LINUX Unplugged - 343: What Linux is Best At

Episode Date: March 4, 2020

We try the Mac desktop for 30 days, find out what we think it does best, and where Linux will always have it beat. This episode kicks off the start of a bigger conversation series. Plus community news, very handy picks, and more. Special Guests: Alan Pope, Maria Komarova, and Michael Aaron Murphy .

Before we hit record, we were talking about Stadia, and what a disappointment it has turned into. And I already regret buying my, I think they called it Founders Bundle or something like that. Yeah, they got you suckered in. I mean, I had fun playing it here at the studio, but I haven't heard you mention it since then. No, I pretty much stopped using it a couple of weeks after I got it. So I logged back in this morning, and I now am the proud owner of one pay game.

I bought that early on, Red Dead Redemption, for like a hundred dollars. But you got to play something. And then I got Metro Exodus now for free. I don't know. Where did they go wrong, guys? Is it the price of the games? Is it a lack of games? Is it all of the above? I think it's the fact that you don't get a refund whenever they close the system down. When they inevitably shut it down people are just too familiar with google google got the track record so yeah i think it's a bit late as well i mean i appreciate that the whole streaming services it now is the time for streaming services like we've got netflix we've got uh spotify you know we understand the whole streaming technology but everyone's already got

their game system set up and they've got their steam account and they've got their itch.io for indie games and all those other stores that nobody cares about i think they're just it feels too late it feels like people already bought into all the other platforms. Yeah, there's that. And, but to what Cheese was saying, I just read an article where the journalist went around and interviewed Stadia players and he did not prompt them, but almost all of them said they were concerned

Google would shut it down eventually. Wow. That's become Google's brand. And then, of course, you combine the fact that you need a pretty decent connection. When I want to play video games, I'm at home and I don't have a great connection at home. I have a good connection at work, but I'm not playing video games at work. So you combine all these things, the game prices, the fear of them shutting it down, internet connections.

It just doesn't seem like Google's going to pull it off. We'll see what, we'll see if the next, we'll see what the next thing has. I mean, who knows? Maybe they'll keep it going. People are putting a lot of money into this. They kind of have to stick with it. Hello, friends. Welcome into your weekly Linux talk show that makes its own Super Tuesday. My name is Chris. My name is Wes.

Hello, Wes. Nice to be podcasting with you for 343. Yes, I'm casting my vote for Linux Unplugged. We have a lot to cover this week. My name is Wes. when I tried using a MacBook at home for about a month. What? And I came away with a couple of interesting notes that I want to share with the audience, observations, and where I think the Mac gets it right, where they get it wrong, and why ultimately Linux is the long game.

I still think that, but in this endeavor, you know, I had to give the Mac a real fair shake. I challenged myself to try to treat this as intellectually honest as I could. At first, I've always, whenever I've used the Mac, I've always kind of had this visceral kind of don't like using it. It's just like all these rough edges response. And it's a different paradigm. Yeah. So this time I said to myself, all right, all right, but what if, what if you gave it a day like you give plasma?

You give plasma a whole day. What if you gave this a whole day? Would it be so bad then if you gave it a day like you give plasma? You give plasma a whole day. What if you gave this a whole day? Would it be so bad then if you took a day? So I'll tell you a little bit about that and what I realized in that process. But we have a bunch of community news to get to as well as some pics. But before we go any further, I got to say in the morning to Cheese. Hello, Cheese. Hey, guys.

How's it going? Hey, buddy. And then, of course, time-appropriate greetings, Mumble Room. Hello. Hello. Good evening. Howdy. Hello.

Hi. Hi. Well, I hear a Popey in there. Hello, Pope evening. Howdy. Hello. Hi. Hi. Well, I hear a Popey in there. Hello, Popey. Well, hello there. How are you? Well, hello. I'm good.

I'm quite well. It's nice to see everybody in there. A great turnout again. I see Ace Nomad, Brent and Byte and Carl and Colonel and Mini Mac and Popey and... Is it Stoon? How do you say that? Is it Stun? Stew-in?

Stew-in, yeah. That's about it. Oh, welcome, Stewin. I like that you're kind of iffy on that. In Dutch, it's stoon. Ah, okay. All right. Okay. I see what you're rolling with. Well, let's talk about some things that are coming down the distribution pipeline very soon. If you're going to grab Ubuntu 20.04 when it comes out or one of the modern distros that ships GNOME Shell, then you're likely going to end up

with GNOME 3.36, right? I think that's what comes in the focal. Popey's been getting me kind of tempted to actually start running it as my daily driver. You're not already? I downloaded it last night. So I might. I just, I don't know. I've been really happy with Manjaro, but I think

I want to wipe it and go with 2004 just to use it for a bit before the release to have some hands-on. But there's some nice things landing in 3.36 of Gnome Shell. A bunch of fixes, a bunch of performance improvements, and these changes are landing just kind of in time because it's going into code freeze. It's going to be all done.

So whenever it does land, in whatever distribution you use, there'll be some things to look out for there are improvements to the css engine but there's also a batch of cleanups and refactoring to how it handles files in general for the s css and gnome shell also now is avoiding certain io operations in the main thread which saves on disk io and this is cute one, but it's a nice fix. The local time zone is now cached, which will avoid reading the contents of ETSI local time more than necessary.

But as a consequence of that, Gnome Shell actually behaves a lot better under heavy I.O. loads. Animations also are now disabled under various circumstances, such as if Gnome Shell is using a software render. Or if you're sharing your screen with VNC, it turns off the animations. That's so smart.

It's nice, huh? And also, it officially is now shipping that extensions management app that we talked about once before that allows you to globally disable extensions. It sort of takes that functionality out of GNOME Tweak and puts it in an actual GNOME Shell official application where users can look at their extension status, toggle them on or off, and things like that. It might be small, but it feels pretty big there. Yeah, I will use the crap out of that. Now, of course, Mudder hasn't been left out.

