LINUX Unplugged - 295: Stay and Compile a While

Episode Date: April 3, 2019

Is there really any advantage to building your software vs installing the package? We discuss when and why you might want to consider building it yourself. Plus some useful things Mozilla is working on and Cassidy joins us to tell us about elementary OS' big choice. Special Guests: Brent Gervais, Cassidy James Blaede, and Martin Wimpress.

Add this to the list of reasons to never run Windows. Office Depot's name's getting dragged through the dirt this week when it was revealed that they will be paying out $35 million in refunds for tricking people with bogus malware scans. Yeah, I don't know. This is just, it's been so long since I've bought computers this way. It felt so foreign. But you got to feel bad for people who just want to get a new machine

and end up as part of this scam. Windows users would come in. They'd say, take a look at my computer. It's running slow. That's because it runs Windows. I mean, I'm sorry. I'm sorry. I know a lot of people.

I'm just having fun. And so then they would bring it in, and Office Depot would scan it, and they had the software rigged to show false detections, false threats. And then they'd be like, all right, $300 to fix that. I mean, how could you not, right? You don't know.

You're bringing your stuff to Office Depot, so you obviously have no idea how computers work. I know. It's such a shame, though, because you could grab a great SSD for that price, and that would make the machine way faster, without a doubt, right? Of course. Also, maybe put Linux on there. This is Linux Unplugged, episode 295, for April 2nd, 2019. Oh hey, y'all. Welcome in to your weekly Linux talk show, Linux Unplugged. This week, it's a special episode. We are packed. My name is Chris.

My name is Wes. Hello, Mr. Payne. So, aside from a full cast of characters this week, we have a big batch of community news that we'll be getting into. Mozilla is doing quite a few useful things. We'll tell you about them this week. NGINX wants to reassure you, everything's going to be okay. And Raspberry Pi has new accessories for you to buy, and they got me with one of them.

I'll tell you which new Raspberry Pi accessory I picked up this week. And we'll get into some big news for elementary OS and the slick update the Ubuntu Mate project has for Raspberry Pi users. There's something really sweet in there. I don't know. For me, it's like a game changer. We'll talk about that in a little bit. And then towards the end of the show, assuming we have some time, we have some deep, deep

thought-provoking feedback to get into. Well, at least some decent feedback. I don't know if I'd say it, deep, thought-provoking feedback to get into. Well, at least some decent feedback. I don't know if it's... I don't know if I'd say it's deep and thought-provoking. Stick around and find out. Yeah, maybe that's the way to put it. But as we're getting close to episode 300, I feel like we're gonna... we've got some great stuff lined up, including,

like I mentioned earlier, a great cast of characters. In our mumble room, joining us this week, we've got Billy, Brandon, Brent, Cassidy from Elementary, Cubicle, Nate, Foursquare, Fundament Utils, Mini-Mech, NE1, Nero Burner, Skull Leader, Social Happiness, awesome to see you, Emma, TechMav, oh man, great to see you, and of course, Mr. Wimpress, time-appropriate greetings, Mumble Room. Hello, everybody.

Good evening. How's it going? Wow. I'm just looking at mumble room. Hello, everybody. Good evening. How's it going? Wow. I'm just looking at the waveform there. This is incredible. That was great. Hello, everybody. It is really good to have you here today.

Our main duty for this episode this week was to avoid putting April Fool's stories in the news lineup. That was job one. And this is Tuesday. April Fool's was yesterday. We this is Tuesday. April fools was yesterday. We were really close, really close. So I had to look and, you know, had to reject a lot of stories. We all, we're, we've got a team now. We've got a team now, including, including folks that are just dedicated to research in this stuff. Like

it's a team. It's pretty great. And there was one story that, it's fake, but it really made me chuckle. Did you see this one? I did. You did see this one? Hackaday had their April Fool's story, GNU slash Linux distros drop support for Vi. It's about time. No more VI. The war is over. They say the reasons for this move aren't entirely clear to us, but from

what we can see on the GNU slash Linux mailing list, the confusing modal interface and the fact that it's been hard for novice users was the lead cause of the decision. Well, you can speak to that, Chris. I thought you would say that. You know, actually, before

I realized it was, because, I mean, I was reading a bunch of fake stuff and real stuff, right? There was still real news that day. Yeah, which we'll get to. I thought this might have just been, like, deprecating VI compatibility mode or something. I thought this was a real change. So they definitely got me.

I've never installed GNU slash Linux. Yeah, and, of course, the only thing that would have made the story better is if everything was switching over to Nano by default. We were here earlier in the studio working on a configuration file, and Wes was standing next to me, and I had to make a point of not only do I, when Wes is next to me, not only do I edit Nano, but I use Nano-W because you've got to have the wide Nano. You've got to go all in. So that's what I do. You're a Nano expert. All right, well, let's get into the real news this week. There is actually some news. Not all of it is fake. So kicking off,

we have an announcement for the Apache Foundation. They have hit their 20th anniversary. So another big happy birthday this week goes out to what they bill as the world's largest open source foundation, which provides 20 billion

worth of software for public good at 100% no cost by now. Happy birthday, I guess. Yeah, I mean, we should take some time and appreciate it. There's a lot of really good software under that and supported by that foundation. So big thank you to the Apache Foundation, and happy birthday.

They were founded in 1999, which makes me feel like an old man and makes me realize that that was 20 years ago. Which we will just move right on from that and continue in. Here's to many more of those years for both of you. I hope I'm still here talking about their 40th. Wouldn't that be amazing? I'll see you there.

And then let's play this clip. Perfect. All right. Hello, future Chris. Now, let's move on to the Mozilla is useful category. And I like this. I like looking at the stuff Mozilla is doing that is either advancing the web in a good way or is maybe even a project outside of Firefox and Thunderbird. Not really context.

