[00:00.000 --> 00:12.000]  Okay, all right, is the mic working?
[00:12.000 --> 00:13.000]  Yeah.
[00:13.000 --> 00:14.000]  Yeah.
[00:14.000 --> 00:15.000]  All right.
[00:15.000 --> 00:16.000]  Great.
[00:16.000 --> 00:17.000]  Okay.
[00:17.000 --> 00:18.000]  All right.
[00:18.000 --> 00:19.000]  Welcome, everyone.
[00:19.000 --> 00:22.800]  I'm going to talk about a program that I've started at Red Hat called Kernel Development
[00:22.800 --> 00:29.360]  Learning Pipeline that I run with a small group of engineers kind of as a side project.
[00:29.360 --> 00:35.280]  So first, I'll talk about what it is, first of all, why it's a good idea, why we think
[00:35.280 --> 00:40.000]  it's a good idea and we're doing it, a little bit about the program and some of the growth
[00:40.000 --> 00:46.760]  that we've had in the program over the past year, two years approximately, and then conclude
[00:46.760 --> 00:50.960]  with some resources about the program.
[00:50.960 --> 00:53.240]  So first of all, what is KDLP?
[00:53.240 --> 00:57.760]  So that stands for Kernel Development Learning Pipeline, credit for that acronym to Julia
[00:57.760 --> 01:05.280]  Denham, actually, and we are building a comprehensive pipeline for Linux kernel talent and low-level
[01:05.280 --> 01:11.200]  talent more generally, because as it turns out, there's not a lot of younger people getting
[01:11.200 --> 01:12.200]  into the kernel.
[01:12.200 --> 01:18.520]  It's not an extremely popular area of study, it's barely really taught in school.
[01:18.520 --> 01:24.920]  So right now, the main component is we have a course that we're teaching at UMass Lowell
[01:24.920 --> 01:28.440]  and also on the Linux Foundation's platform.
[01:28.440 --> 01:33.720]  We are trying to recruit interns from this course and from this program and through kind
[01:33.720 --> 01:35.360]  of our network.
[01:35.360 --> 01:39.040]  We bring people in and then they serve as TAs and help develop the course and improve
[01:39.040 --> 01:40.880]  the content.
[01:40.880 --> 01:46.240]  And then ideally, we bring them in and recruit them full-time, that's the goal.
[01:46.240 --> 01:48.640]  So yeah, why is this a good idea?
[01:48.640 --> 01:54.600]  So like I was saying, a lot of senior Linux kernel engineers are getting somewhat close
[01:54.600 --> 01:55.600]  to retirement.
[01:55.600 --> 01:59.800]  That's kind of, definitely a much higher, I think, average age in kernel and in like
[01:59.800 --> 02:05.280]  low-level engineering, at least in the US that I've noticed, than in other areas of
[02:05.280 --> 02:06.280]  software engineering.
[02:06.280 --> 02:11.480]  And it takes a very long time for people to learn the Linux kernel, especially today,
[02:11.480 --> 02:16.640]  because a lot of people don't even learn C in school and they're generally separated
[02:16.640 --> 02:20.280]  from a lot of the low-level computer science concepts.
[02:20.280 --> 02:25.040]  Things that people 20, 30, 40 years ago, so I'm told, were learned kind of as a standard
[02:25.040 --> 02:26.040]  thing.
[02:26.040 --> 02:30.240]  Today, it's sort of a niche topic.
[02:30.240 --> 02:35.560]  There are people getting into it, but only kind of in-ditch communities at certain schools
[02:35.560 --> 02:40.800]  if you're exposed to certain people or certain online communities.
[02:40.800 --> 02:43.560]  So the kernel itself is barely taught.
[02:43.560 --> 02:46.680]  People take maybe one class in opening systems and a lot of the tools, like some of the more
[02:46.680 --> 02:50.000]  advanced get usage, for example.
[02:50.000 --> 02:51.360]  It's difficult to learn that kind of stuff.
[02:51.360 --> 02:54.760]  How do people learn how to do email patches without trying and failing and just getting
[02:54.760 --> 02:55.760]  roasted online?
[02:55.760 --> 02:58.160]  It's kind of difficult.
[02:58.160 --> 03:02.080]  So for a lot of companies, it's difficult to find talent.
[03:02.080 --> 03:07.160]  I mean, I know at Red Hat, it can be somewhat difficult to find people who have some Linux
[03:07.160 --> 03:11.000]  kernel knowledge and are actually interested in doing it.
[03:11.000 --> 03:14.240]  There's lots of people who are capable of doing it and just maybe don't like it that
[03:14.240 --> 03:15.240]  much.
[03:15.240 --> 03:20.320]  So, of course, the way that we're pitching this to Red Hat and the way that it's a good
[03:20.320 --> 03:24.240]  idea for any company is that we think we can bring more value to the company for a lower
[03:24.240 --> 03:25.240]  cost.
