Episode 826 transcript

FLOSS-826

Jonathan: Hey folks. This week Neil Gompa joins me and we talk about Fedora the new 42 beta, the new KDE flagship news, fedora on arm and Risk five Asaki Linux and more. You don't wanna miss it, so stay tuned. This is Floss Weekly, episode 826, recorded Tuesday, March 25th, fedora 40 twos new flag. It is Time for Floss Weekly.

That's the show about Free Libre and open source software. I'm your host, Jonathan Bennett. And today we have, well, it's kind of a, a, a hark back to I think the very first pair of episodes that we did here on Under the New Hackaday Flag. We interviewed, I think it was the first, the first two we interviewed Neil Gompa.

And I touched base with him this week and I said, Hey, lots of stuff that hap has happened at Fedora and at KDE and at Sahi, all these different things that he and other, other places too. All these different places that he sort of has his fingers into that he works with. And he said, sure, we'll come, we'll talk about it.

It's, it's been a, a long time, almost two years now. So let's do a recap, a revisit. And so I've got Mr. Neil Gmpa here with us and welcome to the show. Hi. It's been, it's great.

Neal: Like, hey, at least it wasn't another 10 years. Yes. Yes.

Jonathan: I think that was our joke when we, when we did the, a couple years ago when we did the episode.

We'll have to wait 10 years to have you back. Well, we, we did not follow through on that when we, we get you back sooner. Ah, I'm fine with that. Yes. That's okay. Yes, me too. So stuff's, stuff's been happening. I'm, I'm trying to think about where to start and maybe, maybe the most interesting immediate news is Fedora 42, that beta is now out.

And that has sort of kept you busy, hasn't it?

Neal: Oh my gosh, yeah. It's been it's been a big release like. Lots of things have been happening in there from a technical as well as non-technical perspective. And I've just been running myself ragged dealing with all the different things along with everybody else who's working in, in fedora kd.

Jonathan: Now, remind us what, how, how do you, how do you fit into this? Like, what is your role at, at Fedora, at Red Hat? Question mark. KDD? No, no, no. I I do

Neal: not. I do not. Red Hat things. I do not. I am not, I'm not a Red Hatter. You know, there was the brief stint where I was in Red Hat as a salesperson. We talked about this in that first Hack a day episode.

And that was it. I, and I wasn't even in engineering, I was in sales, right? Like, it, it, it doesn't, I I am not a Red Hat hat.

Jonathan: It doesn't

Neal: count. It doesn't count. I don't do Red Hat things. No one, no one ever construe anything I say for something that Red Hat does or wants, or cares or anything of the sort.

Yep. Yep, that's it. But with within a fedora context, right? Mm-hmm. So. I am a member of the Fedora Engineering Steering Committee, which is elected every in my election term, basically every spring because I started with Fedora 32 and have been reelected every year, every year since. Mm-hmm. And Fedora 32, by the way, was spring of 2020, so that's, it's five years now.

Jonathan: That doesn't seem like five years ago, but, okay. No, it

Neal: doesn't, but time hurts a little. It's starting

Jonathan: to, yes.

Neal: I am also the lead of Fedra, KDE. I was previously co-chair rather than the sole chair, and before that I was a member of it. And I've been doing Feda KD stuff for going on 10, 10 years now.

Mm-hmm. I am also involved in Fedora workstation, fedora Cloud, fedora Server, Python, go Rust. Lots of different things in fedora space from the development and engineering perspective. Right. And the, but the main, I think the main thing we're here to talk about is kind of like my role within Fedor, Katie, and doing all this stuff around, around that.

But I also, like most recently, I helped mentor new people into Fedora to create their own special interest groups.

Jonathan: Mm-hmm.

Neal: Matthew Kosar, who makes the Miracle Window Manager, a tiling window manager based on Mirror. I helped him create the Miracle Sig and set up the Miracle Window Manager spin in Fedora 41.

Nice. And then for Fedora 42, I mentored and assisted Ryan Brew to build out Fedora Cosmic, which is also part of this release. Yes, so. There is that, and I also, you know, helped people with like getting Fedora Katie mobile and all this other stuff that happens. Like, so like I think you could, it, it's probably, oh, and also Fedora LX Q type kind of revived and now, you know, that's for Fedora 42.

That's Wayland using MiWay based on mirror as well. Mm-hmm. So I think like I have my fingers in a lot of pies in Fedora is I think, the way we want to put this. Yes,

Jonathan: yes.

Neal: So with, and we didn't even talk about as Sahi here because where I created the Fedora Sahi Sig and we'll get there made Fedora, sorry,

Jonathan: remix and all that other stuff.

We'll get there. That's, that's an interesting story in and of itself. So there's, there's something big with Fedora 42. Like we said, we just had the beta release. There's a big change with KDE and the workstation, like flagship. Right. What's going on there? So, historically speaking, since

Neal: Fedora 21

Jonathan: mm-hmm.

Neal: There we, we've had a rejiggering of how things have worked. So Fedora 21 was the first release where we had the change to have this concept of working groups with additions. And these additions are essentially top level curated variants of fedora that are promoted front and center. They are considered the flagships mm-hmm.

Of, of the, of the community. And we started with Fedora Workstation being the OME product or OME Edition. Mm-hmm. And we had Fedora Server and Fedora Cloud. Later on they added Atomic with CoreOS. Mm-hmm. Replacing Atomic later. And then there was Fedora iot added on top of that. And now we are adding fedora Katy Plasma Desktop as an addition.

Jonathan: Mm-hmm.

Neal: And so the additions are. Essentially variants that are in addition to being flagships, they're actually, they're backed by what is known as a working group, which is a special team that has to report both to the engineering steering committee, fesco mm-hmm. As well as essentially being required to be part of the other organs of fedora like marketing and council for the top level leadership and things like that.

Mm-hmm. So these teams have extra responsibilities that they have to fulfill as part of being a top level you know, a top level team. So then below additions are Spence, which is actually an ancient. Ancient variant term that we've used since the beginning of introducing spins in Fedora seven.

Jonathan: Mm-hmm.

Neal: These are essentially deliverables like media curated collections that are done by teams within Fedora, usually special interest groups that go in and take a collection of things that they care about.

Jonathan: Mm-hmm.

Neal: And produce a, a you know, something to, to show off. So, so you can download,

Jonathan: you can download an ISO that has your preferred desktop environment in it.