We talk a lot about Mozilla. Firefox, and Thunderbird. Not really context. We talk a lot about Mozilla. So first for Firefox users, this is a fascinating one that I think we're all going to agree is a very, very good feature they need to just adopt.

Don't even experiment. Just go with it immediately. And that is reducing notifications for permissions to have notifications on your computer. Websites, how do I put this? Because I'm butchering this. Everyone wants to send you push notifications.

And they want to notify you about their intent to send you those notifications. They say, in the interest of protecting our users and the web platform as a whole, we plan to experiment with restricting how and when websites can ask for notification permissions. You can see a little bit of their motivation

as they write, anecdotal evidence, aka people complaining, tells us that there's an issue with notification permission prompts among our user base. As we browse the web, we regularly encounter these prompts,

and more often than not, we become annoyed at them, or, and this was true, I remember distinctly for some reason, the first time I saw one of these, because I accidentally clicked aloud, and then suddenly I was getting notifications, you know, showing up on my browser and everywhere that I was signed in. They're just confusing. And I don't think, especially people already

familiar with the way the web works, they're sort of this weird new out of band system. And you never know what they really want to use it for. So you're asked to make a decision before you even understand what the content is that you're going to get delivered. And then you combine it with the cookie notice policy and like the user data notice policy or the, hey, would you like to sign up for our mailing list policy thing that comes up? Or the, hey, don't use an ad blocker. All these things that come up at once now when I load a website are horrible. Nothing like when I used the web when it was originally started a website are horrible. Nothing like when I used the web when it was originally started.

These are awful. But it's more interactive, Chris. Oh, yeah. I've got to have me in my notifications. That's for sure. You know, they do have some interesting data here. So it's not all just anecdotes. From people running the Firefox beta from December 21st from 2018 to

January 24th, so about a month of data. There was 18 million prompts shown. And these are just people on the beta and not even 3% were accepted. It's negative in the freedom dimension. Not even 3%? Ouch. Not even

3%? I know, right? That's devastating. 18 million prompts and not even 3% of them were accepted. That's devastating. 18 million prompts, and not even 3% of them were accepted. That is devastating. That doesn't work. Wow. Well, I mean, there's always architects in the chat room suggestion.

Break it all, just turn off JavaScript. And that's one way. Yeah, there you go. That'll solve the problem. Actually, you know what else you could do? You could switch everything to AMP. That'll solve the problem, too. Problem solved with AMP, everybody. Now, here's something else that Mozilla's doing that maybe one day could surplant,

although it's probably not likely, but could become at least a competitive alternative to Electron, if I'm not misreading this. And Mozilla's bringing up the point that with WebAssembly now becoming a thing and all major browsers supporting it, more and more folks are trying to develop applications beyond the browser.

They say doing that provides a fast, scalable, and secure way to run the same code across all your machines. But we have yet to have a solid foundation in which to build upon. Code outside of a browser needs a way to talk to the system, a system interface, and the WebAssembly platform doesn't have that yet. Now, Mozilla has been working on a runtime for the desktop

that allows you to run WebAssembly applications as a desktop application. But you needed, I guess, a standard to pull this off, and that seems to be what they're coming out with is this WASI standard. Am I getting that right? What is it? W-A-S-I?

There you go. W-A-S-I. Not wasabi, but WASI. That might taste better. Yeah, that might. What are they trying to accomplish here? They're exploiting the fact that we've made WebAssembly, and it doesn't run on real hardware, right?

We just described this abstract machine that we had already worked out how to run and are actively working out how to run fast in our browsers. So this is just all the little extra details you need when you want to think, okay, well, this could probably run a lot of other places too. Yeah, much like we do with Electron, only using the Chrome runtime, essentially. Or it's the reverse of like Java applets and Java and network.

Yeah, it's nice to see Mozilla doing these kinds of things. Like this, yeah, let's figure out a way to do this. They have been actively involved in all kinds of great WebAssembly stuff. And actually, they've got some of the best documentation. If you're getting started with WebAssembly, you want to learn about some of the concepts. Yeah. We did also recently see Lucid,

which is the Fastly Corporation's WebAssembly runtime. And there's like a couple other open source projects. So there's clearly a lot of stuff going on. It's nice to have a standard now, early in the process, so we don't end up with a whole bunch of divergent runtimes. WebAssembly is coming. And it's going to come to your desktop, too. And maybe that'd

be good. Maybe it'll make a good competitor to Electron and keep both of them very I guess, actively engaged in pushing the boundaries. I mean, yeah, one thing's for sure. It's rare that we see things fade away. We just have more things on our systems.

Yeah, things don't really go away anymore. No. They're too big to fail after a certain point. Now, there's one last thing that Mozilla, along with, I think, the folks behind Vorbis are working on. It's a Mozilla, along with I think the folks behind Vorbis, are working on it. So Mozilla employee who was involved with the Speaks codec and also Opus. So this guy really knows his stuff. Those are some credentials.

And they're working on LPCnet. This is a codec in which they can transmit audio in as little as 1.6 kilobits or kilobytes. Wow. They're doing this with essentially a synthesis. So instead of actually capturing the entire audio waveform, it's almost like taking audio. This is a super rough layman translation, but it's almost like taking it and making it an SVG.

Yeah, you're trying to capture just enough detail to recreate it, not perfectly, but be able to describe it in a way that's good enough. We have Mr. Bacon here, Mr. Cheese Bacon, now part of our team. Cheese, when I, I'm going to get this, I'm going to butcher this, so I want you to kind of help explain this to me, but an SVG is, it's a vector graphic which is scalable, and so it's like, it's like the math for the image, and then it renders the final image, and you can rescale it to any size

you want. You can transmit a tiny SVG, but you could have a huge image that you produce from it, correct? Correct. So I mean, it's a vector image, and stands for scalable vector graphic. And it stands for scalable vector graphic. And it'll allow you to, you can support transparencies with it. Once you've created the SVG, you can open it up in a text editor and edit it. You can actually add like early animation type work to it. SVGs are really cool. And you can reduce their size.