[03:25.240 --> 03:29.680]  But for the community, what we are trying to bring is to train the next generation of
[03:29.680 --> 03:34.840]  developers from all sorts of different backgrounds, as opposed to just people who happen to bump
[03:34.840 --> 03:38.320]  into somebody at a recruiting event, which is how I got into the kernel.
[03:38.320 --> 03:42.480]  I just happened to run into somebody at an event that he wasn't even supposed to be
[03:42.480 --> 03:45.600]  at, in fact.
[03:45.600 --> 03:48.080]  So a little overview.
[03:48.080 --> 03:51.480]  So there's kind of three main sections of the program.
[03:51.480 --> 03:57.920]  So the first is this kernel development course that we've developed that is taught at UMass
[03:57.920 --> 03:58.920]  Lool.
[03:58.920 --> 04:02.840]  So we've created this from scratch using, originally based on the Linux device drivers
[04:02.840 --> 04:05.600]  third edition book, but we've kind of gone further from that.
[04:05.600 --> 04:09.760]  That's actually one of the latest books on Linux kernel engineering and it's published
[04:09.760 --> 04:12.920]  in 2005.
[04:12.920 --> 04:18.040]  But yeah, so we've been working on developing our own curriculum with a number of different
[04:18.040 --> 04:26.440]  labs and some slide decks and kind of a whole, you know, a set of things that people could
[04:26.440 --> 04:28.600]  do to learn the kernel.
[04:28.600 --> 04:32.560]  And then from there, we tried to bring people into internships as we're able to do that
[04:32.560 --> 04:38.120]  or other kind of more second level or hands on experiences where they can further develop
[04:38.120 --> 04:42.920]  those skills and then ideally bring them in as full time engineers at Red Hat or I mean
[04:42.920 --> 04:46.000]  the beauty of open source is if they get recruited somewhere else or if they go work
[04:46.000 --> 04:49.280]  somewhere else in the open source community, that's still, that's a win for everyone in
[04:49.280 --> 04:50.280]  the community.
[04:50.280 --> 04:52.960]  It's a win for all the companies working with the community because everyone working on
[04:52.960 --> 04:57.160]  the kernel is working on the kernel.
[04:57.160 --> 04:59.200]  So a little more about the course.
[04:59.200 --> 05:02.520]  The goal is to introduce students to Linux kernel development.
[05:02.520 --> 05:05.280]  You know, obviously we can't teach them the entire kernel.
[05:05.280 --> 05:09.520]  We can barely even introduce them to one subsystem, but the goal is to teach them kind of what
[05:09.520 --> 05:14.080]  they need to learn to teach themselves and to teach them what they need to know to work
[05:14.080 --> 05:19.080]  in one of these open source communities, specifically the Linux kernel community, but those skills
[05:19.080 --> 05:24.840]  I think are pretty applicable to other kind of more niche open source areas that can be
[05:24.840 --> 05:28.400]  somewhat intimidating for a lot of people to get into.
[05:28.400 --> 05:31.160]  We mostly just require C language skills.
[05:31.160 --> 05:34.480]  I mean that's obviously the most important thing when you're working in the kernel.
[05:34.480 --> 05:41.120]  That's a basic Linux experience, some programming experience in general, but not need a ton
[05:41.120 --> 05:46.680]  of Linux kernel experience to get started, but really the most critical thing to see.
[05:46.680 --> 05:54.480]  So we teach kind of an overview of just various kernel features and subsystems, kind of describe
[05:54.480 --> 06:02.280]  like a map of the kernel, what there is, what, you know, how the different pieces connect
[06:02.280 --> 06:05.280]  together, how you can interact with it.
[06:05.280 --> 06:10.360]  We approach through device driver development, that's a good way to get people into, you
[06:10.360 --> 06:14.040]  know, to work with a lot of the different APIs, because to write a device driver you
[06:14.040 --> 06:18.520]  need to interact with a lot of different areas and you end up building something that's somewhat
[06:18.520 --> 06:24.480]  kind of complete piece that works on its own, and then we introduce people to more advanced
[06:24.480 --> 06:25.480]  usage of Git.
[06:25.480 --> 06:27.800]  I mean people may be familiar with GitHub, but you know, believe it or not, a lot of
[06:27.800 --> 06:31.000]  people don't know the difference between Git and GitHub.
[06:31.000 --> 06:36.140]  So we help out with that and get them started with email patches and rebases and things
[06:36.140 --> 06:37.640]  like that.
[06:37.640 --> 06:42.000]  We also talk about BPF Trace and other tracing, which are things that they, I don't think
[06:42.000 --> 06:46.760]  BPF Trace or BPF in general is taught in any school that I know of, at least not in undergrad
[06:46.760 --> 06:52.400]  courses, you know, NCSCope and other ways to explore large projects and repositories,
[06:52.400 --> 06:57.600]  because you know, in a computer science program people generally don't get exposed to working
[06:57.600 --> 06:59.720]  on like large ongoing projects.
