WEBVTT 00:00.000 --> 00:14.720 Thanks for having me, it's an honor and pleasure to be here again at our community event. 00:14.720 --> 00:21.000 Thanks to Udo and also to Alex for organizing the bedroom this year and let me get into 00:21.000 --> 00:22.000 it. 00:22.000 --> 00:25.000 So today I will be talking about Helena West. 00:26.000 --> 00:30.000 Let me start with a short interaction of myself. 00:30.000 --> 00:33.000 It's not so important here. 00:33.000 --> 00:37.000 It just shows that I have user space components. 00:37.000 --> 00:39.000 It's a multiple form system. 00:39.000 --> 00:44.000 We can't really support these major hardware architectures. 00:44.000 --> 00:47.000 And it's designed and implemented from scratch. 00:47.000 --> 00:53.000 So basically the idea is to test some interesting, 00:53.000 --> 01:00.000 radical new ideas like our asynchronous IPC mechanism with integrated shared memory. 01:00.000 --> 01:06.000 Also, most of our components are natively written for Helena West, 01:06.000 --> 01:08.000 not supported from other systems. 01:08.000 --> 01:12.000 So there are no frankened components so to speak with, you know, 01:12.000 --> 01:15.000 ugly compatibility layers around it. 01:16.000 --> 01:25.000 We also would like to say that our code base is readable and understandable to newcomers. 01:25.000 --> 01:35.000 I'm obviously too biased to really judge, but one objective measure is that 30% of the source code is commands. 01:35.000 --> 01:42.000 And maybe compare to some other of my kernel-based systems that we will hear today during the bedroom. 01:42.000 --> 01:45.000 Helena West is an integrated distribution. 01:45.000 --> 01:57.000 So we just download an image that already contains all the required components to get working desktop-like system. 01:57.000 --> 02:00.000 You don't have to, you know, manually configure stuff. 02:00.000 --> 02:06.000 It will just run things and also discover things and stuff like that. 02:06.000 --> 02:16.000 So what has been new in Helena West since our last status update here at this room in 2020? 02:16.000 --> 02:26.000 Well, the major improvement is the new graphical user interface or redesign and re-implementation of the previous one. 02:26.000 --> 02:32.000 So we have a completely new display server that is more open-ended. 02:32.000 --> 02:39.000 It should support accelerated direct and indirect rendering. 02:39.000 --> 02:44.000 And it also provides better performance in case you render in software only. 02:44.000 --> 02:50.000 It has additional features like configurable, double buffering and stuff like that. 02:50.000 --> 02:55.000 Also there are many more graphical widgets and dialogues. 02:55.000 --> 03:04.000 And one interesting thing is that almost all of those widgets and dialogues can be also used in text mode. 03:04.000 --> 03:08.000 So you can design a user interface using the same widgets. 03:08.000 --> 03:16.000 Whether the final application is going to run in a graphical mode or in a text mode. 03:16.000 --> 03:21.000 Proportion forms is just one, one addition to that. 03:21.000 --> 03:25.000 Then we did some usability improvements. 03:25.000 --> 03:27.000 So to speak. 03:27.000 --> 03:29.000 So there is a font editor. 03:29.000 --> 03:31.000 There is a two panel file manager now. 03:31.000 --> 03:39.000 There is a task bar with a start button and a start menu which you might know from some other rarely used operating systems. 03:39.000 --> 03:41.000 There is multi-seat support. 03:41.000 --> 03:45.000 The terminal can now scroll back and be resized. 03:45.000 --> 03:49.000 And quite recently there is an intelligence. 03:49.000 --> 03:50.000 I mean, intelligence. 03:50.000 --> 03:53.000 It sounds that it actually follows dependencies. 03:53.000 --> 03:55.000 Shutdown mechanism. 03:55.000 --> 03:58.000 So you can cleanly shut down the system now. 03:58.000 --> 04:06.000 We have improved our compiler to change to be currently completely up to date. 04:06.000 --> 04:09.000 With the latest GCC and stuff. 