And since they're vector graphics, of course, they can be scaled infinitely. So if you needed to use that same SVG on a billboard or the side of an 18-wheeler or wherever you needed to use that image, you could scale it without losing resolution or having an issue where it looked aliased or, you know, something pixelated or something like that. Now, I admit my analogy is weak here, but now imagine taking that same concept and applying it to audio where you take the data and you vectorize it and then you can re-synthesize it later on and you can improve the recreation process down the road.

Like you can have better and better models and better and better ways to do speech. And they're getting there now. They're getting there to the sense where they almost can recreate speech using, guess what, machine learning, to have something that sounds very accurate. So I'm going to play you a demo here. Here is the old technology using a great codec. So it's still really good compression, but it's old synthesis. The birch canoe slid on the smooth planks.

Glue the sheet to the dark blue background. Now that's, you know, it sounds like a speaking spell a little bit, but the one thing you have to give it is it's like two kilobits of audio a second. It sounds like a really bad phone call, but it is kind of understandable. And you could use that to get speech in the worst case connections, which is really impressive, because that brings in all kinds of

like packet radio, possibly, like I don't know, like all kinds of solutions. Here is what this LPC Net project that's getting contributed by Mozilla employee that is bringing this further using the machine learning I just talked about. This is their more current sample now.

Add the sum to the product of these three. Thieves who rob friends deserve jail. Now, I don't know what that is about, but that's just 1.6 kilobits of data. Wow. To send that. That's pretty tiny. That's tiny. And it sounds about, but that's just 1.6 kilobits of data. Wow. To send that. That's pretty tiny. That's tiny.

And it sounds really, I mean, not bad. It's totally understandable. Like, I'm going to play just a... Add the sum to the product of these three. That is much better than... The birch canoe slid on the smooth plaques. Right? That's a...

Much better. And you could see where later on, the decoder and the thing that reconstitutes the speech could get much, much better, to the point where it could sound like a podcaster. Oh, yes. Honestly, those are pretty good. Right now, and it's kind of interesting, they've done a great

job of explaining it. All the stuff that most people are familiar with are called waveform coders, and then this is a vocoder. Yeah. And waveform coders are stuff like Opus and all the stuff you might use for music. Everything we're familiar, right? And they basically try to take the natural waveform that is sound and transcribe that as closely as possible and capture all of that detail.

Now, maybe with MP3, you start throwing some of that away, but essentially you're recreating the waveform, right? Vocoders, on the other hand, as you said, they're really synthesizers. They try to, like, in this mode, you have the encoder, and it tries to grab a bunch of high-level stuff, like pitch, shape of the vocal tract. You have this model that you fill parameters for. So instead of trying to, you know,

actually track the waveform, it's more like describing, you could think of a really high-level version would be a transcript, and then feeding a transcript back. Yeah. Right, but it's not, you know, this is a little lower level because it's trying to recreate it at a more accurate level with

tone, character, voice. That's basically what they're doing here, and that's, I think that demo really shows how you can hear it. The difference is they've, now that machine learning has built some high-quality synthesis algorithms, they can apply that here. Before, all this still worked.

Vogue coders had been around since the 70s. Right. They just sucked. Yeah, they sure did. They were a speaking spell. And now they're getting kind of legitimate. It's for people who like to mess with computers. So the folks over at NGINX are protesting a little bit too much.

But they're getting a lot of heat. We had a pretty mild reaction to the NGINX purchase. Yeah, I thought so. You know, I didn't get Wimpy's hot take on it. I don't know, Wimpy, did you have any hot take reactions, first thoughts? What are other really zippy terms people use for immediate reactions?

What are your first reactions to NGINX getting purchased by F5? Huge controversy, Wimpy? No, I think it speaks to the success of Engine X and how it dominates that space now. So I think congratulations to Engine X for all of the hard work that a lot of the people there have put in, and I hope

they're going to benefit from that acquisition. I'm sure some of them will be getting some kickbacks, and I think it will cement Engine X's position in the market as the sort of Swiss army knife of, you know, proxying, reverse proxying and fast serving of assets. Right. Well said. Unfortunately, apparently you are a completely unreasonable shill, according to the open source community, because they are upset.

Nginx have had, you know, the open source and pro version of Nginx for years already. So that open core model is being established for some time just because it's changed ownership. I mean, what does that change? Nothing. I don't think anyone was really thinking about the company behind it before, and suddenly we are. It's the license, stupid. I mean, that really should be the tagline there.

But NGINX writes on their blog, we've heard it loud and clear. We need to continue to demonstrate our commitment to open source. They go on to say, let me take a moment to be very explicit. F5 is committed to not only supporting the development of NGINX, the open source edition, but as well as other NGINX open source projects, but also increasing our investment over time.

Just to repeat that, F5 is committed to NGINX's open source technologies, developers, and community. I swear! Okay, I might have added that last part, but like they're really hitting this hard. They are, and they make an interesting point here. We're also seeing increased acceptance of mainstream companies as stewards of open source technologies.

Microsoft bought GitHub, VMware bought Heptio, and IBM bought Red Hat. The assumption that big fish eats little fish means big fish kills open source project is outdated thinking. Going back to your point, Wimpy, that's the result of success to a degree. As they become more successful, large market players scoop them up. And also, before the acquisition, the pro version of NGINX provided the funding that organization needed such that the open core version of NGINX was available as open source to all of us, so that your website and my website and millions of others are able to run with a well-supported cutting edge web server and you know provide all sorts of nifty stuff like object caching and if you are at web scale

you're obviously going to be making some money. And those pro features you can buy into are probably of, you know, importance to you and not so important to the likes of you and I. So, you know, I'm quite happy for, you know, these organizations to be producing open source software and have a markup on, you know, an enterprise version that, you know, means that they can continue to fund the open source stuff.