Lots of good improvements there, including improved screencasting support. And this part caught my eye. The screencast engine is now implemented on top of Pipewire and exports a Pipewire node that can be read and consumed by other applications. So now you've got that nice jack-like ability in the

Pipewire graph to go connect things up. Until now, also, this API would download the contents of the monitor frame buffers from the GPU to the system memory and then pass it to Pipewire after that. But that was obviously highly inefficient, and that's been improved. That's cool. So that means doing screencasts and stuff like that,

but less of a load on your GPU and on your CPU. Exactly. That's a great improvement. And using Pipewire behind the scenes to do that. And also, back in February, a major change came in on how Mutter handles drawing monitors. This is the first step towards achieving a goal of one frame clock per monitor. And that's a nice architectural change that should eventually translate to real usability wins. Yeah, a lot of rendering improvements in general,

cutting down on just unnecessary rendering in different scenarios, and even improvements in how the wallpaper is handled, which allows Mutter to save resources when scaling is applied to a monitor. And then, of course, various fixes for Wayland support have landed. Small memory leaks were plugged. And Mutter now avoids flickering on X11 when Windows are switching between undirected or not.

GNOME 3.36, looking really, really good. Plasma, looking really, really good. Plasma looking really, really good. Like things are getting in shape. I can't help when we talk about, I can only think back a little bit to, you know, when we express some of our concerns about the underlying architecture and just the amount of work it would take

to sort of get on really solid footing. But I think we're past that. I mean, the pace of development, the changes, the improvements deep down in how Mudder works, those are all coming in. This to me seems like a clear long-term knock-on effect of Canonical putting Ubuntu behind Gnome Shell. And they've done a big portion of a lot of this work, but even just that, it's inspired others to pick up the work there. It really has gotten to a much, much, much better place. It's just about time to throw it all out and start over, isn't it, Popey?

Oh, I don't know about that. 2004 is going to be around for a long time. So, you know, you can install 2004 in April and be confident that you're going to keep getting updates for five, maybe ten years. Yeah, good answer. Dang it. Dang it.

Also, just bringing the slow, eventual, no big deal, low-key arrival of Wayland's Firefox 75, which is due to be released next month, should have its native Wayland support in really good order now. Yeah, merged yesterday, as we record this, were the Firefox Wayland patches for VA API video acceleration support in conjunction with FFmpeg. Support does require enabling OpenGL or WebRender within Firefox, as well as setting the Wayland DMA buff VA API control knobs. And if you don't know what that is, I didn't really either, but we'll have links in the show notes. A control knob, huh?

I imagine that's just in the Firefox about config settings, but I could be wrong. I like it. I like the terminology there for some reason. I think it's kind of funny. Cool to see it. And like I say, it's sort of just this low-key, slow arrival of Wayland that would be no big deal. More fundamentals. And I think, too, maybe what we'll talk a little bit more with your experiences on the Mac, it is these sorts of video acceleration niceties that you just sort of expect on the modern desktop. Yeah. Fair enough. Look at you teasing. Well, before we get there, how about a little housekeeping? Don't forget about the daylight savings monster. It arrives

this Sunday here in the States. Oh no, again? I know. Imagine the time math we do with working with folks in both London, but also in Australia. And then, of course, all the time zones in the United States. So, can we please just knock this crap off? Right. UTC, let's be done with it. I'm good. I can do it. Let's go.

I'm fine with that. Being on UTC, given that's what we're on right now. I know. Even with that, making Popey and Joe happy, I'll tolerate. Yeah, I mean, that's a compromise we all have to make. Absolutely. It's worth it. I want to give a plug while we're talking about Joe.

Foss Talk Live, June 20th at the Harrison, which, as you know, is near King's Cross in London. Of course. I believe tickets are $1,000 each and your first board. What's that? No, that's not true. Oh, okay. All right. Well, you can get the details at Fosstalk.com.

We'll have a link in the show notes. Tickets are free. Donations are welcome, though. So maybe put like $1,000 in that donation box. That's probably. One of those events I would just love to make it to. I know. Someday. Oh, my God. It's so good. It that donation box. That's probably one of those events I would just love to make it to. I know. Someday.

Oh, my God. It's so good. It's so good. It's so cozy. You're in the basement of a pub, a proper pub. And it's really dark and dingy down there. And you've got uncomfortable small chairs. And that's what makes it awesome.

Because you're all a bunch of nerds. Super cozy. Joe pointed out the other day that most of the people who've bought a ticket, or I say buy, have obtained a ticket, have bought one ticket. So most people go on their own. And people do worry about whether they're going to fit in and whether they're going to have people to talk to. Everyone will fit in. Everyone will have someone to talk to.

It's a fantastic event. I highly recommend people go if they can. And there's social lubrication to go all around. I would love to go. I was going to totally go, but then Wes wouldn't let me go because of the whole COVID-19 thing. I just worry about you in the show. Well, all right. So that's plenty. We've probably gone a little too long there. So more links, there's additional things in the show notes. But I want to talk about Linux being a general purpose OS for very specific things. Think about that statement for a second.

A general purpose OS that can be used for very specific things. And while we're talking again about Joe, go check out episode 84 where they covered this in a slightly different track. This has been a wider conversation that we've been having privately amongst ourselves and with friends and other people. And they really kind of encapsulated the conversation really good in episode 84 of Late Night Linux. And really the idea there was, is it time to focus on open source applications and not so much on the desktop? I don't buy that argument. I think the desktop has a very bright future. And in fact, I think Linux is better positioned for that long term for certain use cases than any other platform. And I think we'll get there. But with that statement also,

there's a flip side to it. And that is, there are legitimately use cases that it is not well suited for. We've got to be honest about that. Based on just market dynamics and realities, and the way economics work, the way it all just has shaken out, it's just the truth. It's just... And the expectations of regular users. Yeah. When I look at Linux, I look at it as a platform, really, that is used on servers, but also is almost a runtime these days with WSL on Windows and

really kind of a universal kind of just everybody's writing applications to run it on Linux and these containers. And if they do that or run on any platform, you can run Docker containers on Windows 10 through WSL 2 now. This is something that and they've got Docker desktop that they're working on. Including, I watched a YouTube video this morning of a YouTuber who is using XFCE from inside WSL2. That's fantastic. It's pretty crazy. Of course, Docker is available for the Mac. You can run Docker on a free NAS server, which is BSD.