04:09.000 --> 04:16.000 And we have improved some standards of standard compliance on the API side. 04:16.000 --> 04:20.000 And finally, but maybe last, not least. 04:20.000 --> 04:23.000 There is some additional hardware support. 04:23.000 --> 04:27.000 We have improved the performance of the block device drivers. 04:27.000 --> 04:35.000 And after 20 years we lay a fresh eye on the kernel. 04:35.000 --> 04:39.000 And did some streamlining. 04:39.000 --> 04:41.000 It's not the rewrite. 04:41.000 --> 04:43.000 It's not the redesign. 04:43.000 --> 04:46.000 The basic bits and pieces are still there. 04:46.000 --> 04:55.000 But we simply allow some things that were not being so heavily used after so many years to just disappear. 04:55.000 --> 04:58.000 Like some unused extension points. 04:58.000 --> 05:01.000 We did some performance optimizations. 05:01.000 --> 05:05.000 We have improved the abstractions on some APIs. 05:05.000 --> 05:11.000 So for example, some of the kernel internals no longer use directly the weight queue. 05:11.000 --> 05:18.000 As the primary synchronization means, but use a semaphore, which is of course built by the weight queue. 05:18.000 --> 05:20.000 But it's still more abstract. 05:20.000 --> 05:23.000 And therefore the code is easier to understand. 05:23.000 --> 05:29.000 And we have removed some synchronization points, which were generally unnecessary. 05:29.000 --> 05:36.000 Because the actual workload was more of a CPU local. 05:36.000 --> 05:44.000 To compare this list, which is of course the stuff that we still do support with the previous list. 05:44.000 --> 05:47.000 It's sort of not quite comparable. 05:47.000 --> 05:49.000 So what's wrong? 05:49.000 --> 05:55.000 Well, this is another view of the story. 05:55.000 --> 06:03.000 And of course I know that the number of commits is not the best metric for everything. 06:03.000 --> 06:10.000 But it shows something that it definitely shows that since the last update, 06:10.000 --> 06:14.000 there was not so much activity as there used to be. 06:14.000 --> 06:17.000 Let me give you another picture. 06:17.000 --> 06:26.000 So previously we were exploring very interesting things going, you know, 06:26.000 --> 06:31.000 trying to solve some challenges. 06:31.000 --> 06:33.000 And now we are living in this serene valley. 06:33.000 --> 06:38.000 Now I'm here where nothing much matters. 06:38.000 --> 06:40.000 I'm nothing much happens. 06:40.000 --> 06:43.000 The river is flowing quite steadily. 06:43.000 --> 06:46.000 So this is where Helena was is now. 06:46.000 --> 06:56.000 And this is, I believe, the reason why those two lists of new features, you know, sound different. 06:56.000 --> 06:59.000 And Helena was essentially complete. 06:59.000 --> 07:04.000 And now you might say, hey, come on Martin, are you kidding? 07:04.000 --> 07:08.000 I mean, how can you claim that Helena was is complete? 07:08.000 --> 07:16.000 I still can't use it as a daily driver to replace my, my Linux or maybe at least Windows. 07:16.000 --> 07:17.000 No, sorry. 07:17.000 --> 07:21.000 But you know, there are two at least two meanings of the word complete. 07:21.000 --> 07:24.000 So you might be thinking about this. 07:24.000 --> 07:34.000 So starting with something, something that sort of works, but it's very, very single purpose, 07:34.000 --> 07:42.000 like a doghouse and completing it into a large city that can host almost anything. 07:42.000 --> 07:49.000 And a number of use cases that could be partially solved, contained or self sufficient. 07:49.000 --> 07:52.000 Yeah, this is one meaning of the word complete. 07:52.000 --> 07:58.000 And obviously, Helena was does is, I mean, this does not apply to Helena was. 07:58.000 --> 08:02.000 What I mean by Helena was is complete is this. 08:02.000 --> 08:14.000 So moving from a family house that is in construction where, you know, walls need to be put and furniture needs to be placed in. 08:14.000 --> 08:19.000 And electricity needs to be wired and whatever. 