Yeah. In fact, this live stream we're doing as we record the show right now is being powered by Nginx in multiple different ways. It is a Swiss army knife of tools, and F5 is savvy for scooping it up. It shows that they're watching the market and they understand where there's overlap there.

So there are moments where it's a little uncomfortable. I think that's particularly in open core, I think it's hard because the test is how engaged are they with the open source project? Because if they're just making money off proprietary features and not developing, then that's just... And then just dropping a gzip or a zip file somewhere, and okay, we've complied with the with the license has technically been complied with.

So it can be hard, I think, especially as an outsider, but I don't think that's ever been the Nginx community. No, as Nginx were transitioning to their pro model, it used to be the case that you had to compile your modules into the Nginx binary at build time. And they went to great lengths to make that a modular architecture as they moved to the Pro model so that they could continue to develop the core of Nginx

independently of the modules that make up the Pro enterprise version. So they've demonstrated in the past that they're committed to maintaining the open source roots of the project. Yeah. I think really the solution here is for Nginx to have said nothing. Just look at the absolute outrage

over the GitHub purchase. And they just, as everybody was doing their mass exodus to GitLab, they just stayed cool. They didn't say anything. While GitLab was doing blog post after blog post about how to migrate and tweeting like crazy, they just sat there and didn't say anything. And now nobody talks about it. It's like, it's going to happen. And by the time the deal is finalized, it'll be like, oh yeah, right. That.

Okay. And like the only thing that can really prove their point is what's going to happen in the future and their continued involvement. So yeah, they say that here in the blog post and it really, that's, that is, I think, your and my overall opinion on these big acquisitions is sit back and see what happens. Just watch what happens

and then make your judgment then. And the great thing about open source is it comes back to that license, stupid. You've got that license.

So if you don't like what they're doing, you're good to go. So let's talk to Mr. Cassidy who joins us from the elementary project, who just had one of those, like it had to be a calculated decision, Cassidy, I imagine, like, do we do

an April Fool's announcement or not? Like that must have been right there, a hard call. Yeah, we've done April Fool's in the past and they've been fun. But this time we had some actual news and we thought, why not just release that on April 1st? You know, have some fun. So you guys are announcing that the App Center will be using Flatpak. And I think it makes sense, you know, local repository that you curate yourself and all of that. I guess my first thought is, why is this an announcement in a sense? And not to take away from it, but why is this an announcement to begin with? And walk me back from that.

Like, why does this matter? How will this impact users and all of that? Yeah, so ever since these newer packaging formats started showing up with containerization and sandboxing, people have been asking elementary, when are you going to jump on board? When are you going to jump on this bandwagon. And we've taken the approach of being, you know, kind of slower, methodical, intentional about not just jumping on a bandwagon. And to your credit, I'd say I've seen the project attend events from both communities, from the Flatpak and Snap community. So you've been active in both communities completely. Exactly. Yeah. We've been kind of surveying the technologies,

watching them really closely and engaging with people, but not just jumping on a bandwagon. So the difference now is that we've been, you know, since about 2015, we've been working on these technologies and talking to people. And I think Flatpak is really in a place now where it's getting more mature. We understand it a lot better and we're starting to see those advantages. We're also starting to grow with App Center to a point where the advantages of a format like Flatpak would benefit both us and our users. So it really just is the right time to jump onto one of these new formats. And after a lot of back and forth and discussion with both our first-party developers

and our third-party developers and the people who are making these formats themselves, we decided that Flatpak was most closely aligned to our needs. So why not just do like a little quiet, sneaky implementation on the backend? Why do the whole announcement about it? I think it's, you know, we've started to actually do the sneaky implementation on the backend in a few places. You know, there's a branch of App Center that supports Flatpak already. And, you know, people are going to start seeing those things and start talking.

So we might as well get out in front of it and really be intentional about the message around it. I was surprised because I wasn't sure how you felt as a project that it seemed like some of the what would be otherwise kind of like elementary OS exclusives were getting flat packed up by the community and spread around to other distros. And it's kind of like at one point, that's great because it means more open source software, native applications on the desktop are available to everybody. But in another way, it's not as great because these applications are designed to be part of a larger environment and they kind of feel out of place on other desktops and I feel like it doesn't do a full service to the complete elementary OS experience

and not to mention you get like these probably crazy bug reports from somebody on Fedora Rawhide that's trying to get somebody who's making an elementary OS app to fix their problem so I thought maybe as a project you guys were not a big fan of that arrangement.

There's kind of three aspects of these newer containerized formats like Flatpak. There's the packaging itself, you know, just like Debian packaging or whatever. There's the containerization or sandboxing. And then there's the cross-platform aspect. We're really adopting that first of three elements to start it's we're just adopting it as a packaging format to start um and that that means it should be really seamless that nobody should really notice any differences on elementary os the sandboxing part is the reason we're adopting it you know that it's the better

the better uh security and privacy we've been doing a lot of work on that front. And that's the kind of future we're moving towards that, A, it'll make app reviews easier on the app center side. And B, it also just makes users' devices more secure and respecting their privacy more. The third aspect, the cross-platform part, isn't really a concern or a driving factor at all for us with elementary. It's neat. It's cool. It's a thing that exists there. But the cross-platform part, while it's very loudly one of the primary reasons for these formats to exist, it's not actually one of the reasons we're adopting it or one of the things we're really investing time in. Yeah, I understand.

things we're really investing time in. Yeah, I understand. I think there's probably some advantage, too, that Flatpak is very closely tracking GNOME features and GTK features, which elementary OS uses a lot of GTK, so there's some advantages there, right? When they make a change

that affects the overall GTK ecosystem that is overall maybe like something very nice, like a new way to access a device or something like that, you as a project automatically get to take advantage of that. Exactly, yeah. It's really closely aligned with the stack of technologies we use. As we work more and more with folks that work on GNOME,

we see projects like GTK actually get better because there's multiple projects using GTK. If GTK was only better because there's multiple projects using GTK. You know, if GTK was only used by GNOME, it might be a little bit restricted in what, you know, how it can be used. But because you have XFCE and elementary and other desktops that are using GTK, it makes GTK better. We feel the same about Flatpak. You know, it's not technically a GNOME project, but a lot of people from GNOME are working on it. And so by having yet another project kind of throw their weight behind it, it will make the technology itself better as well.