[06:59.720 --> 07:04.160]  You know, you work on some, you know, one and done thing, and you know, then you send
[07:04.160 --> 07:06.880]  it in and no one else looks at it and you get your grade and you move on, right?
[07:06.880 --> 07:11.520]  So we're introducing people to work on larger code bases, which is something that's somewhat
[07:11.520 --> 07:13.560]  unusual for a university course.
[07:13.560 --> 07:17.960]  So of course, all these course materials and the assignments completely open source.
[07:17.960 --> 07:22.920]  We've linked it, it's all on our website, and we're continuously improving it.
[07:22.920 --> 07:27.360]  You know, the source for the website is also on GitHub, you know, if anyone has any suggestions
[07:27.360 --> 07:30.440]  or changes, we're always open to them.
[07:30.440 --> 07:34.240]  So we have undergraduate and graduate level university courses that run at UMass Lowell
[07:34.240 --> 07:36.960]  that we've created over the past couple of years.
[07:36.960 --> 07:41.320]  It became kind of a full-fledged course in the fall semester of last year.
[07:41.320 --> 07:43.000]  So it's relatively new.
[07:43.000 --> 07:47.560]  And this semester we were also running it as a kind of dual program with the Linux Foundation's
[07:47.560 --> 07:50.000]  mentorship platform.
[07:50.000 --> 07:53.960]  So we have people applying on there to, I will also talk a little bit more about that
[07:53.960 --> 07:54.960]  later.
[07:54.960 --> 08:01.560]  We run through the course and kind of work with at the same schedule and at the same
[08:01.560 --> 08:05.840]  pace as the people at UMass Lowell, but they're from around the world and they don't need
[08:05.840 --> 08:07.200]  to be enrolled as students.
[08:07.200 --> 08:09.040]  They can come from anywhere, right?
[08:09.040 --> 08:13.560]  And kind of the bottom line is we try to reimagine how to teach this kind of material from first
[08:13.560 --> 08:16.920]  principles because it doesn't really make sense to teach it from a typical, like memorize
[08:16.920 --> 08:20.560]  this and do the exam kind of perspective, right?
[08:20.560 --> 08:25.820]  The goal is to get people to be able to explain things in their own words and to work on open
[08:25.820 --> 08:29.320]  source projects and work in open source communities.
[08:29.320 --> 08:33.760]  So we've replaced exams with presentations where they actually explain the work that
[08:33.760 --> 08:38.480]  they've done in their own words to their peers.
[08:38.480 --> 08:43.480]  So from there we recruit people as interns and try to bring them in.
[08:43.480 --> 08:50.440]  So we find people in the course who are enthusiastic and capable about so by the time we bring
[08:50.440 --> 08:53.560]  them in to the internship, they also have a lot of prerequisite knowledge that they
[08:53.560 --> 08:55.320]  would need.
[08:55.320 --> 08:59.400]  And so they're able to hit the ground running and do a lot more in their three to six month
[08:59.400 --> 09:05.160]  period or longer than they otherwise would have because actually, so I had a, the manager
[09:05.160 --> 09:09.440]  who hired me at Red Hat originally as a co-op, he said he didn't like to do summer internships
[09:09.440 --> 09:13.480]  in the kernel because by the time three months are over, they've just gotten on boarded.
[09:13.480 --> 09:18.000]  They've just started to understand how to work with a Linux kernel and then it's over
[09:18.000 --> 09:19.600]  and then they got to go.
[09:19.600 --> 09:21.000]  So he only really did co-ops.
[09:21.000 --> 09:24.360]  But the goal with this is if we kind of vet people ahead of time and we give them the
[09:24.360 --> 09:27.360]  skills that they need and give them kind of, you know, the get skills and the, you know,
[09:27.360 --> 09:30.800]  they compile the kernel and they have the background, they actually can get some value
[09:30.800 --> 09:36.000]  from an internship and we're more likely to bring people into internships who actually
[09:36.000 --> 09:39.080]  want to work in the kernel and then they're more likely to go off and, you know, and work
[09:39.080 --> 09:43.380]  in the kernel in general or work in open source communities.
[09:43.380 --> 09:47.560]  So we pair them up with rail engineers and, you know, they work on various initiatives
[09:47.560 --> 09:53.600]  within Red Hat that are, you know, that are, need people to learn these kind of new and
[09:53.600 --> 09:57.040]  old combinations of old and new skills.
[09:57.040 --> 10:01.320]  But the goal is to train these new people to work in those areas.
[10:01.320 --> 10:04.840]  So we also bring them in to help with our program and we've had people come in, they
[10:04.840 --> 10:07.760]  take the course and then they come in and they turn around and then they're TAs and
[10:07.760 --> 10:09.200]  they work on improving the course.
[10:09.200 --> 10:12.080]  Maybe they do a couple lectures here and there.
[10:12.080 --> 10:17.000]  They do some grading and they kind of service TAs.