So there's the runtime Linux, there's the server Linux, there's the embedded OS Linux, there's desktop Linux. Now, open source has clearly been successful. There's a lot of really massive applications out there like VLC and Firefox and Blender and lots of desktop applications that are very successful. But there's not many distributions of Linux or desktop environments that are as successful as some of those applications are. And I think, as time goes on, this will shift. But today in 2020, there's still a very valid case for a specialist OS. Remember I said Linux is a generalist OS that can be used for specific purposes,

like embedded devices is always a go-to. But then you have specialist OS that are made by commercial vendors that really serve one master purpose. And this is the train of thinking I started to get down when I started using this MacBook at home. I had this 2017 Touch Bar MacBook and it's not getting used for anything. And ultimately, I just enjoy technology, whatever it might be. That's a nice piece of hardware. And so I thought,

all right, well, let's try I was kind of inspired by something Alex and I have been talking about. Let's try when I go home using a totally different computing paradigm to get a broader sense of at work and at home. Because then I could have all of my work stuff on the ThinkPad, and when I'm at work, I'm

on the ThinkPad, and when I'm at home, I'm on the MacBook. It's a totally just like jarring difference. Very clear indication of your switch. Immediately hated it. As every time I sit down to use macOS, I just can't stand it. It's a pretty foreign environment at this point. It's like a hundred small little things. Like the trackpad's great, but natural scrolling's turned on by default.

Like there's just, everything has a little bit of a rough edge for me that I have to go finesse. And I always would get sort of frustrated and be like, I'm not doing this. This is a $2,000 computer. I shouldn't have to bother with this, right? But yet I have no problem setting up a brand new laptop or a Dell XPS or a custom-built PC with Plasma or Gnome Shell and spending a day or two tinkering to get it just the way I want. And I can cut that down, especially on Gnome Shell. But like with Plasma, it usually takes me a couple of days because I kind of fix it as I go along. Right. You realize, oh, yeah, I do want to tweak this, that, and then. So I said, okay, well, if I'm going to be honest here,

I have to afford the same allowance for the Mac. I can't judge it because I don't judge Plasma without modifying those things. I don't judge Gnome Shell. So I need to give the same allowances for the Mac. And you've got to reach that everyday use steady state that is really what it would be like to use it. I will make just a side note the keyboard is atrocious it's really bad it's this is one of the butterfly yeah it's really really bad

and that's one of the reasons i quit using it plus we just didn't really have we didn't need a mac in production anymore so it's like when the when the company sale came i was like yeah we don't use this i'll just keep this we don't use it for anything in the production stuff. Like, I had to make a hard decision there. And so it just sat being unused with the battery dying. And that, like, just kills me. Just sit there looking at it, you know, and I feel guilty. Money slowly wasting away. So every now and then I'd, like, plug it in with a USB-C

cord and then, like, unplug it and then, like, a month later plug it back. Like, I just, I was neglecting it. And so I thought, okay, well, I'll set up a $50 budget. Because on the Mac, you got to pay for everything. Everything costs money. You want window snapping, it costs money. You want to fancy something up, it costs money. You want system meters in your menu bar, it costs money. You want to manage what icons are in your menu bar, it costs money.

So I set up myself a $50 budget and I have not spent all $50, and within a day or two, I think I had it as usable as I would probably say Gnome Shell. That's really interesting, too, because I think, at least when I'm used to the Linux desktop, that seems like a really big philosophical barrier. I couldn't expect those features

without having to pay for it. But when you frame it with a very reasonable, I mean, $50, especially if you're going to invest in tools maybe one time or yearly, that's so reasonable. It feels dumb to pay for something to help me tile a window or snap a window. But if I just say, all right, I'm not going to spend more than 50 bucks, they are good apps. They are probably the best apps at doing that. The one that is really freaking impressive that I got is called iStat Menus. This is one of the

more expensive ones and some of these now want subscriptions which is cute. Which is really cute. But it's menus, it's menu meters up in the menu bar and it is I mean it's really good. Like it's worth $30 or it was $19. It was worth it. Really, I would

buy it again. I would buy it for Linux in a heartbeat. And then you have some insurance that's money going to the developer to continue to develop it. And of course, on the Mac, you have the one platform. You're never going to switch away from the Mac OS environment and not be able to use this anymore. It's just going to keep working. That is true.

Yeah, because if I bought something that only worked with GNOME Shell, it wouldn't necessarily... Maybe they have a Plasma, but maybe not. Yeah, I hadn't thought of that. I generally don't need to do it on those desktops. Like, Plasma has everything built in I need. I just have to turn it on and learn how to configure it. So once I got past this hurdle, though,

I could appreciate a few things that I think the Mac gets right. In part, I started to notice things that Linux can't really do, that they can do on the Mac. And that is, they build macOS with the intention of their pro apps being used on the platform, like Logic and Final Cut. What that means is, the systems work together in unique moments of stress, probably in a way Linux never can. Like the kernel team is working with the audio subsystem team, who's working with the video subsystem team to make sure that things remain

responsive when the system's under serious pressure. And really what got me started on this whole journey was I wanted to see what would happen if I hooked up my eGPU to the MacBook. Like what does macOS, Catalina do? I mean, you've been pretty into eGPUs. Does that make sense? And my first experiences with this were documented in the Chris Last cast where it was called, like, the Max Vista moment because Catalina is a rough turd, let me tell you.