Neal: Right. So Fedora, KDE was the first live CD in Fedora. Mm-hmm. It was where the live CD stuff started. And you know, there's, there's many others now. Like, as I briefly mentioned earlier, there's Miracle Window Manager, there's cosmic, but there's also lx cute, there's a sway and much, much more. Right.

So I think there's like, I don't know. 15 of them or something like that. Goodness. I, I, yeah, there's a lot now. I, I have not counted. I guess if I go to the website and go look at the spins. Let's see. So three, three columns and 1, 2, 3, 4. So three by four. So 12. 12.

Jonathan: Yeah.

Neal: Yeah. I'm subtracting out automatically the fact that for some reason the website still shows the KD desktop as a spin as well as in addition.

But like, that's fine. Takes time. It'll get the it, everything eventually consistent. Yes. Is the way that this works. Yes. Eventually. So like there's, there's 12 spins and then then labs are essentially, you take a spin or an addition and you provide a workload opinion. So like for example, if you are an artist, you probably want the design suite.

Or if you are an audio enthusiast or an or a podcast producer, you probably want jam.

Jonathan: Yeah.

Neal: These take an existing desktop edition or spin. For example jam takes KDE Design Suite takes ome, and then layers changes the application set and some of the default settings to optimize it for a particular workload.

Mm-hmm. So for example, for Jam, they change some of the kernel settings to make it better for real time audio.

Jonathan: Right.

Neal: And things like that. And of course the applications are changed and things like that. So for the labs, there's three by three by three. So it's, it's 9, 10, 11 of them. So there's not that many, but mm-hmm.

The idea is that there is a way for a fedora community contributor if a group of people come together who want to do a particular type of solution, I think is the official word, to create a workload that, to, to express a workload type. There's a path for them.

Jonathan: Can you, can you mix and match the desktop that you want with the lab?

Like is there a way to say, I want an ISO that is jam, but running on cosmic?

Neal: Well, so not right now. Technically you could install cosmic and then install the, what we call composition group. Mm-hmm. Or comms groups for short, which actually describe the workload. Mm-hmm. And then you could just install it on top of an existing cosmic one.

Mm-hmm. So if you wanna do audio production on Cosmic, by All means you can, yeah. You could install the software that is used on those, on that spin, on that lab, and then put it on a cosmic environment. Like, there's nothing particularly nuts about being able to do that. But these are just like opinionated combinations that people have decided that they like and want to put together.

And sometimes it makes natural sense, for example. A lot of multimedia production applications are cube based. Mm-hmm. And they have integration with KD or they're from kd. Yeah. And so it makes a lot of sense to pair that with a KD desktop. It perplexes me that design suite ison based as opposed to KD based, given like half the software for kd.

But hey, I'm not in charge of design suite. If they want to switch to kd, I'm happy to have that conversation. Mm-hmm. But like that is the, the idea is that the people that are working on this basically come up with how they want to express a, the workflow and for the workload.

Jonathan: Yeah.

Neal: But you can always mix and match yourself if you want.

Jonathan: Sure. With, with KDE in 42, what, like, what's the exact terminology? Is it Fedora workstation? KDE plasma? Is that the, is that the title? No, no, no, no, no, no, no. So

Neal: the workstation working group owns the workstation name within Fedora. Okay. And they wanted to represent the GNOME desktop. Okay. For a variety of reasons.

This is how, that is Uhhuh and we've accepted it and that's fine. So for Fedora 42, the edition is the Fedora Katy Plasma Desktop Edition. It is a mouthful. I, I, I know. Yeah. But we get to have the word desktop in it, which is always important. Yeah. We have the word edition in it, which is even more important.

Yeah. And we have the word KDE in it, which is awesome. Yeah, KD is the best. So like, we want to be able to, and, and this has actually been beneficial to us because one of the things that I, we've taken a slightly different approach compared to some of the other additions, is that we have done integrated cross marketing between fedora and kd.

So if you go to the KD edition Katie plasma edition website on Fedora. You'll notice that the learn more links across the page actually take you to brochure sites on the KDE website. Nice. To explain more about this stuff. I'm taking the approach of let's not do it all by myself and let's, let's, let's kind of sharing this, and this was actually, this website was actually helped put together by a member of the KD Fedora community Carl Schwan, who's one of the KD artists that like puts together the websites and stuff for kd.

Jonathan: Cool.

Neal: Actually helped make this, like, basically wrote this website.

Jonathan: Mm-hmm.

Neal: Like, I gave him some headlines and some pointers and did some review of it, but he basically wrote it for us and I very much appreciate that he put in the time to make this work. Yeah. And it looks fantastic, I think.

Jonathan: Yeah. And so, you know, I, I assume this kind of means that you, whenever Fedora has an official presence somewhere, like we were talking before the show about Foz dim or, you know, wherever it is, this means that the, the.

The KDE Plasma Desktop Edition is sort of at the headline just as much as Workspace

Neal: Workstation. Yes. See works. Yes, for sure. That is the, the goal, the plan, the hope. It may take a few reminders to make sure that we get all this straightened out. Like, I mean, of this is brand new. This is, this is the first time we have something like an addition that people can easily demonstrate

Jonathan: mm-hmm.

Neal: Coming that, that we need to start prioritizing in terms of marketing. But yes, like the idea is that in the breadth where we talk about fedora and workstation, we should also be talking about KD plasma. The goal is to make them equal.

Jonathan: Mm-hmm.

Neal: Because that is what they are at this point. Yeah.

Jonathan: Yeah. I, I imagine that you to, to sort of talk about them with a shorthand, do you just refer to them as workstation and desktop? I.

Not officially.

Neal: I call it fedora Katie and Fedora workstation. Okay. That works Because, because while yours is, is also accurate. I, I try not to upset people. I try to keep, I try to keep the peace. I know some people don't believe it, but I do in fact try to keep the peace. Oh,

Jonathan: yes, I understand. Okay.

Something else I've wondered about with this is there is another distro out there that uses fedora as sort of, its, testing and source. So Red Hat, red Hat Enterprise Linux. Yeah. Is is in some ways a downstream of Fedora. Is the change that KDE is now a first class citizen, is that going to get reflected, do you think?

Not speaking as a, you're not a Red Hatter but just sort of, I guess, reading the tea leaves and what conversations you are aware of, is there gonna be any change to Red Hat as a result of that? Or is Red Hat always gonna be a known first distribution?