[10:17.000 --> 10:22.080]  So another thing that we are doing is we have this kind of new kernel devs group within
[10:22.080 --> 10:24.680]  Red Hat, which I thought I'll just briefly mention.
[10:24.680 --> 10:28.840]  It's a little bit tangential, but, you know, once people get into Red Hat and when people
[10:28.840 --> 10:35.520]  are working in the kernel, especially in a remote job, which a lot of these positions
[10:35.520 --> 10:41.560]  are, people get very separated, they get very siloed, there's not a lot of socialization.
[10:41.560 --> 10:46.160]  So we run a group to, you know, just bring people together with, for a meeting with very
[10:46.160 --> 10:47.160]  little agenda.
[10:47.160 --> 10:51.160]  You know, occasionally we have a presentation here and there, but just to talk about things
[10:51.160 --> 10:57.160]  and share concepts and tools and questions and, you know, it's a Red Hat specific thing,
[10:57.160 --> 11:01.120]  so they can ask about specific things related to their job and a place where, you know,
[11:01.120 --> 11:05.200]  there's, I mean, managers are allowed, but it's not a specific structured meeting for
[11:05.200 --> 11:06.200]  a business purpose.
[11:06.200 --> 11:10.160]  And I think people who, you know, you might get that in an office somewhere when you're
[11:10.160 --> 11:14.520]  just walking around and talking to people, but it's a little bit more difficult to get
[11:14.520 --> 11:15.880]  when you're in a remote position.
[11:15.880 --> 11:21.520]  So we also bring interns who we've recruited into that group, you know, and people from
[11:21.520 --> 11:25.640]  elsewhere within Red Hat who are interested in the kernel and they can ask questions and
[11:25.640 --> 11:30.840]  find resources and, you know, if they enjoy it, they can switch into the kernel.
[11:30.840 --> 11:34.280]  Because there are actually a number of people who are interested and they just don't know
[11:34.280 --> 11:35.280]  where to start.
[11:35.280 --> 11:39.980]  They're just not a lot of good resources.
[11:39.980 --> 11:45.440]  So now I'll talk about some of the growth of this program over the last couple of years.
[11:45.440 --> 11:51.160]  So first of all, we've partnered with Red Hat's main educational initiative, Red Hat Academy,
[11:51.160 --> 11:54.720]  which we found out about only through doing this program.
[11:54.720 --> 12:00.400]  But they work with universities mainly on delivering and facilitating kind of systems
[12:00.400 --> 12:06.400]  management, you know, what do you call it, system administration and cyber reliability
[12:06.400 --> 12:12.400]  engineering, kind of the more, you know, the standard kind of Red Hat certifications that
[12:12.400 --> 12:13.920]  they do.
[12:13.920 --> 12:15.680]  But they don't have much software engineering.
[12:15.680 --> 12:18.840]  In fact, I don't think they had any software engineering component at all.
[12:18.840 --> 12:24.400]  So they were happy to work with us and, you know, we're happy to work with them.
[12:24.400 --> 12:28.480]  And they made us these nice posters, which we have down here if anyone's interested.
[12:28.480 --> 12:33.840]  Or I guess they're called like leave behind sheets, technically.
[12:33.840 --> 12:39.880]  We also ran a workshop last year for interns and co-ops in Ireland.
[12:39.880 --> 12:44.000]  And that was just kind of every other week, kind of a casual thing, you know, over time
[12:44.000 --> 12:45.960]  we had fewer and fewer people, to be honest.
[12:45.960 --> 12:50.240]  But the people who stayed around really enjoyed it and I think hopefully some of them will
[12:50.240 --> 12:56.480]  end up working in the space or, you know, in open source generally, hopefully the feedback
[12:56.480 --> 13:00.160]  for them was generally fairly positive.
[13:00.160 --> 13:03.600]  We've also been connecting to various educational programs in different countries.
[13:03.600 --> 13:07.560]  A lot of this stuff's very preliminary, but we're hoping to kind of package what we've
[13:07.560 --> 13:12.480]  done in our course in a way that can be replicated at other places and other universities and
[13:12.480 --> 13:16.200]  other countries to bring people in from different places in the world.
[13:16.200 --> 13:20.080]  And then we've also partnered with the Linux Foundation, specifically their mentorship
[13:20.080 --> 13:24.480]  platform, which I will talk about next.
[13:24.480 --> 13:27.760]  So like I was saying, the Linux Foundation.
[13:27.760 --> 13:36.600]  So we have just put our course on the Linux Foundation's platform essentially as a mentorship,
[13:36.600 --> 13:39.680]  but it's the same thing as the course.
[13:39.680 --> 13:43.400]  We've worked with Shua Khan, she's been very helpful.
[13:43.400 --> 13:48.360]  She runs this Linux Foundation's mentorship platform and, you know, great person to get
[13:48.360 --> 13:52.600]  in touch with if you're interested in learning the kernel or have questions about this.
[13:52.600 --> 13:57.280]  She's also a kernel maintainer if you're not familiar.