It is a rough rider, and it is UAC all up in your face constantly. It's worse than Vista. And so you got to get past that. That takes like a solid two weeks to get through all those damn dialog boxes. And once you do that, you hook up the eGPU,

it was remarkable. So I would work on a project in Final Cut and the entire system would become aware of the presence of an eGPU, and applications could elect to use that eGPU, or I could go in and tell them to prefer it even. And what would happen is I could render something, and Final Cut would dispatch work to the eGPU, it would dispatch work to the AMD GPU that's built into the laptop, and it would dispatch work to the Intel CPU. It would manage all of that in the background,

effectively utilizing all of those cores. And on this 2017 MacBook, I am crushing out video that you wouldn't think should be possible on this thing. And then the thermal throttling kicks in, and then it slows down a little bit because it's a Mac after all. But fundamentally, they have something wrong. There's a sickness in this platform that is not competitive with Linux. But it doesn't mean it is not competitive.

What is hard for us to understand and appreciate is Apple's style here has been really easy for third-party manufacturers, hardware makers, people that are making interfaces, people that are, you know, like high-end audio interfaces like Apollo that are several thousands of dollars and only work with Macs, or audio or video editing surfaces that work with the Mac, or software that comes out on macOS, there is a limited scope of problems that they have to solve. There's a known set of hardware, and there's a known set of OSs.

And Apple only supports a few OS releases at a time, so it's a very narrow spectrum you have to support. And the flip side of that support means you can optimize better for what you are supporting. And if you're writing something for a pro user, you can almost guarantee they're on the latest OS. In fact, if you buy a Mac Pro today, you can only run Catalina. You can't run an older version of macOS. So you can guarantee that

your customer is on a certain set of hardware and software. And that that software gets revved every year on a yearly cadence, which is a built-in reason to then prompt your users to buy the next version which adds support for the new OS so you have a guaranteed revenue bump every year. So there's a

real healthy third-party ecosystem around it. It's a ecosystem around lifecycle you can just sort of latch right onto. Hardware and software. Well, I'm not so sure you can be that confident that people are on Catalina or whatever the latest Mac OS is, because the secondhand market for Macs is so strong that people will buy older Macs. There's a whole series of YouTube channels of people reviving old Macs and keeping them running with whatever the latest version of Mac OS is that they all run.

And for some of them, it's older than Catalina. And people have these Macs kicking around, like I've got a MacBook Air kicking around here and that's not running Catalina. So I don't think you can be super confident. Like if, sure, if you're following the gravy train of people

who are always buying and revving to the next device every year, if they're doing that annual update, but I think your average Joe buys a Mac and keeps it until they run it into the ground and then buy another one. I agree. Your average Mac user, absolutely. But where Apple gets you is certain hardware requires Catalina and certain software like Final Cut to have eGPU support also requires Catalina. So if you're a certain class or Xcode, for example, if you're an Xcode developer and you want live Swift previews,

when you're writing an iOS app, you have to be on Catalina. That's true. And I also, while I'm here, I want to mention, I think it's interesting you set aside a budget of like 50 bucks to get yourself what you might call a working desktop. Right. And I think it's interesting

because that's the kind of model that I think the elementary guys are going for, which is, you know, give you your basic desktop that has nice features, it's beautiful, but expect you to pay for the little applications that you need here and there. Now, their model is slightly different. It's not 30 bucks per application. It's a pay what you want. But they're certainly striving to ensure that developers are being paid for the software

that they develop. And I don't think that's unreasonable. And I think it's a good thing that on Linux, we're starting to see more of that prevalence of people being paid for work they do, which is really super common on the Mac. And it's just second nature on the Mac. You need a thing, you pay for it, right? But on Linux, there's that resistance. And I think it's slowly coming i think with elementary

improving their software store story and flat packing proving their software store story i think it will only improve on linux as well people will up the game i agree in fact that's a bit where i was going to go with this is what's great about linux is you can have stand-ups where something is tried like exactly that, like what elementary OS is doing, which you don't get with the Mac platform. But I thought I should give at least a fair crack at what you do get with the Mac platform. And I think it's that third-party ecosystem, hardware, software that they've really nailed, and it's a real tight experience. Right. I mean, we've talked about some of the advantages of the, you know, the SDK supports that are out there, that sort of very consistent platform and internal services that are all unified.

Yeah. But you remember at the start of this, I said Mac OS is a specialist OS, and Linux is a general platform that can be built into something. I think it's easier to go from a general platform into a specific thing than from a specialist platform into a general thing. And what MacOS specializes in, and I think you need to be really clear about this, is it fundamentally lives and dies by Apple's interests. When they're focused on other things like the iPhone or maybe AR glasses one day,

the hardware and the software at the same time suffer substantially. You can still see the knock-on effect of this evident on the platform today. This 2017 MacBook Pro is a living example of when they took their eye off the ball for a while. And that, I think, betrays the biggest weakness long-term of the platform, is it will always live and die by Apple's interest. What it specializes in is being a strategic product for Apple, which has complementary benefits for third-party ecosystems because of the dynamics of that.

But it doesn't mean it shall always remain that way. In fact, the transition to ARM could be a great opportunity to lock things down even further. Maybe every app now has to be signed by Apple. Maybe that's enforced at a hardware level in the Macs. So now Hackintoshes aren't as viable. And just somebody writing software, building it, and running it on a Mac becomes no longer tenable.