Neal: Red Hat doesn't have the staff right now to add a second desktop back into the portfolio.

Jonathan: Hmm.

Neal: The desktop team has, you know, this is the fir, I don't know if you've noticed, but like, for the first time in many, many years, red Hat is hiring in the desktop team. Hmm. They're hiring for known people to work on G stuff. I, I don't know. Whether this will eventually result in some kind of change on the Red Hat level to reintroduce KD plasma at in Red Hat Enterprise Linux.

Honestly, that's more of a conversation with Red Hat customers. Sure. If Red Hat customers start really pushing back and wanting to have KD plasma as their preferred desktop in on Red Hat Enterprise Linux, then I'm sure that there will be a change. But as of right now I don't, I don't see that happening.

That being said the up the immediate upstream for Red Hat Enterprise Linux is sent off stream,

Jonathan: right.

Neal: And pretty much all the cool people in CentOS Stream use Katy Plasma. So if you look at, for example, CentOS Hyperscale and the alternative images like KA plasma is a big part of, of those as well.

You know, like Cintas Hyperscale has not yet put not yet put out. It's Cintas Stream 10 based deliverable, but as the one making those deliverables, I could tell you that Katie plasma is very, very important to me. Right. And Troy Dawson, who works on Fedor, Katie e and supports Fedora k supports KD plasma on enterprise Linux distributions in Fedora kde.

Mm-hmm. He makes the alternative images and KD plasma is, you know, a very important part of it for him.

Jonathan: Sure.

Neal: And so I think if you look at the community of people who actually use Enterprise Linux in a way where they're part of. Developing this stuff or putting together things, I think you'll see a higher degree of KD plasma usage than you, than you would expect.

Jonathan: Mm-hmm.

Neal: There are a lot of people, even within Red Hat, who are KD fans, even if they're not in charge of the product.

Jonathan: Right.

Neal: KD plasma is very important to a lot of people, and it is a high quality experience that deserves to be shown out more and, and commercially supported more than it is right now.

So, so I'm just doing my part to make it better.

Jonathan: So in, in the realm of RL there's nothing official, but one, one might say that there are some undercurrents that are at play.

Neal: Maybe I, I, I can't say one way or the other 'cause I don't know. Yeah. But if Red Hat customers start asking for KD plasma and keep seeing or, or keep deploying it and, and, and ignoring Danone.

Maybe that will change. I don't know.

Jonathan: Yeah.

Neal: As it is, you know, like it was unthinkable five years ago that Red Hat would even provide any kind of engineering support for extra packages for Enterprise Linux. And here we are now, they have two people on staff that help work on supporting Apple and community people to work with re people mm-hmm.

On making Apple workout. So like, you know, that's a Red Hat is culturally very slow. We, we've talked about this before Yes. That they're, they're like this mishmash of the eighties and nineties cultures.

Jonathan: Right. And

Neal: they don't, they haven't really taken in a whole lot of the newer newer community cultures.

Yeah. So it, it takes time.

Jonathan: Sure. Sure. That's fair. Let's get, let's get back to Fedora and looking at the Fedora 42 beta. Is there, is there anything really exciting in there that you're just, so we, we talked about KDE becoming a, a. You know, flagship level. But other than that, is there anything in 42 that you're just, you're really excited about?

Neal: Oh, I mean, there's a bunch of good stuff in, in Florida 42. Like, I mean, honestly like the, one of the cool things is that we now have Fedora, Katie e specifically out of the box, we have support for running X 86 applications on Arm.

Jonathan: Mm-hmm.

Neal: Because we integrated the f stuff from the Fedora Sahi people.

You know, I say that as if they're separate people. It's all the same people, but like, you know, the stuff that we did for Fedora Sahi remix in 41. Mm-hmm. We are bringing forward into 42 in Fedora kde. Right. So if you're using Fedora KDE on Arm, you know, we have that stuff integrated so that it, it becomes a much easier experience.

Another thing that I'm really excited about, I mean, from a user facing point of view, is. The, the stuff that we've done to introduce the cosmic spin, right? Mm-hmm. The cosmic spin is very exciting to me because the system 76 folks have done a lot of great work mm-hmm. To build out a brand new ecosystem.

Yes. Nobody, people are not ballsy enough anymore to do stuff like that. And I'm very, very humbled and honored to have been part of help putting that together. Mm-hmm. And the co the folks that at System 76 working on Cosmic, I've met them in person multiple times now and we've talked a lot about Fedora Cosmic and they've been extremely helpful

Jonathan: Yeah.

Neal: Every step of the way. And they're very, very pleased with the progress we've made as well. Mm-hmm. Supporting Cosmic and Fedora. And they really do wanna see how it grows on, on its own. And they're very, they're very much more in kind the kind of spirit I have about open source where I feel like. Where you're building all this stuff and doing all these things to create these you know, these things that are different, the different options, opinions.

Mm-hmm. Like, you know, people decry, fragmentation, whatever. But like open source is all about you being able to do you? Yes. And, and if you can't do you then, like, what's the point? Right. Right. Like the and, and cosmic is a desktop that is made by people that embrace that philosophy. Mm-hmm. They are very, very happy to see Linux distributions adopt cosmic.

They're very happy to see Linux distributions adapt. Cosmic. That's a very big deal. So like, you know, this is in contrast with ome, which would generally prefer that you don't touch the ome experience. You keep it very close to what they're shipping. Yeah. Upstream and, and deviations are not welcome.

Jonathan: Right. Nome has always been very opinionated.

Neal: Right. And cosmic is opinion aid too, but it, but in many respects, they're more flexible. Mm-hmm. And their idea is we want you to show us how you can make cosmic awesome for you. Yeah. And I think that that is a wonderful philosophy. It's basically the reason why I use KD plasmas is because this philosophy's built into it.

Like you look at what we have for the marketing material for Fedora Katie Plasma Desktop, and we call it the Next Generation Personal Desktop, and I mean it. Mm-hmm. The idea is this is a desktop that you get to own. It is, if there's something you don't like about it, you can change it and make it more fit your needs.

Jonathan: I, I'm reminded of a conversation I had just the other day about h the way HDR works, but we'll, we'll set that to the side for the moment.

Neal: I mean, okay. Look man, sometimes technology. Sometimes sand doesn't work the way we want it to. Just give me the slider. But you know, they're working on it, right? Oh, yeah.