08:19.000 --> 08:25.000 And ending with a family house that can be, that can be considered as finished. 08:25.000 --> 08:27.000 I mean, there is nothing missing. 08:27.000 --> 08:30.000 But it's still not self sufficient. 08:30.000 --> 08:33.000 I mean, it still needs something else. 08:33.000 --> 08:36.000 So in this sense, Helena was is complete. 08:36.000 --> 08:41.000 And this is, this is the, the thing, you know, 08:41.000 --> 08:46.000 Helena was a pure open source project. 08:46.000 --> 08:49.000 It's completely community driven. 08:49.000 --> 08:56.000 Meaning that it's driven by individual interests of contributors. 08:56.000 --> 09:01.000 Maybe by their academic research or maybe by their education. 09:01.000 --> 09:09.000 I mean, that there have been something like 30 masterpieces written around or on top of Helena was. 09:09.000 --> 09:15.000 We did have some corporate sponsorship like Google summer of code or the season Nick. 09:15.000 --> 09:18.000 I'm playing one of our contributors. 09:18.000 --> 09:20.000 But there was never a business goal. 09:20.000 --> 09:22.000 There was never a monetization goal. 09:23.000 --> 09:27.000 Meaning, since there are no major building blocks missing. 09:27.000 --> 09:29.000 I mean, there is a file system. 09:29.000 --> 09:31.000 There is sound support. 09:31.000 --> 09:32.000 There's not working. 09:32.000 --> 09:39.000 The only bits and pieces that are currently missing and that would lead to the completeness in the other sense. 09:39.000 --> 09:43.000 In the completeness of, you know, running new Linux. 09:43.000 --> 09:49.000 Is all the device drivers file systems, standard APIs, usability features. 09:49.000 --> 09:55.000 And it turns out, at least in our setting, this is no longer rewarding. 09:55.000 --> 10:02.000 I mean, at least no longer so rewarding as it used to be before when you can say, 10:02.000 --> 10:05.000 I have implemented a networking stack. 10:05.000 --> 10:12.000 I have filled in the, the large building block that was missing in the system. 10:12.000 --> 10:15.000 And is it a cautionary tile? 10:15.000 --> 10:17.000 I believe it is. 10:17.000 --> 10:24.000 I mean, do you remember, Linux 3, that used to be very famous that had this. 10:24.000 --> 10:29.000 They might end up being complete in the sense that that I have described. 10:29.000 --> 10:31.000 And not complete in the sense. 10:31.000 --> 10:39.000 And a very user would use an operating system or would like an operating system to work. 10:39.000 --> 10:42.000 And again, there's nothing wrong with that. 10:42.000 --> 10:48.000 But you either have to, you know, expect that is going to happen and embrace it. 10:48.000 --> 10:52.000 This is what we do in how an OS, at least at the moment. 10:52.000 --> 11:01.000 Or if you don't like this outcome, you probably need to have a plan and explicit well fought plan to avoid it. 11:01.000 --> 11:07.000 And at the end of the day, everything boils down to money and time. 11:07.000 --> 11:14.000 So you probably just need some reasonable long-term source of revenue, 11:14.000 --> 11:19.000 whether it's a paying customer or some paying use case or charity. 11:19.000 --> 11:27.000 I don't know what it is, but it needs to be there to cover the, the, the costs of the, 11:27.000 --> 11:35.000 of, you know, finishing those tasks that are no longer fun, like implementing a new subsystem. 11:36.000 --> 11:43.000 So that's, that's all what I wanted to say about this completeness issue. 11:43.000 --> 11:47.000 So what about Helen OS and our vision into the next 20 years? 11:47.000 --> 11:50.000 I mean, so far we are embracing the Syrian Valley. 11:50.000 --> 11:52.000 There are still contributions. 