That's possible, as long as Apple's customers don't mind. They have to walk a line there, but you can guarantee they're going to walk the side of the line that we're not going to like very much. It's a very quantifiable, predictable ecosystem, and it has a premium customer base, but it will always suffer that fundamental flaw. Its master is not the owner. Its master will always be Apple because their whole platform affects your Mac. So whatever their long-term version of the market is will affect your Mac. Over time, the ecosystem will get locked down,

and general operating systems like Linux that have been built up into specialists like elementary OS or Ubuntu will really be our last refuge. They really will. And so that's why, even if we can never quite nail integrating with all the eGPU manufacturers or the people that are making cool editing hardware, even if we can never really nail that, the Linux desktop is going to be more important than ever

as these platforms continue to evolve. But it might only appeal to certain types of users. Different distros will appeal to different types. But I think when we're thinking about the future of Linux and trying to advocate for the platform or trying to market for the platform or trying to make it the best platform for gamers, we need to be realistic about what it can accomplish

and what it can do really well. And I think you just look at the success of things like open source software, like VLC and Blender and Firefox, et cetera, et cetera, et cetera, or Docker to prove how successful free and open source software can be and then ask why Linux is still continuing to shrink in a lot of areas of the market on the desktop. Like the Steam stats, I know everybody hates these, but they just came out again, and we're down again.

Meanwhile, the Mac's up. The Mac went from 0.14% to 3.15% in February. Is that possible? And Windows dropped slightly to 96%. Linux is hanging in at 0.83%. It was at 0.9 in January. It's gone down. But what's even more frustrating is it's been years now.

Protons have been around for a year and a half. Like it's been a while and we're still struggling to even crack these basic numbers that computers that barely ship with functional GPUs can kick our butts on. Like, we just have to be honest. There's just certain use cases we're not the strongest at. There's others we dominate. We're awesome at.

And, I mean, I think they touched on this very well in Late Night Linux. Computing moves on, right? We add new paradigms, and most people just don't care that much about the desktop. I mean, if you do, then Linux is there, and it's a great option that you can specialize and tweak and make your own. But for the most part, since you can get all these open source apps on pretty much every platform, or you're using web-based apps anyway,

why switch what you have now? It works. And I mean, most normal users don't care that much about the specifics of the tool that they have to use. Yesterday, you and I were kind of making a really bad comparison to like a Tesla. You could say like maybe the Mac is a Tesla. It's got self-driving.

It's, you know, it's very, very expensive. Any maintenance you do on it or like, you know, wheel replacement, anything. You got to go back to that specific vendor. Very expensive versus a, what would be like a, what would Linux be, a Jeep? What would it be, a truck? It's a versatile vehicle that you've kind of built up yourself over time.

You have easy access to the engine, you can swap things out, no problem. But if you want things like self-driving, well, there's maybe not a standard solution for that, and you're going to have to assemble it yourself. Right, exactly. And there's different products for different types of people, or some jobs require something that you can customize or some people prefer to work in an environment

that's more tailored to them. The Mac users are happy to set aside a budget of 50 bucks to solve their problems. The Linux users are like, well, that's just built in. Or if it's not built in, it's an extension away. I want to chime in here a little bit on the Mac conversation and the Steam survey conversation. First off, I've never used

a Mac to play Steam games on, but I have used Linux and Windows. I have a 2015 MacBook Pro that actually runs elementary Pop! OS Ubuntu pretty good. So you don't have to, I think they still make decent hardware. I think you could debate that with their latest butterfly keyboard and touch bar and all that. But that 2015 that I have is still a great little machine and it runs Linux great. And me personally, for my workflow, which is daily on a Mac, I haven't spent any money on the store to get any of these little personalization apps. I literally have the Adobe Creative Suite, Mumble, what we're recording on now, Audacity, Quassel, and VS Code. I wanted to have it like my workstation, like I, you know, where I have

menu meters and I like it's something where if I was going to live here, how would I really want to customize it? Like I can customize an OS that I prefer to use. That was my philosophy there. And it's clear that it's doable, but my home for that kind of mindset is better on Linux. And I kind of like the idea of embracing Linux as a professional power workstation. That's why I find where Gnome Shell's going and what it's fixing. hour workstation. That's why I find where Gnome Shell's going and what it's fixing. In the back of my mind, when I read that news story we did today, I hear this is going to make it a more professional grade workstation. Something that's faster, cleaner, leaner, better put together.

And I'm surprised, but kind of pleased to see that I think System76 is on a similar line of thinking. We sat down with two individuals, Maria and Michael. Maria is a UX architect, and Michael is the product engineer and one of the maintainers of Pop!OS. And I think they're seeing some demand from power users to add features to the Pop!OS experience, and they're starting with tiling in their 2004 release. So both Maria and Michael from System76 are joining us today to tell us a little bit about the advances they're making in PopShell. Michael, could you describe it to us?

PopShell is advanced tiling window management for the GNOME desktop in PopOS. That is a nice, concise elevator pitch. So we're bringing with 2004 version of PopOS. It's tiling inside of Gnome Shell. It's not replacing Gnome Shell? Yes, inside as an extension. This is going to be something that will really appeal, I think, to power users.

I'm kind of curious about that because normally when I think tiling, I do think sort of advanced users, maybe people who are using a more niche distro than Pop! OS or who really like customizing their desktop. But this seems, I played with it a little bit, it seems way more accessible. And I'm curious, what are the motivations for adding these quote-unquote advanced features to an otherwise very friendly shell? Everyone at System, or at least most people at System76

are using tiling window management. And many of the members in our community have also been using i3. So it's popular inside the team, huh? Yes. Maria, I thought Pop! OS was, this is my bias, but I thought it was a distro for new customers that were buying their first Linux laptop.

I don't think we necessarily defined it that way, but there is certainly a large subset of our users who are very new to Linux. Apparently, Pop!OS is also, we found out, very popular among those newcomers, if I can say so. So part of the motivation, I think, behind trying to make tiling more accessible and create maybe a better, more easily approachable system to actually enable not just power users, because power users obviously is our largest concern. But we also make it easy for anyone who wants to try using the tiling manager because we see a lot of value in it.

Sure. And so is it fairly easily invoked, Michael? If I'm a regular Gnome Shell user, can I just start experimenting with this and then turn it off if I don't like it? How is that process going to work for users? Yeah, there's a pop-up menu that you can toggle the features on and off. Nice and easy. I noticed that you've also added a launcher sort of facility, and I'm wondering, you know, what was the motivation there? The motivation is simply that if you're already doing searching for Windows, you may as well just throw in applications launching as well.