Like it's, yeah, yeah, yeah. And I, this is all, so here's, this is all new stuff.

Jonathan: Yes. Yeah. And we could, we could talk about H-D-R-I-I, I'll give everybody just a little bit of a, of a background as to what I complaining about privately and what I was told. Right? So like, I've got the big, I've got the OLED TV behind me and my normal use case for that.

I, it's, it's a 4K. And I discovered very early on after getting that, that it splits up into 4, 10 80 p windows very nicely. Right? And so I, I end up using it like, as if it were a two by two monitor grid. And there is a, there's a little plugin for KDE that lets you use keyboard shortcuts. Moodier is what it's called, lets you use keyboard shortcuts to like full screen things and then put things full screen into these tiles.

And so I use that a lot. And one of the things that I like to do, because it's HDR and there's now support for HDR and KDE is I'll have like three places I'm actually working on stuff. And then I'll go grab an HDR video. Usually it's like a nighttime walk through some big city and I'll throw it on the, in the fourth quadrant.

And. I do not like the white windows being blindingly bright. And in KDE 6.2, there's a nice little slider where you can tone down the brightness of your SDR content and yet have your HDR content at full brightness. And for my particular use case, that is amazing. And they took it away in 6.3. And so I've gone and I've complained and I was told, no, no, no, we're not gonna add that back.

But at the same time the, the, the, I think it's Xavier, it's aver, right? It's Xavier Hug Z

Neal: Huckle. Yeah, Z

Jonathan: Okay. Yeah. At the same time though, like he, he listened, he understood what I was talking about. And, and I feel that there is, you know, if enough of us go in and say, man, we really like this feature, like there's a possibility of getting it back.

I've also. Contemplated a little bit like, well, what would it take for me to patch my own KDE install, you know, my own plasma install so that I can get the, the cool stuff in 6.3 and yet still have my slider back. Yeah. I mean, that's the kind, that's the kind of thing I love about Fedora though, and KDE, like, I can go in there and find the patch.

It, honestly, it wouldn't be that difficult to go and undo whatever they did in 6.3 and make it work the way I want it to, and then recompile my own plasma packages like that is not sure the hardest thing in the world to do.

Neal: Also, you could make an applet or an extension that does it, because I'm pretty sure the underlying, the underlying setting mm-hmm.

Still exists in Quinn. Because you can manipulate all these settings using the tool case screen. Doctor,

Jonathan: I went and looked in case screen doctor and couldn't quite find it. But that's, that's an interesting point that it might, well, case screen doctor's terrible about documenting itself. So like there is that too,

Neal: like I, I wind up struggling figuring out how to use it.

'cause it doesn't have a man page and the help is dynamic and it's like, and it will crash if you try to use it outside of the desktop running. So like it is, it's a

Jonathan: very special

Neal: application.

Jonathan: Yes, yes. It is not, it is not intended necessarily for regular people to work with. I get that.

Neal: Yeah. But like essentially it, it just talks to, it talks to Quinn News against private, the private interfaces.

Yeah. But it is possible, you know, if the underlying infrastructure still exists. For the controls.

Jonathan: Mm-hmm.

Neal: There's no reason you couldn't make a pla, a plasma applet or extension that does it.

Jonathan: Yeah. That's an interesting idea. I may have to look into that in all of my copious spare time with all of the, all the other things on the to-do list.

Neal: All of us with our copious spare time can get so much done, right? Yes,

Jonathan: yes. I've, I've discovered though that I'm capable of a nearly infinite amount of work, so long as it's not the work that I'm supposed to be doing at a given time. Oh, geez. I know. It's the same thing. It's so, so very hard. Yes. Yes. Okay, so let's, you, you guys did something kind of crazy.

You, you brought a sahi in to KDE et su, these two fedora fedora, I guess the kde too, probably. What, what's the, what's the story with Fedora and Ashi? Fedora is kind of the flagship for Asaki now, isn't it? Yep. What, what happened there?

Neal: Well, I mean, so originally Hector Martin was working on this with ar, with Arch yeah, I had actually approached like near the beginning, just after the announcement on Christmas 2020

Jonathan: mm-hmm.

Neal: About working with him to produce a fedora ASA distribution. Sure. There wasn't enough stuff ready to be able to do it. So like early days, in very early days, yes. But he made his original prototypes on arm on Arch arch Arm, which is an unofficial derivative that rebuilds the arch packages for arm architectures.

Mm-hmm. Arch as a project does not support arm. So this was just one guy rebuilding everything for Arm because he wanted to very wildcat sort of approach. Sure. And that's fine and all. Sure. The problem was that a Sahi alarm, as it was called. Was not very stable and kind of broken a lot. Mm-hmm. I think the thing that, that popped his cork at the end was the fact that Firefox web r TC was turned off because it was too hard to figure out how to make it work on arm at build time. Yeah. And so that's like, well, okay, now I can't video conference, I can't do any of my work job things.

Jonathan: Yeah.

Neal: This is not good. And I was still plugging away at Fedora Sahi, and I'd offered him that, Hey, you want to, I'm, I'm happy to like, kind of essentially take over the distro part of this for you so you can focus on all the rest of it.

And I gave him an early prototype where I had. Things like kind of stroll together.

Jonathan: Mm-hmm.

Neal: And he was like, this is already better than what I have with, with arch alarm. So the Sahi alarm. So then we made the plans to switch to it. I started formalizing it. My partner in crime on this, David Alka helped with getting all the infrastructure in place and putting the SIG together.

And essentially we poured it over the, a little bit of the things that came from a Saki alarm to Fedora. And then the rest of it was All right, we, we just packaged the stuff and then put it together and we created a new opinionated base integrating Fedora features into it.

Jonathan: Mm-hmm.

Neal: And I think we've had a really good relationship since then because one of the things that has kind of happened since is that desktop Linux packages that are in arm on Fedora are actually really getting used.

And like there were fun surprises in the beginning where it turned out. No one had ever actually used these on real computers and they didn't work and they would crash in bad ways because Apple, Silicon CPUs, particularly the M two and newer, had features that fedora packages were built to, to, to use such as pointer authentication and branch protection and all this other stuff.

Jonathan: Yeah. Yeah.