[13:57.280 --> 14:03.160]  So we are running the same course, and this is kind of experimental, we're running the
[14:03.160 --> 14:08.720]  same schedule with the same assignments with a group who are doing the course for credit
[14:08.720 --> 14:15.560]  at UMass Lowell, and at the same time, people who are just, they just apply it online, and
[14:15.560 --> 14:16.720]  they're just doing it.
[14:16.720 --> 14:19.800]  There are people, you know, just on random continents, some of them are students, some
[14:19.800 --> 14:24.560]  of them are working, some of them are doing, you know, I have no idea, honestly, some of
[14:24.560 --> 14:31.120]  them are just probably, you know, having a good time, but they're doing the same assignments
[14:31.120 --> 14:35.600]  and on the same schedule, and, you know, they submit their assignments to a different mailing
[14:35.600 --> 14:42.040]  list than the UMass Lowell cohort, but they're in a shared Discord server and they're interacting,
[14:42.040 --> 14:46.960]  and, you know, we're seeing how that goes, and it's an extremely diverse group of mentees.
[14:46.960 --> 14:52.480]  We have people from, I think like five different continents, you know, people from various places
[14:52.480 --> 14:57.920]  in Europe, a couple in America, South America, someone in Mexico, like three or four from
[14:57.920 --> 15:00.360]  Africa, Nigeria, Kenya.
[15:00.360 --> 15:06.720]  So a very interesting group, a couple in India as well.
[15:06.720 --> 15:10.080]  So a couple of statistics on what we've been able to do.
[15:10.080 --> 15:14.400]  It's a fairly new program, but I am happy to show some results.
[15:14.400 --> 15:20.280]  We were able to hire two people full-time from internships who originally did the, went
[15:20.280 --> 15:21.780]  through the class.
[15:21.780 --> 15:23.720]  This was during a year that was very difficult to hire people.
[15:23.720 --> 15:29.000]  There was a hiring freeze, so I think two is still, it's a lot more than zero.
[15:29.000 --> 15:33.340]  Two people who did co-ops or internships with us were recruited to Amazon and Microsoft,
[15:33.340 --> 15:36.760]  but one of them liked Red Hat so much and liked working with Open Source so much better
[15:36.760 --> 15:41.240]  than Amazon that he is pretty interested in coming back, leaving Amazon and coming back
[15:41.240 --> 15:43.920]  to Red Hat, which is pretty cool.
[15:43.920 --> 15:47.640]  But overall, we've had seven interns and co-ops who have been trained and have gone
[15:47.640 --> 15:52.800]  through this particular program via this KDLP thing.
[15:52.800 --> 15:57.400]  And last semester, we had our biggest class, well, until this semester, we had 12 students,
[15:57.400 --> 16:01.640]  undergraduate, mostly graduate students, graduate and undergraduate courses are the same, which
[16:01.640 --> 16:04.920]  I found out, same content.
[16:04.920 --> 16:08.680]  And we have about a dozen, a couple more than a dozen students this semester, and the course
[16:08.680 --> 16:09.960]  is actively going on right now.
[16:09.960 --> 16:15.800]  It just started a couple of weeks ago, and I think about a dozen in the LFX mentorship
[16:15.800 --> 16:16.800]  as well.
[16:16.800 --> 16:23.680]  And this is by far the most diverse group by gender and location, which is pretty interesting.
[16:23.680 --> 16:27.680]  So now, just briefly, some program information.
[16:27.680 --> 16:30.000]  So the team is, well, it's me.
[16:30.000 --> 16:31.280]  Julie Denham is a program manager.
[16:31.280 --> 16:36.360]  I think she's in the chat answering questions, if anyone's interested, back in Boston.
[16:36.360 --> 16:41.000]  Charles Marabelle, he's the content lead for the course.
[16:41.000 --> 16:45.920]  And Dennis Alexandrov, he is an intern who's been extended and also TA.
[16:45.920 --> 16:50.880]  He's been working on the course since last summer, originally went through the course
[16:50.880 --> 16:54.480]  from McGill University.
[16:54.480 --> 16:55.480]  And yeah, acknowledgments.
[16:55.480 --> 17:00.640]  I just want to give a shout out to two Red Hatters in particular.
[17:00.640 --> 17:03.720]  First Heidi Dempsey, the research and innovation director in North America.
[17:03.720 --> 17:07.480]  She's been a longtime supporter of this program from the beginning, from like, you know, before
[17:07.480 --> 17:11.200]  it was even an idea, you know, when we were talking about doing something in, like, February
[17:11.200 --> 17:12.200]  2019.
[17:12.200 --> 17:20.240]  So her support's been great, and, you know, recommend, you know, doing research and innovation
[17:20.240 --> 17:22.960]  with her if anyone's, you know, interested in doing that.
[17:22.960 --> 17:26.960]  And of course, Mike McGrath, the vice president of RELL, who is the executive sponsor for
[17:26.960 --> 17:29.920]  this program within Red Hat.