So this is going to be something you guys work on probably for a period of time because it sounds like it may not even land as the default way to package software for Juno itself. So this could be something that takes a little while to implement? Right, yeah. So today nothing is changing as far as how apps are being developed and distributed on elementary OS. It's still using the app center dashboard with the Debian formatting, but we wanted to get this out there to kind of let our developer community know what's coming and start to get feedback from them as we work through these issues. So there's two main things we have to do. We have to implement it in app Center itself, the client-side app, which work is

progressing there really well. And then we also have to implement it on the backend, what we call App Center dashboard, where developers actually submit their applications. That work hasn't started yet, but we're starting to look at that and kind of investigate how we're going to do that. So it's going to be a long process, and it may not even come to Juno itself. As an update, we're still deciding how that migration will work. But we wanted to get the news out there for our app developers to know what's coming. So I guess my last question would really be with this. Is there anything in this that really inherently prevents elementary OS from continuing to just use snaps if the user

wants to install a snap themselves? No, of course not. This is, you know, we're deciding to use Flatpak for the out-of-box experience of AppCenter and elementary OS. Developers are always free to, or developers or users are always free to do whatever they want on their own systems. You know, they can install Snap or they can install Debs or they can sideload whatever they can compile from Tarballs. So, you know, whatever the user wants to do, it'll still work. This is just about the supported default experience for Map Center. Right. And in a sense, it's an implementation detail, isn't it? It's really not a user facing concern. Exactly. Yeah. If we do our job correctly, then most users shouldn't even really realize what's going on under the hood.

Yeah. And it makes total sense. You know, you want your own repo. You want to curate it yourselves. You've got your own app center. You, and it makes total sense. You know, you want your own repo, you want to curate it yourselves, you've got your own app center, you want to publish applications for that. Like, it totally makes sense.

And it seems like you guys are thinking a lot about security too, which I really appreciate. Especially as this is the distro I recommend to friends and family.

It does seem like a particularly good match in that sense. The same abstractions that are trying to fight the same problems. Yeah, exactly.

Well, thanks, Cassidy. Thanks for joining us and explaining it to us. And we'll have a link to the entire Medium post, which goes into a lot more details. Was there anything else you wanted to mention to make clear before we continue on? Or is that, are we good?

I'm probably going to write up a follow-up post about this, but we've been getting a lot of great feedback on social media. You know, people are having concerns about, oh, you know, I've had issues with accessing external devices with flat packs, or I've had issues with app sizes being huge or downloading giant SDKs with flat packs. And I do want to clarify that the way we're implementing flat pack will be very, very like

specific and considered implementation. So to start, you know, we're going to make sure that those permissions can be very permissive, because we already review and curate and test all apps that are submitted to App Center. So we don't have to have the locked down permissions to start that might break things like external devices. So to start, there shouldn't be any big difference there as far as issues that other people have seen with other Flatpak implementations. And things like app sizes. Similarly, we'll have a specific SDK that developers must target when they submit to App Center. So it's not like you'll download the entire GNOME 3.22 SDK for one app. It'll use the SDK that's already installed on the system,

so it won't have a 2-gig download or whatever, like people have seen with a lot of things on like FlatHub. So it's, you know, there's Flatpak is just a technology. It has lots of different implementations and we're being really careful with our implementation to make sure that it solves these issues. You're going to get it right, it sounds like. That's a very good detail

because it's a totally separate implementation. And so if you might have had a bad experience, that's no reason to write off all the hard work you guys are doing. Yeah, all right. Well, Cassidy, thank you very much. Cassidy from the Elementary OS Project. Check out the link in the show notes. It was good to have you with us.

Thank you, sir. You're welcome to stick around for the show. We'll just cover a couple more community news stories before we get out of here. First up, the Raspberry Pi Foundation. Actually, I'm not actually positive because there's two different branches. before we get out of here. First up, the Raspberry Pi Foundation. I'm not actually positive because there's two different branches, so I don't know if this is the foundation or not,

but Raspberry Pi folk have an official keyboard and mouse that is available for sale. And one of the selling pitches of the keyboard, besides its small size, is it's integrated with a USB hub. I mean, you're going to need one of those anyway, right? There's no way your laptop has anywhere enough ports. I mean, actually, this one's pretty good. Or Raspberry Pi. Yeah, I mean, whatever device you're plugging into. I was just immediately thinking that this is like, it looks so nice that I would probably be using it.

You could. Not just on my Pis, but on all my systems. Yeah, you could. It's just a cute little three-button scroll wheel optical device with Raspberry Pi logos for the mouse. It's a little USB cable, but it's white and red, which is a cute little color scheme. Yeah, it looks nice. And then just like a pretty standard 78-key matrix for the keyboard, which you probably would see on a laptop keyboard. on like a laptop keyboard.