Neal: That was enforced by the CPU. Well, it turned out no one had ever actually verified that those would work. And so these packages would just straight up crash when they were used. And it took, and these kinds of things, these kinds of things are the inhibitors, the real inhibitors, to being able to use another architecture as a daily driver for a desktop.

Like if you don't have people actively using it and you don't have a platform that's performing enough for people to use, you're just not going to see these things get fixed.

Jonathan: Yeah.

Neal: All the things are theoretical until someone uses it. Yeah. So, you know, in, in many ways this has allowed us to leapfrog, you know, fedora Katie has seriously benefited from it being the flagship variant for Fedora Sahi remix.

Jonathan: Yeah.

Neal: Because now ARM support in K is pretty much the best of all the open source desktops.

Jonathan: Mm-hmm.

Neal: And we have support for, you know, some of the weird things that you do with arm graphics and things like that, that pretty much none of the other desktops have.

Jonathan: Mm-hmm.

Neal: And because none of the other desktops have it, the experience is not as good.

And it until, and some of those desktops just don't want to care about it. And that's fine. That's their prerogative. Sure. But Katie is, was perfectly happy to have us contribute and fix things and get things working. And so that's, and like things like supporting HDR, making it a priority to handle high color depth

Jonathan: Yeah.

Neal: And stuff like that, those are important features that you need because the Apple platform. Pretty much mandates all of it.

Jonathan: Yeah. If somebody wanted to have an arm, laptop or desktop for that matter to run K to e fedora on what, what's the go-to right now? I wish

Neal: there was a go-to right now. Right now this is this, it's, it's a rough world outside of Apple silicon because to put it bluntly, not many people care.

Like you look at, you have Qualcomm with the X one extreme. Yeah. Right. The Windows Onar devices that they released last year. Right. Those devices have the beginnings of Linux support, but they're not in a way that is usable. Yeah. And this is in part because I don't know of qual, I am not assured Qualcomm really cares about making this work.

Jonathan: Right. So for some, for some devices, you get the what do they call it, like server ready, where it's where Arm has UAFI and boy that helps. Yes. That helps a lot to get boot working without having to go in and and mess with Device Tree. Oh my goodness. I've lost so many hours of my life fighting with Device Tree.

I still lose hours fighting with Device Tree. Yeah. U-U-E-F-I makes it a little bit better.

Neal: So it's UEFI with a CPI because what Qualcomm has done for their, for lytic support for their devices is they've taken the worst road in the middle, which is UFI with Device Train.

Jonathan: Okay.

Neal: Which you can do now.

Jonathan: Yes.

Neal: Because Yes, because ARM has made it a optional feature to have A-C-P-I-I

Jonathan: see

Neal: uhhuh. Yep.

Jonathan: Yep. That's, and

Neal: so system Ready now lets you validate a platform without a CPI support and all the ARM embedded people are like, oh, but A CPI is terrible anyway. We should device tree's. Awesome. It's like as a person that has to ship something for generic hardware, as a software integrator, as an operating system vendor, I.

Device Tree is the worst possible choice that you could have made for everyone because Device Tree is not standardized or specified. It's actually controlled by the Linux kernel project.

Jonathan: Yeah. And

Neal: changes based on what Linux kernel developers want. And that makes a ton of sense in the embedded place where no one cares about anything.

Jonathan: Yeah.

Neal: But like once you get past that and you start talking about scale out long lived devices that need to be updated continually, this breaks down fast. Yes.

Jonathan: Yes.

Neal: And and like

Jonathan: it is, it is not designed for consumers. It is, it is built for developers or kernel developers in particular, not for consumers.

Yes. That is exactly what I found with device three.

Neal: Yeah. That is, it's a big problem. And like, so right now, if any, any, all I should say, all current Windows on arm devices have incomplete UEFI implementations and thus cannot boot, S-U-E-F-I on Linux. Oh fine. Yes. It is something that I have complained about and is something that no one else seems to care about.

So it is the current state of affairs. As far as a potential go-to device, Lenovo has actually been fairly good about their proof of concept, as they like to call it. Mm-hmm. Work for supporting Linux on their windows on arm devices.

Jonathan: Hmm.

Neal: So there is a couple of them that you could buy where probably by the end of this year you will be able to reasonably load Fedra k arm on yeah.

The Lenovo ThinkPad T 14 s Snapdragon version. Mm-hmm. Gen Six is one candidate.

Jonathan: Yeah.

Neal: And I think there's like a idea pad slim that's being worked on as well. I think the Slim seven X or something like that has a Snapdragon version. Mm-hmm. And that one's also being worked on those two. And of course the ThinkPad X 13 s Snapdragon version.

All three of those devices have varying degrees of you can do stuff.

Jonathan: Yeah.

Neal: With it. Because the firmware that is required for those things has been made available. So you have full hardware enablement and combined with someone wrote device trees. It, it turns on and, and actually kind of works.

Jonathan: Yeah.

Neal: Performance, your mileage may vary, but between the difficulty of getting them set up the middling performance of the, of the socks, in reality I, I, there just isn't, I mean, the best of the go-to is still not great. Mm-hmm. But hey, it's an option. And we now have, we, we have UEFI enabled. Live ISOs for arm for fedora kd, as well as the raw disc images that you typically flash and you work with in the first for, for things like raspberry pies and whatever.

So like they're available you can try. If someone is actually interested in, in driving this more. I am, you know, within Fedra KD we are very interested in making this work. Mm-hmm. But getting access to hardware is the difficult part. Sure. Because they're not cheap and they are not fun to work with.

Jonathan: Don't hate me for this observation, but we're kind of in this weird place where we need windows on arm to do well enough for enough people to buy these, for the price to come down, for us to really be able to support Linux on arm on them. You're

Neal: right. But I'll also point out that even Windows is now making concessions for Qualcomm.

Craziness.

Jonathan: Right?

Neal: Like, so the way that it works for Windows to boot is they have just enough a CPI for the NT boot environment to work.

Jonathan: Mm-hmm.

Neal: And then they use drivers to pun fill in the holes, essentially Device tree style on Windows to fill out the rest of the platform enablement. That means that these devices are effectively going to be orphaned as soon as Qualcomm stops caring about them.

Right. Which is probably within two to three years. Yeah. That's the real, real dangerous part about the way that things work right now with ARM, is that unless someone cares and has the ability to make things work, like this is go that that stuff's gonna turn into E-Waste way faster than any X 86 machine will.