[17:29.920 --> 17:32.520]  Now, a couple resources.
[17:32.520 --> 17:37.360]  We have a mailing list, which is linked up there, which we pretty much just used for
[17:37.360 --> 17:41.320]  giving a quarterly update newsletter, which we just started this quarter.
[17:41.320 --> 17:45.520]  And it's read only, so you're not going to get spammed with a ton of information.
[17:45.520 --> 17:50.360]  If you subscribe to that mailing list, you get invited to a weekly office hour session.
[17:50.360 --> 17:56.480]  That happens Tuesday at noon Eastern Standard Time, which just for asking questions about
[17:56.480 --> 18:01.320]  the program or the kernel in general, or you can just send me or one of us an email.
[18:01.320 --> 18:03.120]  And then on the right, we have our website.
[18:03.120 --> 18:08.520]  We have a section of our website that talks about the information, specifically the structure
[18:08.520 --> 18:10.720]  and content of the course.
[18:10.720 --> 18:14.040]  We have the page on the UMass Law Catalog that talks about the, you know, that just has
[18:14.040 --> 18:18.640]  a course description, so in case anyone's interested in checking that out.
[18:18.640 --> 18:21.600]  And the mentorship page on LFX.
[18:21.600 --> 18:26.080]  Then down there on the bottom right, we do have a crowdfunding page on LFX.
[18:26.080 --> 18:32.320]  If you want to support the program development and diverse engineers, it's a very, you know,
[18:32.320 --> 18:37.160]  we'll be able to bring people on through that platform as TAs, potentially give them, you
[18:37.160 --> 18:41.760]  know, bring them on and continue them kind of beyond what we're able to do with just
[18:41.760 --> 18:42.760]  Red Hat.
[18:42.760 --> 18:49.760]  All right, any questions?
[18:49.760 --> 19:08.280]  Hi, I think there is an offer and demand problem.
[19:08.280 --> 19:14.520]  If you consider every Linux developers, it will be like 1,000 JavaScript developers.
[19:14.520 --> 19:19.000]  And also people at the university, if you give them the choice between JavaScript and
[19:19.000 --> 19:23.280]  Linux, they will have more opportunities going with the JavaScript.
[19:23.280 --> 19:24.280]  They cannot do a mistake.
[19:24.280 --> 19:28.960]  However, if they go with the kernel, like, I don't know, in my country, for example, there's
[19:28.960 --> 19:33.620]  a lot of, there's a lot less jobs and you have less opportunities.
[19:33.620 --> 19:35.200]  So what's your opinion on this?
[19:35.200 --> 19:44.040]  Does people know the choice between like classic software development and kernel software development?
[19:44.040 --> 19:51.400]  So if I understand that correctly, you're saying there's fewer jobs in the kernel.
[19:51.400 --> 19:58.760]  And so, you know, what would I say about, just about what, it was like 1,000 to one
[19:58.760 --> 20:05.640]  or something, you said, well, I mean, that's true, but the competition for those jobs is
[20:05.640 --> 20:06.640]  much more intense.
[20:06.640 --> 20:11.200]  It's much kind of broader, you know, skill area.
[20:11.200 --> 20:15.600]  I mean, it's definitely, you know, I'm not condemning JavaScript developers.
[20:15.600 --> 20:16.600]  JavaScript's fine.
[20:16.600 --> 20:20.920]  It's necessary, you know, the web runs on it for all its problems.
[20:20.920 --> 20:26.760]  But it's also true that, I mean, working in the kernel, you know, the Linux kernel, there
[20:26.760 --> 20:33.640]  are fewer jobs, but there are a number of very solid jobs, and it doesn't, learning
[20:33.640 --> 20:38.480]  Linux kernel skills is somewhat transferable to other areas of development.
[20:38.480 --> 20:40.480]  You don't have to work as a Linux kernel developer.
[20:40.480 --> 20:44.680]  You don't have to, you know, go to Red Hat and work on RHEL and do backports or whatever.
[20:44.680 --> 20:45.680]  Right?
[20:45.680 --> 20:46.680]  You can kind of just do it on your own.
[20:46.680 --> 20:51.760]  You can do, you know, just some other low-level project and just having knowledge of the kernel
[20:51.760 --> 20:55.560]  and how it works and the knowledge necessary to work on the kernel, I think, could benefit
[20:55.560 --> 21:01.760]  people in many other areas and just is an overall boost to people's skills that is somewhat
[21:01.760 --> 21:03.520]  rare to find these days.
[21:03.520 --> 21:06.800]  So even for a JavaScript developer, if they know how the kernel works and if they know,
[21:06.800 --> 21:11.480]  you know, behind the scenes, you know, what's going on at a system level, I think that is,
[21:11.480 --> 21:22.240]  you know, can't be anything but a good thing.
[21:22.240 --> 21:23.240]  Thanks for the talk.