Sure. Okay. Yeah, no, that makes sense. I have kind of a technical nitty-gritty question that surprised me when I noticed it. And totally correct me if I'm wrong here, but is TypeScript the primary language being used to design this and write it? Yes, it's the only. Now, that is an interesting choice. I've heard a little bit about TypeScript, and what I've heard has impressed me. But it wasn't probably what I assumed going into it. I thought it'd be something super crazy like Rust or really kind of high level like that.

Can you talk a little bit about using TypeScript and why it works for this? Well, ideally, we would like to use Rust, but you can't for GNOME extensions. And using JavaScript is very painful if you're used to Rust. I bet. TypeScript is the best choice. Have you run into any issues getting that all to work nicely? I mean, I know TypeScript can, you know,

compile down to JavaScript, but any issues making that play nicely with GNOME's implementation of JavaScript? I have a said script, which fixes the little discrepancy. Oh, that's perfect. Yeah, that'll solve it. So what's the long-term play here?

Is this going to develop into a full-fledged tiling extension? Yeah, it's planning to do that. What should our expectations be going into the 2004 release? Much of what you would expect from i3, it should work in here with this extension. Wow. And is it portable outside of POP? Yeah.

Well, there you go. That's good to know. You just have to be on a more recent version of Gnome Shell. Definitely. Yeah. Well, that is the essential trick right there. If people can try it out, it might draw more people in. Maria, did you want to add something?

Go ahead. Yeah, I wanted to add to the question of what to expect. something? Go ahead. Yeah, I wanted to add to the question of what to expect, because I think one of the relatively big things we are trying to achieve is to make it easy to switch between floating mode and the auto tiling. So I think that's, we see that as somewhat of a big difference compared to how most of the other tiling we know managers work. So you're saying it's kind of nice because you can just easily switch in and out of the mode. So whichever one works better for that moment. Exactly.

Yeah, I can see that. You're in work. You got a lot of terminals up. Maybe you want that to be nicely laid out. And then you're at home on the couch and you just got a couple and it's easy to just operate normally. Can I apply it per workspace? Currently, you can. Oh, now that I would use the crap out of for sure.

Oh, now that I would use the crap out of, for sure. Because I'm essentially doing it now manually on Gnome Shell, on like Workspace 2 and 3, dedicated to email and chat. That's great, guys. Yeah, that's the plan. Well, thank you for coming on and just telling us a little bit about it. We just wanted to kind of get some of these questions answered, so that way we could talk about it intelligently and inform the audience.

I appreciate you taking a few minutes to share it with us. Oh, of course. Thank you for inviting us. Oh, yeah, guys, for sure. Maybe we'll chat when it's all done, too. Is there anything else you want to add? No, it's just only that we are very excited about this extension coming up, and we keep working on it

to make it the best user experience possible. We're excited about it, too. It's nice. Definitely. I think it's going to be a nice differentiator for Pop! OS as well, which is pretty cool to see these things developing over time, which I suspected would start happening.

Yeah, that was one of our main motivations. Yeah. Well, very cool. Well, I look forward to trying it out. Thank you. I'm feeling the hype. I want to try out that. I want to try out 2004. It's pretty nice. And just having this as a little extension on top of GNOME,

if you want some tiling, the auto-tiling, I was kind of impressed. It just popped up a couple windows. It has the nice little edges, you know, the gaps between windows. Even for being pretty new, not really released yet or ready for anyone to use it, they've come a long way and I think it shows a lot of promise.

So are we all cool with extensions now? Is that... I'm never sure. There's an official tool, Chris. I'm always behind. Like, I'm always the guy that's not for extensions and everybody's like, use an extension. Then I use all the extensions and people like, don't use extensions.

I mean, it might depend a little bit too. System 76 has, you know, has thrown some resources behind this rather than a random one-off extension someone wrote two years ago. Is that it? That's it. I mean, that's a part of it.

There are seemingly some fundamental issues, but I guess it does come down to, does it work for you, and do you have issues with it? It kind of feels like things are shifting, though, with Gnome Shell shipping that official app and whatnot. Mm-hmm. I'd try it.

I think I'd try it. Especially if I could just go to Gnome Extensions and just turn it on, turn it off. And if anyone does want to play with it, it's really easy. You just need TypeScript installed. There if anyone does want to play with it, it's really easy. You just need TypeScript installed. There's a make file that'll build it and install it for you. So that's really easy.

And you need to be on Pop! OS 2004. Well, aren't you already? I just assume. There's that little detail as well. I did get a sneak peek at the Superfan event, and I was really blown away by what I had seen there, just even in that early development. So I think this is interesting, especially with extensions and the thought, for me anyway, that KDE, for me, kind of comes with everything baked in.

You have to dig and hunt and find it and apply it and get it all set up. And that's cool. No problems with that. Whereas to me, GNOME kind of takes a different angle and here are extensions that you can add on to your system to add this additional functionality. So I think it's just two different ways to look at the same problem personally. Yeah. I mean, I'm glad they're developing it. I'm glad they're putting the resources behind it. It's open source. It's going to be cross distro. I think that's a net win overall. But I think it kind of makes our overall point that this is an OS that is really well-suited for advanced users,

much like of previous eras where there were people who loved to build radios or there were people who loved to tinker on cars. There's another generation now that is of a similar wiring and has found Linux, I think. It's interesting. I mean, how else could System76 do this, right? I mean, they're not going to build their whole own operating system.

That's probably too much to ask, you know, secure and implement all these features. And you can't really rebadge Windows or Mac. Yeah. But Linux have at it. I'm dying to hear what you think of it, Popey. I think it's great that they're innovating on top of all the work that others are doing and being able to deliver what their customers ask for.