Jonathan: Yeah.

Neal: Or even Power PC machine. You know, and open and, and Device tree came from freaking open firmware like, and, and Open Firmware device. Trees are not this bad because they're expected to be on the board. They're expected to. They have, they have like restrictions on what, how they can be filled out. It's only on Arm and Risk five where device tree is a mess because it's controlled by the kernel rather than by a specification.

Jonathan: Yeah. Although, to be, to be fair, this is in some cases a pro. So this sort of flavor of thing is a problem on X 86 sometimes too. Sure. I've, I've got a desktop right over there that I'm working on for a, a local customer and it needs, it needs a Windows 11 reinstall, right? Oh boy. It's got, it's got some kind of proprietary NVME thing in there that need like, so Windows will not see the drive.

I know it's there. Linux can see the drive, no problem. Mm-hmm. Windows will not see the drive. So it needs, you know, it needs the driver installed during install time. Yep. That machine will not boot into recovery. It blew screens because it does not have the driver.

Neal: It's like it's, it's broken. So back there, there's an HP desktop that I have

Jonathan: mm-hmm.

Neal: Which has the opposite problem. The disc controller cannot be seen by Linux because it's sufficiently weird and not initialized properly by the bios. But Windows can see it because, because for some reason there, the windows, the windows, early boot stages seize the hardware differently than, than Linux does.

Jonathan: Yeah, yeah.

Neal: So like, yes, this is not exclusive to those platforms, but I would argue the problem is far Oh yes. It, it's far worse on arm and, and, and risk five than it is on X 86 on power. Like, and you know, speaking of power, right? Like Fed Door 42, Katie has now power PC support and it's Power nine and, and newer, basically it's little Indian power.

And like, I'm sure most people don't care. But like, hey, if you're one of those people that have a Talu workstation, hey,

Jonathan: you can now run KDE on it. Yeah, that's true. I guess there is Telus I was gonna ask whether there is any, any existing power devices out there, but there is Talu I've found about that.

Neal: I mean, in theory you could also install it on IBMP series servers, but like, I don't know why you want to put a, you know, door K on a P series server.

Right, right, right. But hey, you do you, it's, yeah, like the, the whole reason SUSE added like remote desktop support to gno for GNO Wayland was because people run no on mainframes. Yeah. So like Yeah. People do some

Jonathan: strange stuff. Yep. Speaking of the KDE. Arm fedora experience. What about like the raspberry pie?

I know it's not, well, you can make it into a laptop. In fact, I've got the the Ella Crow crow view, I think they call it, which is like a laptop form factor with ports to be able to slap a pie into the side of it. That's neat. Yeah, it, it is. I've actually, I've actually talked to them and I said that when you make the version two of this, just give us the give us the slot in the bottom to be able to slot the, the compute module into it.

None of this popping into the slide. Just let me plug a compute module into it in the bottom, and then well, if the,

Neal: if the, if the thing has like portage for like, you know, the N VMEs and, and all the other things like in the, in the enclosure. Yeah. Then the compute module makes sense. Otherwise, you might actually still want to use the standard F form factor board.

Jonathan: Yeah.

Neal: I, I would think that the, I, again, I know nothing about this product. I, I you, you're, I'm literally hearing about it just now from you. Let's see. Here

Jonathan: do. It's, it's neat. It's the crow view note, that's what they call it. And so, oh, it looks cool. It is, it's neat. And it's, it's literally just got USBC for power.

It's got an hd, like an HDI mini connection and A-U-S-B-A, and then they have this little carrier board. I, I literally just happened to have it here, this little carrier board that snaps in beside the pie and then the whole thing. Oh, and ports over the interfaces. Yes. And so the whole thing just kind of goes together like this, which it's, it's super cool that it works, but you have a laptop, but it's super ja.

You have a laptop with a sidecar, right? Yeah. It's a super Jake setup, but it's so cool. Exactly. Yeah,

Neal: I, I, I think I agree with you that you'd probably want to use the compute module there, but I would also probably say that if they go the compute module route, they need to make sure that the main board actually exposes all the interfaces that the CM five has.

Like for example, you probably want to be able to add MVME storage, definitely want NVME to make that work. Yes. Yes, absolutely. You also probably want to plumb through the interface to plug in the wifi card because Right. If you're using the compute module, you no longer have built in wifi.

Jonathan: I think some of them do.

Neal: Pretty sure some of them do. Wow. I am so out of touch with the raspberry pie people. Like this is, let's see, let's go look at their products. I'm just gonna quickly, the compute module is, oh yeah. Wow. I'm impressed and horrified. But point is right, the, you know, when you do it with the compute module, you gotta do all the interfaces that are normally on the board itself.

Right, right, right. But yeah, like the Raspberry PI four is supported. The Raspberry PI five is still being up. The work is still being upstreamed. My understanding is the Raspberry Pi Foundation has paid Sosa to do the work. Okay. So eventually it will show up and then we can do something with it.

But right now we can't. But you can do with a PI four, you could use the Raspberry PI 400 computer

Jonathan: mm-hmm.

Neal: And, and put it on there.

Jonathan: Yeah.

Neal: That's actually how I test on the rasp for, for Fedor Katie on arm is 'cause the Raspberry PI 400 is easy to get. It has all the things. I don't have to MIT fuss with it and it's fine.

Yep. Yeah. So that is probably the closest to a go-to arm device I could give is the Raspberry PI 400 right now. Which is, which is depressing,

Jonathan: but I mean, it's a, it's a really cool, awesome device. But yes, that's, that's less than ideal. Is it any better or any worse on risk? Risk? Five.

Neal: So risk five, I think there's been pluses or minuses with risk five.

Mm-hmm. One of the pluses for risk five is that I think people have made it a point to not make the same mistake that ARM did and say that. You know, arm arm's mistake was they said that they're not the stewards of the ecosystem. It's not their job.

Jonathan: Right. To make

Neal: sure people that are integrating arm IP do the good job of doing it.

I think they've seen the folly of their ways, but it's too late for arm to fix it.

Jonathan: Yeah.

Neal: For risk five, I think people are trying to be better about this. There are far more products that use standard UEFI for risk five now than there were for, than there are for ARM today after 10 years of it. So that is, that is a blow to the heart.