[21:23.240 --> 21:29.040]  I have a question about the audience, so is it mainly targeting students at the college
[21:29.040 --> 21:34.080]  level and then partnering with university, or do you also consider a conversion program
[21:34.080 --> 21:38.960]  like existing software developers that want to switch to more low-level kernel development?
[21:38.960 --> 21:44.920]  Yeah, so the course at UMass Lowell is, you know, a course at UMass Lowell for people
[21:44.920 --> 21:48.720]  enrolled at UMass Lowell, but the program on LFX, the whole idea of putting it on there
[21:48.720 --> 21:54.720]  is that we could accept people from anywhere in the world doing anything, students, non-students,
[21:54.720 --> 21:57.080]  you know, U.S., non-U.S., whatever.
[21:57.080 --> 22:02.320]  So any background, as long as they have some knowledge of C, good internet access, decent
[22:02.320 --> 22:06.920]  English and a computer, then we can take them.
[22:06.920 --> 22:12.160]  The only limitation is our ability to grade the assignments and handle that many students
[22:12.160 --> 22:15.480]  and like the resources that we have to work on the program, because this is, you know,
[22:15.480 --> 22:19.800]  for the three engineers and, you know, potentially a couple more people who are interested and,
[22:19.800 --> 22:23.240]  you know, a couple of interns, right, it's, this is all a side project for us.
[22:23.240 --> 22:25.680]  Like, I'm mainly a RELL engineer.
[22:25.680 --> 22:26.680]  That's my main job.
[22:26.680 --> 22:30.720]  Like, this is just a side project, so we don't have a ton of resources and, yeah, we're basically
[22:30.720 --> 22:35.160]  limited by how much time and resources we have to run the program.
[22:35.160 --> 22:39.080]  But we're, yeah, we encourage people to apply from all sorts of different backgrounds.
[22:39.080 --> 22:49.680]  Okay, thank you.
[22:49.680 --> 22:55.960]  So back in the day when I used to participate in the development of kernel, it required
[22:55.960 --> 22:59.880]  us to basically build the entire mainline.
[22:59.880 --> 23:03.800]  Your slide showed something with Raspberry Pis.
[23:03.800 --> 23:08.960]  Is it feasible to build the entire kernel using a Raspberry Pi?
[23:08.960 --> 23:13.120]  So, that originally comes from, we used to do a lot more stuff with Raspberry Pis before
[23:13.120 --> 23:14.120]  the chip shortage.
[23:14.120 --> 23:17.880]  It became more difficult to get them and then, you know, I think generally like a lot of
[23:17.880 --> 23:24.720]  the introductory stuff we were doing didn't require them as much, so we moved somewhat
[23:24.720 --> 23:25.720]  away from them.
[23:25.720 --> 23:30.040]  But, yes, we did work with Raspberry Pis and we were compiling the kernel on the Raspberry
[23:30.040 --> 23:31.040]  Pi.
[23:31.040 --> 23:36.920]  So I was able to do it on, on the three, it took a lot of weird tricks and it took like
[23:36.920 --> 23:41.720]  a day, you know, because I think it was like single threaded, and on the four, I could
[23:41.720 --> 23:42.720]  do it.
[23:42.720 --> 23:46.720]  It was a little bit faster, but if you didn't have cooling, the system would overheat unless
[23:46.720 --> 23:49.640]  you used fewer than four cores.
[23:49.640 --> 23:54.200]  But you can do it, and in fact, you can install Fedora, you can install CentOS, and you can
[23:54.200 --> 23:55.200]  install Rail.
[23:55.200 --> 23:56.200]  Oh, yeah.
[23:56.200 --> 23:57.200]  Yeah, I know that.
[23:57.200 --> 24:04.200]  Yeah, all right, that was my question, thank you.
[24:04.200 --> 24:27.200]  Hey, he had said, with the, the alternate, the non-university rate for doing it, you've
[24:27.200 --> 24:32.680]  got people here working at the same time, possibly, I was just wondering, since it's
[24:32.680 --> 24:36.600]  also, you've got the same assignment schedule, what is the time commitment like for someone
[24:36.600 --> 24:40.160]  trying to get into it, but might have full-time work?
[24:40.160 --> 24:42.320]  What is the time commitment like?
[24:42.320 --> 24:47.360]  Because it depends how quickly you complete the assignments.
[24:47.360 --> 24:57.480]  Okay, so it's, well, the, the lectures are the, the sessions that we have, which are,
[24:57.480 --> 25:02.520]  you know, on, on Google Meet, they're also recorded and we post them there about their
[25:02.520 --> 25:08.040]  two of them per week, generally, and they're, they're each an hour and 15 minutes, and it
[25:08.040 --> 25:11.520]  goes for, you know, like a standard 12-week semester.