11:52.000 --> 11:59.000 There are still, it's still a valuable tool for me and for many other people to test new stuff, 11:59.000 --> 12:03.000 to debug other systems, for example. 12:04.000 --> 12:07.000 And there is some value in the codebase itself. 12:07.000 --> 12:13.000 So we can stay in that Syrian Valley for, for some time. 12:13.000 --> 12:17.000 And we are still on a lookout for some new challenges. 12:17.000 --> 12:25.000 There are people who still have this, I don't know if I should say a pipe dream or a weather dream or whatever kind of dream, 12:25.000 --> 12:32.000 that there will be a company around Helen OS and that will push it into new areas. 12:32.000 --> 12:36.000 But somebody needs to step up. 12:36.000 --> 12:45.000 The asynchronous will be more important than ever, at least according to the development of the hardware. 12:45.000 --> 12:48.000 So I believe there is something to build on. 12:48.000 --> 12:57.000 There are even many features that have not been integrated yet into the mainline branch of Helen OS. 12:58.000 --> 13:00.000 That could be useful for somebody. 13:00.000 --> 13:06.000 So just to point to, to examples, there is task, check pointing and migration implemented. 13:06.000 --> 13:12.000 There is a manual, less support implemented, just not in the main branch. 13:12.000 --> 13:18.000 So if anybody has an idea, how to use that, just let me know. 13:18.000 --> 13:22.000 So to sum up, Helen OS is still alive and well. 13:22.000 --> 13:27.000 Helen OS is complete, but in the sense that I have described. 13:27.000 --> 13:32.000 There are still many ways how to contribute to it and it makes sense in some ways, 13:32.000 --> 13:35.000 because we do things differently. 13:35.000 --> 13:42.000 And Helen OS is not going to move away, at least until there is still somebody contributing to it. 13:42.000 --> 13:45.000 And there is nothing wrong with being in that Syrian Valley. 13:45.000 --> 13:53.000 All I'm trying to say is that many existing projects, very ambitious projects, 13:53.000 --> 13:59.000 might eventually end up in the Syrian Valley as well, unless they do something against it, 13:59.000 --> 14:01.000 if they don't want to be there. 14:01.000 --> 14:03.000 So have a plan. 14:03.000 --> 14:05.000 And that's it from me. 14:05.000 --> 14:06.000 Thank you. 14:06.000 --> 14:07.000 And are there any questions? 14:07.000 --> 14:14.000 Does it have a real mercy? 14:14.000 --> 14:15.000 No. 14:15.000 --> 14:17.000 So I should repeat the question. 14:17.000 --> 14:19.000 The question most does it have a real mercy? 14:19.000 --> 14:23.000 No, but it has tattoos. 14:23.000 --> 14:24.000 Any other questions? 14:24.000 --> 14:25.000 Yes. 14:25.000 --> 14:26.000 I believe you are first. 14:26.000 --> 14:36.000 So the question is about the manuals stuff. 14:36.000 --> 14:40.000 Yes, it essentially means that it can run. 14:40.000 --> 14:46.000 I mean, I'm not quite sure if every microcontroller would be a good fit, 14:46.000 --> 14:52.000 because there are other requirements for the payroll size and whatever, 14:52.000 --> 14:53.000 but generally yes. 14:53.000 --> 14:56.000 It could run on some on a CPU that does not have a memory. 14:56.000 --> 14:59.000 It does not have any page and stuff like that. 14:59.000 --> 15:01.000 I can give you the details later. 15:01.000 --> 15:02.000 Thanks. 15:02.000 --> 15:03.000 That was still. 15:03.000 --> 15:05.000 Yes, though. 15:05.000 --> 15:07.000 I'm sorry. 15:07.000 --> 15:09.000 Can you can. 15:09.000 --> 15:13.000 The envelope running on top of Helenowas. 15:13.000 --> 15:16.000 Not not yet. 15:16.000 --> 15:20.000 I mean, so the question is how to get and we are running on Helenowas. 15:20.000 --> 15:23.000 Not at the moment there is no virtualization support at this moment. 15:23.000 --> 15:26.000 The answer is complicated. 15:26.000 --> 15:27.000 All right. 15:27.000 --> 15:28.000 Thank you. 15:28.000 --> 15:29.000 Thank you.