If I ever had any, if I ever said any, if I ever saw any, yeah. From the perspective of, you know, device trees and stuff like that. I think for a lot of the single board computers, they're still doing it this way because. They're still doing the open SBI and they're doing the, the crazy jump route. But like for example, the Milk five, the Pioneer Machine is full on U-E-F-I-A-C-P-I.

So it works like a normal computer.

Jonathan: Yeah.

Neal: I think the framework 13 risk five board while slow as balls is actually a standard UEFI setup. Mm-hmm. So like, I think we are seeing this mistake not being repeated as badly. Yes. Now, what is a problem right now is that risk five chips themselves, their feature sets are fragmented.

Jonathan: Yes.

Neal: Because the instruction set architecture is a puzzle, and, and people have not, I think my understanding is that the, they have just agreed on how to baseline for Linux, like just like a couple months ago.

Jonathan: Yeah.

Neal: So it's so new that there's no hardware with that pro, with that configuration.

Jonathan: So yeah, they, they need, they needed Intel to say, you know, this is the, the pen two feature set and then mm-hmm.

And then everybody else, but nobody's there. Right. There is no intel in the, in the, well, I mean, Intel works with Risk five, but they're not, they're not doing this. They're not,

Neal: they're not pushing anything.

Jonathan: Right. They, they, Intel uses risk five internally. They don't necessarily tell anybody about it. They, they put it, they put it like, I mean everybody does.

Yeah. Yeah. Well, I mean, that's kind of the point of it, right? Like all your Intel chips have little risk five cores in them. That's what runs the, the intel management stuff.

Neal: Right. And so does the the Nvidia GPUs, they're all risk five cores. That makes sense. Makes sense. Yeah. So, and same for Western digital hard drives.

The microcontrollers are all risk, five cores, stuff like that. It. Yep. And I, I don't care about those things, right. Because they're not, but, but they are endemic to the problem because, because there is no one who says, I care about this from a gen pop perspective, and I want PC experiences to work well, and I need Linux to actually not be crap on these platforms.

Jonathan: Yeah. How do

Neal: we do this? Yeah.

Jonathan: What's, what's the workflow look like for a regular person to do this install and use it as a daily driver?

Neal: Flash it to an SD card?

Jonathan: Yeah.

Neal: Right now, that's pretty much it. I mean, to be fair, right? Risk five in most distributions is brand spanking new. Nobody has any workflows.

Nobody's built any tooling. No one's done any serious testing. Fedora is one of the early builders of risk five. Fedora risk five has been around for a while, just quietly in the background. And then of course there's a buntu risk five and there's open source risk five, and there these are all efforts that exist.

Mm-hmm. All Mo Linux is working on a risk five build, and they'll be the first, you know, my, my, my bet is they'll be the first enterprise Linux based platform to have a risk five offering. Mm-hmm. And that will, that will change things because essentially hardware makers have to figure out how to work with the software vendors.

Because what happened with ARM was most of the ARM products for a long time were vertically integrated, and that's not what's happening with Risk five because the ecosystem is not vertically integrated, you are seeing much more collaboration on these fronts.

Jonathan: Mm-hmm. And so it, when, when we say vertically integrated, like let's just talk, let's dive into for just a second about what we're talking about.

Like, so this, this is Google makes Android and makes a pixel handset. And is paying, you know, Qualcomm to make the specific CPU that goes in it. Right. And so, like, they have control over the entire thing. And so they don't necessarily care about what gets upstream to the Linux kernel and what the experience is because they, they control the whole thing so they can just make it work for their, their platform.

So, well, they

Neal: don't have to ca they don't have to care about a generic case. Yeah. Is essentially what it

Jonathan: boils

Neal: down to. Yes. Yeah. That's although to be, to be clear, Google's actually one of the better Android handset manufacturers. That's, that's fair. Actually do upstream their platforms. And you can use them with standard Linux after you beat them to death a little bit.

But mm-hmm. You know, like a better example probably would be like Motorola, Sony Sure. Like the other Android manufacturers just don't do these things.

Jonathan: Yeah, yeah. Yeah. Interesting. Alright, man. We have we have been going for almost an hour here. Is there, is there anything that you wanted to talk about that we didn't get to?

Neal: Well, I don't know. I mean, like, this was mostly like, this was kind of spur of the moment, so I didn't exactly have a plan of things to talk about. But like, I always enjoy coming on Oh, yeah. And talking, even though we've only done it twice and now thrice, so

Jonathan: yes.

Neal: Let's not, let's not make it take too long the next time.

Jonathan: Indeed. Indeed. I did want to ask you about, I'm trying to go through my mental list too. There's been sort of a shakeup recently with Ishi Linux.

Neal: Oh yeah, sure.

Jonathan: And, and even some drama with Rust in the Colonel. And I'm, I'm curious, one, what your take is on all that, and two, is that gonna have an impact on the Fedora Sahi remix?

Neal: I'll answer the questions in reverse order because it's faster. It has no impact on Fedora Sahi remix. Mm-hmm. Because Fedora Sahi remix just chugs along, just doing the thing. I work on it. I take the patch sets, I build the kernel and produce the images and push them all out. That all continues as it does now.

So for the upstream components, Hector Martin essentially

Jonathan: burned out, right? Like that's, that's he got tired of fighting the fight and like that's understandable. That happens. Yep. When it comes

Neal: to rust in the Linux kernel it has always been contentious because mm-hmm. As that community doesn't churn a lot.

Right. So, like, one of the things I like to point out to people is it is very rare for someone to leave the Linux kernel project

Jonathan: mm-hmm. External project.

Neal: And that is actually a bad thing because there is no churn, which means there is no fresh blood, which means there is no new leadership, which means there is no new ideas coming in very often.

Jonathan: Hmm.

Neal: So this is actually the first like really new idea that has come into the Linux kernel in a long time. Right,

Jonathan: right.

Neal: And I mean, we're also talking about a kernel project that doesn't really use ci. They don't have they don't have good workflows for continuous testing and integration and delivery.

Their code review style is painful to use because the modern internet does not really handle having tons of patches being sent as emails. Mm-hmm. You have, you know, these people that have been doing the same thing for a very long time and things have, and, and they're all paid by their companies to just basically do the thing.

Jonathan: Yeah.

Neal: And they are in a position of which if they disagree with their company they can kind of do whatever they want. Sure. Because. They have their pickings of places to go work.

Jonathan: Right.

