of continuous integration, because you know the level of system was extremely helpful, 13:39.040 --> 13:44.080 but also some super tools that I've been creating along the years to help. 13:46.400 --> 13:54.160 About maintenance, that's actually the structure of the library with all the files involved, 13:54.160 --> 14:02.800 actually, you've got the links to the different files here, but that's the core library, that's 14:02.800 --> 14:09.200 Riley library, and that's just the architecture, maybe it's not that clear how it maps to code, 14:09.200 --> 14:17.600 but that's it, actually it really, there are seven C files, some bigger, some smaller, 14:18.560 --> 14:24.880 just 30,000 lines of code all together, so it's quite portable, self-contained, 14:24.880 --> 14:30.960 it has no external dependencies, so you can build the library in a single static library, 14:30.960 --> 14:36.320 actually just recommend it to be used that way, and you can link it with your executable for the 14:36.320 --> 14:41.440 back-and-platform you like, and yeah, that's it, you got the library. 14:41.840 --> 14:48.960 Okay, about automation, well, as I commented, continuous integration continuous delivery system was 14:48.960 --> 14:56.640 really helpful, another point about Riley is that learning how to use the library is based on 14:56.640 --> 15:03.200 examples, Riley has a very big collection of examples, has at the moment it has about 200 examples, 15:03.200 --> 15:10.400 and at some point it became quite difficult to manage that collection of examples, because every time 15:10.480 --> 15:16.960 any example was added to the library, it was not that easy to add the example dot c file, 15:16.960 --> 15:21.440 it required to be out in several build systems, it required testing, it required to follow 15:22.640 --> 15:32.240 conventions, so actually it required modifying about, I think it was 27, 28, separate files in different 15:32.240 --> 15:37.920 places, I created a tool to do that automatically, it's just an example of some 15:37.920 --> 15:44.560 manager that allows to add, remove, modify, rename, examples, build examples in a very easy way, 15:44.560 --> 15:50.000 and then another point it's a bindings, at some point I noticed that many people was creating 15:50.000 --> 15:54.800 bindings for Riley to many, many languages, actually there is even people that are creating new 15:54.800 --> 16:02.880 languages, and as a test, they bind Riley for the new language, so it has many, many bindings, 16:02.880 --> 16:10.720 even some languages like Oding or Zik include Riley as a part of the standard packages available 16:10.720 --> 16:17.360 for users of the language, well for that reason I also created a support tool at some moment 16:17.360 --> 16:24.720 to simplify binding's creation, in what way? Well, RL parser just parses all the Riley 16:24.720 --> 16:30.560 header and generates a semantic file with a function name, number of parameters type of parameters, 16:30.560 --> 16:38.240 it can be exported into TXT, JSON, XML, multiple formats to simplify the bindings generation, 16:38.240 --> 16:43.680 the automation, automation of binding generation for the user creating a maintaining bindings. 16:45.760 --> 16:54.800 About motivation, a question, how long can motivation last? Well, it's difficult to stay 16:54.800 --> 17:01.680 motivated all the time, especially with an open source project, but my recommendation is that 17:03.680 --> 17:10.080 your primary source of motivation for a project must be that you enjoy genuinely what you do. 17:10.080 --> 17:16.320 I think that's the most important thing, when you stop enjoying what you do, even you can try 17:16.320 --> 17:23.040 many things or try to get money or whatever, it's difficult to stay motivated. 17:24.000 --> 17:32.560 Say that, there have been some elements that have kept me motivated more or less along the years. 17:35.120 --> 17:43.520 First, users, users, I created Riley for my students and as soon as I start using it in my classes, 17:43.520 --> 17:49.280 I start receiving a lot of feedback from them, they tell me, oh, I like that, I don't like that, 17:49.280 --> 17:56.320 better do that in that way. Well, they start giving me a lot of feedback, but also they start showcasing 17:56.320 --> 18:03.920 many amazing projects created by them. I think that's very motivating, or at least for me, 18:03.920 --> 18:10.080 that was very motivating and was one of the reasons to keep working on the library. 