What's interesting though is they've worked on some custom firmware so that hopefully with a little magic, it can auto detect what country the keyboard is configured for and just sort of work right. So you'll be able to get a keyboard that's manufactured in a simple, you know, easy way

that'll work wherever you are in the world. Yeah, it's kind of worth reading the Raspberry Pi blog post in the show notes because they talk about the care they went to to make it a good keyboard experience. And they're like, this turned out to be a little more challenging than we expected, turns out. The Raspberry Pi official keyboard and mouse

are available now from approved resellers. And I decided to do it. I got the keyboard. The mouse wasn't available from the place I wanted to buy from. So I got the keyboard. I ordered it this morning. Oh, I'm going to be curious to do it. I got the keyboard. The mouse wasn't available from the place I wanted to buy from. So I got the keyboard. I ordered it this morning. Oh, I'm going to be curious to try it. It wasn't bad. I think it was like maybe after shipping and tax

and everything, it was like $27. Yeah, that's not bad. And the mouse is something like around $20 US. Maybe I'll pick it up later, but I was thinking about doing a segment about like a cheap Amazon LCD HDMI screen and using this and just like the cheap, easy bare bones Raspberry Pi setup that makes it easy to just bang out Raspberry Pi projects like a segment around that.

And so I thought, you know, I was actually literally when I was thinking about that segment, I was like, I don't have a great keyboard. And then, I'm not even joking, I woke up and I saw this announcement. So I thought, maybe it was a sign I should try this. That or it's a good excuse to order a $20, $30 keyboard. So I'll let you know what I think about it. Also something that I thought was kind of neat is, I think this might have been sent into the show too, Make Pro Audio.

This is a DIY kit that you can get for the Raspberry Pi and you build modules around it, turn it into a professional audio. I'll just leave in the show notes. We don't need to go into it a lot. I think the thing that's really cool about it is it's a company that's well known for make your own audio gear. That's the MakePro

Audio company, that's what they're known for. And they have all kinds of modules they've already built for other things. So they're sort of known for creating these platforms to build very professional audio kits around. And they're doing a build-your-own-audio-gear platform for the Raspberry Pi. Right, so you already have a huge access to just a huge number of modules, and now they have the code needed to fit the Raspberry Pi into that existing platform.

Yeah, and it snaps in there, and then you've got a fancy Linux computer controlling all that kind of stuff. It's really cool. It's really neat. Well, go check it out. For those of you who care, go check it out. It'll be in the show notes. I already want like a little Raspberry Pi mixer for my home studio closet. I'm not going to lie. Yeah.

I know, right? I know. You and I geek out on this kind of stuff. Well, our last bit of community news for this week is from our own Mr. Martin Wimpress in the Mumble Room right now. And it is Ubuntu Mate 1804 Beta 1, so this is the LTS, for Raspberry Pi.

It's a beta pre-release, so it's hot off the presses right now. It's a great opportunity to give Martin your feedback. And Wimpy, you know what I'm most excited about is that hardware accelerated VLC. That's huge. You like that, do you? I like that a lot. I mean, I thought maybe

you would get there, but when we were kind of chatting off the show, my expectations had been set for this maybe not to happen in this release. And then I see the announcement, and you nailed it. So I'm curious how we went from, I don't think it's going to happen, to, oh, by the way, it's a feature now. We were rather fortunate in that as I was entering the final stages

of preparing these images, the Raspberry Pi Foundation dropped patches for current versions of VLC and some other software, which made it possible for me to enable this. So all down to the work that they're doing, yeah. Open source is amazing. Isn't it? That is so great. Jeez.

So not only VLC, but FFmpeg as well. So kind of the standout here is maybe now the web browser. Is that right? So I'm a fair way along in doing that. I actually have a build of Chromium that is fully accelerated for the Raspberry Pi, again, because literally this weekend, the Raspberry Pi Foundation dropped patches for current versions of Chromium. Now, the issue that I have is a rather complicated one in that the builders for arm

in launchpad are based on arm 64 and obviously the raspberry pi is a 32-bit arm and the cross build environment is leaking a bit of the 64-bit nature of the host underneath and i'm running into a few issues where it's picking up um oh this should be 64-bit when actually it's 32-bit so i'm just having to patch my way through that but the problem is um chromium takes about 26 hours to build on arm even in launchpad oh man so my my iteration cycle is currently around 11 hours to find out whether or not my fixes are working. So, you know, I had a failure condition at about four hours and a half and then at seven and a half hours. And now the current problem I'm fighting is about 11 and a half hours into the build.

So I'm slowly getting there, but I hope maybe not for release, but it'll certainly be a post-install option, an accelerated version of Chromium. I'd like to have it baked in, but we'll just have to see how it goes. Like I say, the iteration time is pretty long. That is brutal. So even more kudos to you for fighting through that. I love that FFmpeg and VLC at least got this, because if you think about it, this is the 1804 LTS, so this could be running on some Pis for a while. This could stick around for a while, so it's really nice to see that. I think the other thing that's really clever that you've

included in this release is Steam Link. How'd you get that idea? Well, yeah, so that was obviously made available for the Raspberry Pi shortly after Valve discontinued the hardware devices of Steam Link and it just occurred to me that you know that is a really good use case for a Raspberry Pi to create you know a head-end unit for your Steam streaming so yeah I put a bit of work in there this is actually the opposite kind of problem where I had to pour older libraries to the system and not have them tread on the newer versions that you know are in 1804 so there was a little bit of delicate you know refactoring of some packages to provide some old older versions of things that sit alongside current versions of things but yeah I'm pretty happy with that. Something I would like people to test.

I can't obviously fix Steam Link itself because that's all built from upstream binary blobs. I can't recompile it, but I'd certainly appreciate some feedback on how it's working. You know what strikes me about creating a distribution for the Raspberry Pi? Most distros have to be compatible

with almost all of the x86 machines in the market in theory, right? You just, you expect that if it's an x86 machine with standard hardware, a distro will run on it. But a distribution built for the Raspberry Pi, there's really only like four or five, maybe three devices you should expect it to run really well on that you have to test on.

I'm not trying to say it's easy. I'm saying it is an opportunity. If this was a full-time job and you had a team of developers, you could buy each one of them the entire range of Raspberry Pis. They could do all of their quality assurance testing on that. They could stamp it with their quality seal approval and say, yes, Mr. Wimpy, we have tested on every single Raspberry Pi. Like, it is a manageable, scopable problem.