Neal: And so things are relatively stale and stable up in terms of personalities, in terms of opinions and things like that.

Jonathan: Mm-hmm. So

Neal: that's the stage here.

Jonathan: Mm-hmm.

Neal: It's also like overwhelmingly paid people. Like there are very, very few hobbyists that work in the Linx Turtle project these days. I believe that Greg Crow Hartman once said that, like even back in 2003, it was like 89% you know, these, and, and the churn rate from then to now is almost zero.

Right. So, you know, if, if in a 20 year time span you have all the same people doing all the same things in all the same ways and nothing has changed.

Jonathan: Yeah. And so, and then

Neal: you try to introduce change.

Jonathan: It's not surprising when you try to introduce a change, it's gonna be sort of cataclysmic.

Neal: Yeah.

Jonathan: And cataclysmic and cataclysmic is overselling it, but just to make the point.

Neal: Right. It, it is going to just be difficult. Yeah. Right. It's gonna be a difficult process. And I, the, the problem is that when you have a really old project who has really old opinions, who really doesn't get a lot of new people

Jonathan: mm-hmm.

Neal: And then you get a bunch of new people who are working in their free time to do something interesting

Jonathan: mm-hmm.

Neal: That is kind of hearkening back to the early days of how Linux was being built out. But they have different opinions of how to do things than you do. Yep. You're, you're just a culture clashes everywhere. Yeah, exactly. And, and to be clear, not everybody who works on a Sahi Linux has. Is a complete outsider in the, in, in the, in those things.

Like for example, Alyssa has been working in graphics drivers for decades at this point. Yeah. And she has very good experience working with both the kernel and the user space people on graphics.

Jonathan: Mm-hmm.

Neal: You know, Jean Grau has worked in multimedia for a long time and is a fff m peg contributor and fff m peg is a very similar cultural mindset.

Yeah. Warts and all. And you know, it kind of goes on like, it's not like these, and you have me who exists in a lot of places. Right. Like, so you, you, it's basically a story of burnout and the fact that like. You know, there's only so many fronts you can fight. And he tried to take them all on.

Jonathan: Yeah.

Neal: And it's just, you know, you can read the story.

He has written a blog post about it. Mm-hmm. And, and that's pretty much like the, I mean, I'm just happy it, it went well, the transition because mm-hmm. Like, it could have, it could have been different and yes, thankfully everything went fairly well as it has. Yeah. So

Jonathan: there you go. HEC Hector stepped down.

There are other people still at the Ishi project and Yeah. Someone else has already stepped up and said, okay, I'm gonna help maintain what's in the kernel. Yep. So it, it's gonna, it's gonna continue on maybe a little bit slower with the, with the loss of one or two of the people at the head, but,

Neal: well, we don't know yet because the current focus is now around Upstreaming stuff rather than writing, you know, new, I mean, some new features are being made.

Like we just released microphone support for like, just this past week, I think.

Jonathan: Yeah.

Neal: So. But, you know, we'll, we'll see how it goes. Right? Like the, it, it's still early days in the transition. We've only made the change like a month ago. Mm-hmm. So we'll see. Yeah.

Jonathan: Yep. Alright. Very cool. Let's see. I think about the only thing we have left is to is to ask you what your favorite text editor and scripting language is.

Neal: Well, it's still Python. Still Python, but I now use Kate as my text editor for things now. Ah, I fell in love with being like, so I told you before that for simple stuff, I tend to use Nano. And then for big programmatic stuff I use emax. Mm-hmm. But what I didn't tell you before was that in my, my original heritage of doing software development, I always used IDs.

Because big, big projects, I can't keep all the context in my head. I need a way to manage it. And k is actually a nice middle ground here. It has a language, language, LSP support. Mm-hmm. It has you know, many of the features that you would want in id, it's very much like vs code except not written an electron and actually fast.

Ouch. And like, if I wanted a full blown IDE, I could move up to k develop, right. Like so. Mm-hmm. Kate has actually been really nice and I've been using it for the past year or so for, for bigger projects and I've been very happy with using it. So Kate is now my preferred editor of Choice graphically and I still use Nano for small things.

Mm-hmm. Because, eh, I don't need the whole id, I don't need an Id like experience when I'm just editing a file. Sure,

Jonathan: sure. Alright. If somebody wants to try out Fedora KDE specifically, say the 42 beta, where's the, where's the best place to go? Look for it. KD dot fedora project.org Super easy. And if somebody wants to get involved with sort of the backend tinkering, is that the same place or where, where should

Neal: all that, all that, all that information is present on the, on the [email protected].

Jonathan: Yep. Yeah. Very cool. And then there's, there's various places like the Matrix chat room and, and all of that where people can come and hang out, ask questions Yeah. And bug you about features that are not there yet and all of that good stuff. Oh my

gosh, yes.

Yes, there are. I try to, I try to behave myself, but I've been lurking in Yeah, you're fine.

Chat for a very long time. You're fine. I've been on both sides of that equation. Right. I have been the guy saying, man, it doesn't work. Just fix it. And I've also been the guy saying, yes, we know it doesn't work. We're trying to fix it, or this is the reason why we're not fixing that. I've been on both sides.

Right. Yes. Empathy. Yeah. It helps. It helps. All right, Neil, thank you so much for being here, man. It has been, once again, a, a joy and a pleasure to have you and we appreciate the time. Yeah, sure. And I'm happy. I'm happy to be on

Neal: the, on the show and I'm happy to come back again in the future if you so desire.

Jonathan: Yep. We will definitely have to do it. I appreciate it. All right, so that was Neil Gmpa talking about KDE in Fedora. And definitely I can recommend it. It's what's running on the machine behind me. It's not what's running on this one for reasons, mainly because this is this laptop ship with Papa West.

And so inertia kind of got me there, but it's what I run on the main desktop behind me. But we've got some fun stuff coming up in the future on Floss Weekly. So next week we're talking the plan is to talk to one of the developers behind YT Dash dlp. That is one of the u the new YouTube downloader.

And then the week after that we're talking with Stefan about LXC, the Linux Containers project. And then a couple weeks after that we're working with someone from the Kuna Project, which is a Linux sort of, I don't think it's exactly an antivirus, but it's a Linux security project and platform. And definitely looking forward to that as well.

We appreciate everybody that is here, that watches and listens and whether you get us live or on the download, we will be back next week for another episode of Floss Weekly.