So their customers clearly ask them for bespoke, beautiful computers that are high performing and have open source internals and they've delivered. And now their customers are clearly asking them for features in the desktop that don't ship by default in either Ubuntu or upstream GNOME. And so they're delivering them to their users. It's exactly what Elixir is all about. You know, you build a solution for the problem you have. They do have a pretty tight loop there because they're the support for the problem you have. They do have a pretty tight loop there because they're the support for the hardware as well.

So they're, in theory, probably getting a decent amount of direct customer feedback, maybe in a channel that's more unique to traditional distros. They have a really active Mattermost chat room, chat client thing where customers and users who are clearly running Pop!OS on something that's not a System76 hardware, and they're asking technical questions

and the developers are there. All the familiar names you know from System76 are in there answering their questions. So yeah, they have a direct line from their developers to users. Yeah, interesting. Would you give it a try? Yeah, interesting. Would you give it a try? Yeah, sure, I'll give it a try.

I didn't really get on super well with tiling window managers, and right now I'm in a mode of trying to stay as close to upstream stock Ubuntu as possible because we've got this little LTS release around the corner, so I kind of need to test that. Yeah, maybe an after release kind of thing. Yeah. Or on another machine,

one of my other many, many, many ThinkPads. Yes. Do you think it's daily driver for somebody who's willing to risk it? What, 2004 or? Yeah, yeah. Yeah, totally. I mean, the standard caveat is don't

because it's not production ready yet and it's not supported but if you're an adventurous kind of gung-ho linux enthusiast then yeah totally do it totally feels like things are shaping up pretty good um obviously the last couple of releases have been really solid but the zfs supports getting to a special state you've got this new release of Gnome Shell that's looking like it's going to be really good. Could be another one of those... I mean, you guys got to knock this off.

You got to ship it late or something. This is a good one. I've been using it for a while, and since I've formatted my drive, I've got a clean install, and by crikey, this is faster than it was before. It really is. It's much better. So I'm really happy with this.

Yeah, Wimpy tweeted recently, I think there's a lot to be optimistic about for the desktop, for Linux desktop. The other evening, I watched a live stream using Ubuntu Focal with Node 3.3.5 and Firefox. The streamer was using OBS on Ubuntu, and I was having a video call with a friend at the same time. And that's just it.

There are areas where Mac and Windows are going to have us beat for a bit, especially when it's working with an ecosystem of software developers and hardware. Industry that's already bought in. Maybe the elementary OS guys can help change some of that.

But over time, we managed to check off another use case box and then another one and then another one. And then pretty soon, we're sitting here in the future using our Linux boxes

to basically do everything all day long, every single day without issue. I mean, okay, maybe more people aren't flocking to use Proton, but that doesn't mean I can't use it. It works great for me. Yeah, and the truth of the matter is,

is you and I are able to do our jobs full-time in Linux, and it's not a compromise. It's not like bending over backwards like it used to be. I don't have any non-Linux devices. I mean, as long as you're including Android in there. Yeah, and whatever is on your router. I mean, who knows? It's probably some net BSD. Yeah. And

so for a lot of us, those boxes were checked a while ago, or for some of us just checked recently. Thanks for listening. Glad you just found us. Every year, it's like we check another one or just a few more days. Sometimes it happens so subtly, we don't even notice. So if you take a really, really long-term look at it, it's the best it's ever been right now, and it's pretty great. But if the momentum even continues at the pace it's been, eventually a few more boxes will get checked off. Another thing. I mean, look at these improvements with, like, Pipewire and GNOME.

Who knows what's possible? Mr. Payne, before we go, what do you say we do some pics? Oh, yes, pics. I like this first one. Cat, cast all the things. Allows you to send videos from many, many online resources

to your Chromecast. Or local files, it turns out, too. Yes, I mean, so there's some tools that are great for sending local files. There are other tools or websites that work well if you want to stream from some URL that you've got, like maybe a HLS stream to the JB feed. Yeah, that works great. So that'd be a really super easy way to just take that URL and just get it blasted to your Chromecast. Yes, without, you don't have to futz with Twitch or YouTube or any sort of thing that might show you ads or take over your display.

Chromecast can just play links. And much like YouTube DL supports a lot more than YouTube, Cat seems to support all kinds of online services. Yeah, so it has like integrated support for YouTube DL, so you can paste a YouTube link. Oh, it's literally using YouTube DL. Yes, and then YouTube DL will go, you know, grab the actual URL to that video,

and then Chromecast plays it directly, which again means you don't have to fuss about with using YouTube. Yeah, or no browser with an extension, no Chrome required. Yes, and if you have an MP4 file that you'd like to play, including subtitle support, let's say you've got those subtitle files already sitting on your disk, Cat can do that too. And that's just over the land then?

Yes. Hmm. I wonder if that even goes through the Google servers when you do an MP4 file like that. No, I think it just talks over, you know, the Chromecast advertiser itself

over with MDNS. And the syntax is super simple, right? It's like cat command and then what? Yeah, you can install it with pip because it's Python and then

cat, it's got a helpful little, you know, help doc to tell you, but cat cast and then either your URL or a path to the file on your file system.

How does it know which Chromecast to send it to? It will pick one. I don't know how it picks the device. It might be like a Google Home type setting involved. Yeah, yeah. But it does have support for specifying the Chromecast. Okay. So you could just give it a name of a Chromecast.

It's all using probably MDNS, right? Cool. Nice one. That's cat. We'll have a link to that in the show notes at linuxunplugged.com slash 343. You can also tell it to just display, like if you want to display a web page on your Chromecast without having to

cast from your laptop, say, it can do that. And there's a config file so you can make aliases for your Chromecast if it's got a long name or something like that and you just want to use that all the time. So yeah, this is my new favorite Chromecast tool and I strongly suspect it will be a part of my daily life. That's pretty cool.