Yeah. And that's interesting. I have all four of the Raspberry Pis that this will run on, you know, and it's not an expensive, you know, outlay to buy all four of them. It's, you know, under $100 to own all four of those computers. But unlike x86, everything about this is extremely bespoke yeah it's got its own bootloader it's got its own custom kernel it's got its own um bespoke way of enabling um bluetooth and wi-fi it's got its own considerations regarding gpio then you've got things like you

know python module support adding things like usb boot support frame buffer support and then we've got all that we've talked about with hardware acceleration and then you have people who have some expectation of what you do with the raspberry pi because of raspbian and the tools that exist in raspbian you know and if those tools are not available in a raspberry pi distro then that is alien because a lot of the time people using the raspberry pi don't think of themselves as linux users they think of themselves as you know raspberry pi users so you have to then you know make sure that those tools are available and ported because there are in fact quite you know

considerable differences between raspbian and ubuntu So there's a tool called RaspiConfig that a lot of people will be familiar with, and I've ported that from Raspbian to Ubuntu and catered for the differences that exist between the two platforms. Yeah, it is not as easy as I just made it sound, but it's a fascinating, it almost, Billy, doesn't it kind of remind you of like the Android versus iOS development paradigm? Yeah, exactly. Maybe a little bit lesser because on Android we're working with Java, but definitely.

Yeah, and Billy does a little Android development. And he says sometimes, you know, he feels a little jealous when you look over at the iOS side. There's less devices to target. Yeah, it's not just a big old mess. Imagine if the Raspberry Pi was much, much, much, much larger and it was maybe much more mainstream. Maybe that's where the bulk of the users were, so it wasn't so bespoke.

I don't know if that'd be a great future, but in a way it would be less fragmentation. I think it's a really great device to target. I mean, I think they announced just recently they've sold 25 million Raspberry Pi. Now, I don't know how many of those are the raspberry pi 2 3 3 plus and um 3a plus but i imagine it's probably the lion's share of those so it's a large audience to hit you know so it's a great way to do this and the one thing i've done this time of course

we've got 32-bit arm and 64-bit arm images, there's some caveats with the 64-bit version, but it does mean that now on an expensive device, if you want to create projects to target Ubuntu on ARM, you've now got a ubiquitous platform that you can now test your applications on 32-bit ARM and 64-bit ARM. They're good enough for that. The 64-bit systems are missing pretty much all of this hardware acceleration

that we've discussed because it's simply not available for 64-bit ARM. It's really, I mean... That's a caveat. It's amazing that all of this works as well as it does on such a different environment. I know. It does feel like it's moon computers.

It's a totally different system, and we're making contact, and they're building the spaceships to get us there right now. Well, I think it's looking like a solid release. It's in beta right now, so folks, go give it a try. And it's just one SD card flash away,

so why not? That's one of the things, I think, that makes the Raspberry Pi so approachable is you can get a fairly reasonably priced SD card, throw an image on there and try it and swap it out. Or use that fancy new USB boot. You know what it's like? It's like floppy disk. Back in the day, I would have my entire operating system on a floppy disk and I would have custom bespoke floppy disk for like different games with optimized environments and I would boot from a floppy drive

and now we're doing it with SD cards and Raspberry Pis and I just now made that connection. Like why is it taking me so long to realize that these are floppy disks

all over again? Now, where can we find a floppy disk reader for the Raspberry Pi? That's my next question. Yeah, well you know what?

It's got to be out there. There's a floppy disk that has an SD card, like it's like an SD card adapter like the floppy, that's got to be a thing. That has got to be out there. It's a floppy disk that has an SD card. It's like an SD card adapter, like the floppy. That's got to be a thing. That has got to be a thing. I'm not trying to get all the secrets out of System76, but I'm wondering if Emma would tell us, like, you know,

on a scale of 1 to 10, maybe. I don't know. I mean, just like if there's any kind of hint, maybe, that System76 has looked at these ARM devices and thought, you know, maybe there's something here we could ship. Maybe there's a product out of this. What can you reveal Emma? I really don't think that they have been looking into that at all.

So no, like all pink, a raspberry pie, you know, arm laptop or something like that. I mean, it's possible, but I haven't been part of any of the discussions.

I know the, the arm servers is our biggest performer with the ARM market. So you do sell an ARM product then, I guess, in the server space? Yeah. It's a bit of a beast too, if I recall. Yes, the IBEX. Yeah. Are you looking it up right now?

Yeah, Wes is going over there right now to go check it out. I'm going to go over to System96 real quick. Yeah. As you do, you know. I don't know what I'd ever use an ARM server for, but I wish I had a use for it. It just seems super cool. I guess a NAS. I could build a NAS out of an ARM server. Well, I'm

going to hold out. I see a future one day where there is a laser engraved System 76 aluminum laptop with an ARM. I don't know why I care, actually. I think x86 is doing great. Long live x86, I say. Hey, long live x86 is doing great. Long live x86, I say. Hey, long live x86.

All right, a little housekeeping. Some big things to talk about, including some awards and whatnot coming down the pipe. Award shows. Did you know that the open source community has award shows? I did not. Not only does it have award shows, apparently it's had them for years. I probably even talked about them on the show before. God, what is the matter with me?

Clearly, you place a very high value on awards. Yeah. I really, yeah, okay. Anyways, but you know, some people do. Some people do, and it's totally fine. Are we going to have a viewing party here at the studio? That's what you're saying? Oh, yeah. You know what? If they live stream it, we should. I'll tell you what, if they live stream it, that would actually be pretty cool. But we do have a couple of big things to talk about. Wes just got done with an Ansible stream it, that would actually be pretty cool. But we do have a couple of big things to talk about. Wes just got done with an Ansible study group, just did one before the show. Oh, yeah, yeah. Demos and all.