[25:11.520 --> 25:17.760]  In terms of the assignments, I mean, if you're, you know, I think at the beginning, setting
[25:17.760 --> 25:21.680]  up the environment, send email patches, if you haven't done that before, and compiling
[25:21.680 --> 25:25.920]  the kernel, and, you know, we have one assignment where you write a shell, you know, that can
[25:25.920 --> 25:31.520]  be, I mean, yeah, it depends how familiar you are with, with some of the concepts.
[25:31.520 --> 25:38.480]  But, you know, I think of it as, it's just a, like, same commitment as a, you know, maybe
[25:38.480 --> 25:44.880]  medium to, like, junior to senior level undergraduate computer science course, approximately.
[25:44.880 --> 25:49.200]  I mean, that's, that's how we're designing it.
[25:49.200 --> 25:50.200]  Sure.
[25:50.200 --> 25:53.520]  Hiya, thanks for the talk.
[25:53.520 --> 25:58.880]  Any of these courses, like, I'm, I'm curious about back, back-porting and forward-porting
[25:58.880 --> 25:59.880]  stuff.
[25:59.880 --> 26:04.000]  So, when, when, when you're doing this course work, are you working on a specific version
[26:04.000 --> 26:05.000]  of the kernel?
[26:05.000 --> 26:10.640]  Like, yeah, this course is going to be on 6.2, because I find that a lot of problems that
[26:10.640 --> 26:15.920]  I run into is, okay, this driver was just introduced, let's say, in October, but I need
[26:15.920 --> 26:22.160]  it on 5.13, so I need to back-port it, and that might be trivial, right, might just be
[26:22.160 --> 26:27.440]  copying some files, but if the API has changed in some way, it can be really difficult to
[26:27.440 --> 26:32.760]  figure out how it changed, because the documentation, to my knowledge, is not great there.
[26:32.760 --> 26:39.280]  So, is that some, something that you actually sort of discuss in these courses?
[26:39.280 --> 26:42.560]  We haven't done back-porting specifically.
[26:42.560 --> 26:45.200]  We don't, I don't think we even, we don't have any assignments that's specifically
[26:45.200 --> 26:50.000]  asked for, for back-ports, because we have to, we design them in a way that, you know,
[26:50.000 --> 26:53.520]  each student is doing, like, relatively unique work, so it's kind of, you know, and they're
[26:53.520 --> 26:58.920]  all posting it in a, in a public mailing list, so, you know, we need to figure out how we
[26:58.920 --> 27:04.240]  can kind of either generate assignments or, or figure out assignments that are basically
[27:04.240 --> 27:09.560]  by design, difficult to impossible to, to just, to cheat in.
[27:09.560 --> 27:11.960]  And so, so with back-porting, I mean, we could figure something out.
[27:11.960 --> 27:13.120]  We haven't really looked into it.
[27:13.120 --> 27:17.480]  We talk about, you know, get cherry-pick and rebase, and we talk about the, the background
[27:17.480 --> 27:22.640]  that you would need to do that, and I think we may have done a demonstration of it in
[27:22.640 --> 27:23.640]  the course.
[27:23.640 --> 27:26.640]  But that's a good idea.
[27:26.640 --> 27:41.440]  Do you anticipate that the recent admission of Rust as a programming language, at least
[27:41.440 --> 27:46.280]  for driver development in the Linux kernel, will cause renewed interest, especially among
[27:46.280 --> 27:47.560]  younger talent?
[27:47.560 --> 27:50.680]  And are you already anticipating for that in the program?
[27:50.680 --> 27:54.200]  Potentially, yeah, we've, we've discussed it, and we've heard some people talking about
[27:54.200 --> 27:55.200]  it.
[27:55.200 --> 27:58.360]  We had a few people who were in the course, who were excited about Rust, and Rust in
[27:58.360 --> 27:59.360]  the kernel.
[27:59.360 --> 28:01.920]  So we, we have seen some interest in the kernel.
[28:01.920 --> 28:07.320]  Probably, yeah, it's probably increased interest from young people, I would say.
[28:07.320 --> 28:09.120]  But personally, I need to, I need to learn Rust.
[28:09.120 --> 28:10.120]  I don't know Rust.
[28:10.120 --> 28:31.640]  Do you mind if I do a little bit of, oh, sure, yeah, they're here, but people take us
[28:31.640 --> 29:00.760]  with them.
[29:00.760 --> 29:08.000]  where you were asked, you know, to work on vice-drivers and then, you know, non-handling
[29:08.000 --> 29:14.000]  veterans and so on, actually, when we fly, so they were teaching us since, I think it's
[29:14.000 --> 29:20.000]  never been more, you know, for the people, maybe like, what are you, and you're a serious
[29:20.000 --> 29:21.000]  person.
[29:21.000 --> 29:22.000]  I mean, I'm probably fine.
[29:22.000 --> 29:23.000]  Sorry, let me get the mic away from you.
[29:23.000 --> 29:24.000]  Oh, okay.
[29:24.000 --> 29:25.000]  Yeah.
[29:25.000 --> 29:26.000]  Yeah.
[29:26.000 --> 29:33.000]  All right.