All right. I'm going to get that, too. Thanks for finding that. And then Mr. Bacon rolls in with the Silver Searcher this week. It's a code-searching tool similar to ACK, but says it has a focus on speed. I like that. I like it fast. Yeah, so—

My searchers. Yeah, so a buddy of mine recently reached out to me and he was like, hey, have you guys ever used Silver Searcher or covered it in a pic? And I was like, I don't believe that we have. But essentially, it started off as a clone of ACK, and they claim that it's five to ten times faster. So get on it. Get your code searching done quickly. What do you think, Wes? What do you think? You're going to do some code searching?

You have something else you like. You want to do a third bony pick? Yeah, no, I definitely like the Silver Searcher. It's a fantastic program. But, I mean, Chris, you've just got this rule now. Silver Searcher has a bunch of C involved. Oh, yeah, no C. You don't allow that.

You're Rust only as you make clear. Rust is what everybody knows. I say that. There's no C. You don't allow that. You're Rust only as you make career. Only Rust is what I, everybody knows I say that. There's no C code allowed in the studio. So for that, I have RipGrap, which is in exactly the same sort of vein, but is made in Rust. And it's so fast.

It's really just great. And because it's Rust, it's really easy. The release page, you can just go download a binary, stick it on your path somewhere, and away you go. I can't believe you managed to bring it. You know, we almost went the whole episode without our epic Rust watch. You know, technically, you got to warn me. Epic Rust watch on Linux Unplugged.

An update from Wes Payne on something else that uses Rust. There we go. See? It's obligated. Ever since we hired that new soundboard guy, it's in the contract. Well, that brings us to the end of this week's program. Thank you so much for listening. Thank you to the mumble room for stopping by and hanging out with us here on a Tuesday.

Next week is WSL Conf, which has just been converted to an online only conference, but it's come virtually say hi to us. We'll be there. Yeah, we'll be there. We're still going to be live. We've decided we'll still do the show at the regular time. So Tuesday.

No excuse. Show up. jupiterbroadcasting.com slash calendar for the lifetimes. I'm at Chris Lass. He's at West Payne. The show's at Linux Unplugged.

Go check out the Ubuntu podcast and we'll see you right back here next Tuesday. The Unplugged Program. Dabba dabba dabba dabba dabba. All right, well, so JBot was down for a bit. It's back up now at jbtitles.com. Yeah, we didn't lose anything, so go vote. Oh, really?

I keep suggesting we've only got, it looks like, six of them right now. Okay, okay. So we were talking about this earlier. I want to kind of go around and see if maybe people in the Mumble remember. Popey, what was the first expansion slot standard that you recall when you started working on computers that you dealt with? A standard or just a slot that I would plug things into? The slot type. Mine was when new bus, and then I remember ISA, I had lots of ISA cards.

That was sort of when I got into it. Yeah. I watch retro gamers building 486 machines, and every time I see them with a Visa local bus on the end of the ISA slot, it makes me sad because I backed that, and I bought loads of Visa local bus cards, like super long ones, and that was a dumb idea. I should have waited and got PCI

because that was obviously the winner. But yeah, ISA was probably... I remember my first PC was an Epson 8086 with a turbo button and it had ISA slots. And what were you putting, a modem in there or a sound card or something, a sound blaster? I actually put an external floppy drive because the internal floppy drive was five and a quarter inch. So I had to get an external three and a half inch

and had to put a card in for that. Right, sure, you got to put a card in for that. Right. Sure. You got to put a controller card in there with a big old cable to connect to the external drive. Oh yeah. It was beautiful. Beautiful white, pure white floppy drive.

You need that full bus speed when you're reading those three and a half inch floppy disks. In fact, that first PC I had came with an MDA graphics card, monochrome display adapter, which did no graphics. It was text only. And I managed to acquire from where I worked, my very first job out of college, there was a Hercules graphics card kicking around. It was just kicking around in a cupboard, and I got that.

And the upgrade from text only to graphics was just like chef kiss. So good. Oh, my first big video card i invested in was a matrox agp oh yeah it had an expansion ram module as well and it came with the first tomb raider pc game as like a see what you can do with your upgrade i was so dumb i made the mistake of using of choosing a motherboard that had visa local bus but I also made the mistake of buying an IBM PS2 that had the micro channel architecture so I couldn't buy any cards for that thing that was just like came with what it came with you're never going to upgrade this thing yep MCA buddy MCA which they thought was going to be the big

thing yeah I think my first graphics card was the uh riva tnt2 or the i think it was the rage tnt2 for agp yep um and then shortly after that it was a pci trident eight megabyte card so i could run linux because it wouldn't run into trident video Trident video cards. Oh, that name I've seen a million times. You know, the IBM PS2 was released April 1987, 32 years ago. Wow. Yeah. Which, as the name suggests, is where the PS2 port came from. And it was a rocking way to run Windows 2.0.

They also had a port of AIX to the PS2 back then, which sounds horrible. Could you imagine AIX on a 386 processor? No, I don't want to. DOS was bad enough. It's a better world now. Yeah, it is. What about you?

Do you remember what your first interface was? I remember the thing about AGP, the thing that dawned on me was, oh, they're going to be changing this all the time. Like, they just kept getting like... Yeah, I think AGP was the first time I was really sort of bought in and had got like an expensive card

to attach to my machine before I just sort of stuck with the defaults. What about MumbleRoom, Carl or Brent? Either of you have old PC memories or byte? I stepped into the computer world when Windows 95 was around with the AGP. And I still have

an AGP video card around. Yeah, that's coming in handy one day. You might need it. You never know. Surely there's like a Thunderbolt AGP adapter. It still has the VGA and the SVD out on it.

Oh, yeah, of course. Man, VGA is still around. I think I had an S-Video out once my AGP card at the time, and I had an old TV I was using with a second monitor. It was gross. Yeah, yeah. I remember when I finally got my hands on a Sony Triniton CRT, which was the flat CRT.

Yeah, really nice. Really nice nice screens yeah i remember those i still have a trinitron at my mom's house my very first tv i bought my with my own money was in 1988 and it was a 14 inch sony trinitron and i watched the matrix on dvd when that first came out on that 14 inch widescreen letterboxd oh happy days