18:10.080 --> 18:15.920 Another source of motivation for me, were a site projects. When you keep working in a project 18:15.920 --> 18:23.680 for so long, at some point, it feels kind of repetitive and just reviewing bugs, fixing issues, 18:23.680 --> 18:27.680 it's not that motivating. For me, it's way more motivating, creating new things, 18:28.880 --> 18:36.000 having new ideas and creating new libraries. So I decided to, as I got that base library, 18:36.000 --> 18:42.000 I decided to start creating tools using the library, and it was motivating to create the tools, 18:42.000 --> 18:48.480 and I was also growing an ecosystem around Riley. Well, another source of motivation, 18:48.480 --> 18:56.720 are recognitions. When you receive some award, some recognition, or some people ask you for an 18:56.720 --> 19:04.000 interview, or participating in some podcasts, or go to some event, that's really motivating, 19:04.000 --> 19:11.200 and that also helped me to stay motivated while working on the library. And finally, 19:11.200 --> 19:18.240 I think it's also important for motivation, self-care. In my case, it helped me to have 19:18.240 --> 19:23.600 no-screen hobbies, to have some hobbies that I wish I knew right now, that are not related 19:23.600 --> 19:31.200 to working on a screen, and are more related to doing kind of hand work. For example, 19:31.200 --> 19:38.160 about the site projects, the Riley ecosystem, well, here is Riley, but around the years, 19:38.160 --> 19:45.520 I created all those support libraries for UIs, for PNG management, for initialization files, 19:45.520 --> 19:52.560 and all those tools to do many, many things. I like to investigate different fields, 19:52.560 --> 19:57.440 and while creating those tools, I had the opportunity to investigate about many, many different fields, 19:58.400 --> 20:09.360 sound generator, textur packaging, UI generation, well, text editors, the timeline creator 20:09.360 --> 20:15.760 that it's what you have seen in a previous slide. So, I created a bunch of tools and 20:15.760 --> 20:21.520 support libraries, most of them free and open source, but some of them also commercial, I'm private. 20:22.480 --> 20:28.320 About self-care, those are some of my hobbies, I play guitar, I really enjoy playing guitar, 20:28.320 --> 20:36.480 I like cooking, and lately I've got a new hobby that it's a good work and restoring things, 20:37.200 --> 20:45.200 well, here I restore a bunch of hammers that I got at home, so if that's what I did last summer, 20:45.280 --> 20:52.160 it was really refreshing and gave me motivated to work later on on computer again. 20:54.400 --> 20:58.880 Finally, let's talk about the third challenge that has been morning. 21:02.480 --> 21:08.320 That's a question that I asked myself many times while working in Riley, how long could I keep working 21:09.280 --> 21:13.200 in Riley's course for many, many years, it has been very, very difficult. 21:14.400 --> 21:20.800 There has been moments that I've been working full-time, that has been getting some contract work 21:20.800 --> 21:26.000 or working at the university teaching to get some money to then allow me to work for some 21:26.000 --> 21:34.240 months full-time again on Riley, but it has been very difficult to keep doing that without a secure 21:34.320 --> 21:46.160 source of income along the time. My experience as a solid developer, it's extremely, extremely 21:46.160 --> 21:55.440 difficult to make money with open software, that's my experience. I imagine that maybe if you are 21:55.440 --> 22:03.840 a corporate or a big foundation, it would be easier to get funds, but as a solid developer, 22:03.840 --> 22:13.280 it's really difficult. What were my sources of money? Well, first it was donations. I'm getting 22:13.280 --> 22:22.240 donations from people around the world in several platforms. It's not enough to live from that, 22:22.720 --> 22:32.400 obviously. Along the time, I got some grants and lately I got those NNF funds that really 22:32.400 --> 22:39.280 helped me for the first time in 12 years to be able to work in the project full-time with a kind of 22:39.280 --> 22:46.560 salary and also some corporate sponsors that also for the first time in 12 years, I managed to get 22:46.640 --> 22:54.560 two corporate sponsors in the last four months or so, that it was also very, very nice. And then 22:55.920 --> 23:04.400 I also try to sell tools and services, but it's very difficult on the amount of money that I got 23:04.400 --> 23:15.440 from that was really, really low to live from that. But even managing to get some money, even 23:15.440 --> 23:21.440 managing to get some funds for your project, I think that's a solid developer, there are some 23:21.440 --> 23:29.440 other challenges. First, the uncertainty, how long will donation last? Maybe one month you receive 23:29.440 --> 23:35.040 a lot of money or enough money to live and another month you do not receive a send. So, 23:35.360 --> 23:48.400 one minute. Okay, well, the little entity, how you manage to receive that money if you are 23:48.400 --> 23:54.800 not a foundation or a corporate, like a salary, a stability, you cannot plan for a future and 23:54.800 --> 24:02.720 all those concerns make you ask you if I should be able to find a rule of job. So, some point to 24:02.720 --> 24:09.840 remember, just to earn simplicity, I think it's key for long-term maintenance. I recommend 24:09.840 --> 24:15.520 showing your project and build a community, I think it helped me a lot, so it's a recommendation. 24:16.480 --> 24:20.800 Money could be a bonus, but it should not be a goal, because it's really difficult to get money. 24:21.360 --> 24:28.240 And the most important point for me is that you should enjoy genuinely what you do. So, that 24:28.240 --> 24:38.240 would be the take-out ways from this talk. And, yep, that would be everything. Thank you very much.