And he and Elle went through the basics of Ansible, described a little about use cases of Ansible, and then you even gave a demonstration, a nice one, too. A nice, clear, easy-to-understand demonstration, a demo, if you will. And it's done now. The live stream has concluded, but we will release that video in the future on the YouTube channel. Something to look forward to.

Yep. We'll have probably a link next week. In the meantime, if you missed it and you just prefer to take it, Elle has now released the Ansible Quick Start course for free, along with a bunch of other things, including using Ansible for configuration management and deployments, Google App Engine Deep Dive, DevSecOps Essentials. See, what's so dangerous here is

you start with the Ansible Quick Start and you get a whole bunch of servers set up, and then now since you can configure them with Ansible, you take all the rest of that free content and suddenly you're hooked on the cloud. I know, that's true, yeah. PowerShell Core for Linux Admins. I know, that's true, yeah. PowerShell Core for Linux admins. You know, I got to admit,

I'm just a tiny bit curious. Like, I kind of am curious. So we have this community edition thing and that's what Al and I are in charge of. We release these courses every month. There's a whole bunch that are always free and then there's new ones that get released every month.

Some get turned back to paid only. But you can sign up for a community account and go try them out. It's pretty good. I'll have a signed-in link in the show notes for the Ansible one, so if you're logged into the platform and want to go right to the course, I'll have a link in the show notes. That'd be pretty cool. So I just wanted to let you know about that.

So we'll have that posted up on the YouTube channel, Wes's study group, soon. It'll be a little bit because we've got to encode it, and that takes a long time. It's a little post-processing. We got rid of all of our video computers, and then all of a sudden we decided to do a video thing. So we've got to go back into the garage.

No, I'm kidding, I'm kidding. It'll just take a little bit. Now, we also do have these open source awards, the UK Open Source Awards 2019 in Edinburgh, Open Source Awards, the UK Open Source Awards 2019 in Edinburgh June 12th 2019, 70 days, 8 hours 54 minutes and 15 seconds

from when we record the awards event will be at a university and you can go there, there's going to be I think a little drink section, a little show thing, it's kind of like a show and an event kind of thing, so there'll be booths and all that

kind of stuff, I think, and as well as awards. And the nominations are still open as well in five categories for individuals and organizations who have, in the last year, shown exceptional initiative with free software and open source software. Obviously, Wes Payne, you should get the

individual award, company award, and public sector award. Yeah, I'm all right with that. I assume that comes with a free ticket to the event, which sounds lovely. But really, there are so many projects that can fit that bill, so go out and nominate. Yeah, okay, maybe not. But I feel like, you know, maybe somebody could go and nominate Wes,

because the nominations are still open, right? So somebody should nominate him for something. You're saying I haven't done that for myself already? You know, we do have some stuff up on that GitHub repo. We do have some stuff. You know, Cheese, while we're still in housekeeping, you know what we ought to just tease a little bit is the big look change that's coming within days.

Within days as we record this. Yeah, we're nearing the first rollout phase. Which means we're right now in crunch time. Cheese right now is in crunch mode. I'm working on stuff as we're in the show right now, actually. We've just strapped the microphone to his face. He's still designing with his hands. You're podcasting and designing at the same time, huh?

Well, there's so many. I wish we had an account. We don't have time to count, but there's maybe 1,000 art assets in total. Way more than I expected. There's a ton of stuff out there. Plus, we're also getting the swag. Most of those orders have all been placed,

so I think people are going to be happy with that. Linux Fest Northwest swag. I've had some people asking me if I show up or do I not get the stickers? Can I not get them? We do have some stuff in the works, some things in the works to make those items available to other individuals.

Initially, they're just going to be at LinuxFest. Sort of a trial run here. Down the road. Down the road. It should be available to more people. That's going to be really cool. Also, along with all of that, we're probably doing too much. But it's so exciting to have our own designer now. It's amazing.

She's just the best. We kind of went crazy. We also are working on revamping the website a little bit. Cleaning it up, trying to make it a little faster, updating the look. There will be a couple of features that get removed initially, and then we'll work them back in in a new way. Stuff like that is coming soon to the website. It's definitely evolving.

I think people are going to find it's a lot easier to identify the shows that they like and they want to listen to. I think it's going to be, you know, people have been asking for it. We've been listening. I'm here to try and help this this all happen we got a new look the individual show pages will get updated as well with the new look i mean it's just it's a full spectrum new look we may start uh just to get it right we might like roll it out in phases so you might see it like update in a few places and then there'll still be some art assets out there that we haven't gotten to and then a couple of days we'll get to those and a couple

of days after that we'll get to the rest but there's so many um like just today i was like logging into twitch i'm like oh yeah right twitch we got twitch we got a twitch art thing we got to do and ustream's got a background for when we're not streaming that has to get updated like there's so much stuff because we're doing the network logo. We're doing everything. It's very, very exciting. Of course, nobody likes a logo change. Give it a little time. Let it sink in. I think everyone

will find something to like. I hope so. I think it'll make a lot of sense to people, the brand change for JB. One of the best benefits of this job is that after I'm done grinding away on this stuff all day, I can actually go take Linux Academy courses. That's true. Yeah, you know, one of the best benefits of this job is that after I'm done grinding away on this stuff all day, I can actually go take Linux Academy courses. That's true. Yeah, you did that last night, you maniac.

Another free one out there that I really recommend people check out is the one that Kenny did. And I think you mentioned it on a previous show, the Linux Operating System Fundamentals. I'm just starting from the very bottom uh basically all of the community courses and i'm going to work through that and and hopefully be able to give you guys some feedback on those as well right on well come say hi to cheese wes and myself as well as wimpy and many others at linux fest and get the new swag items there and then we'll uh we'll get them up in other ways as well but that's the that's the housekeeping. Thank you, everybody. Now, let's get to some feedback.

This is something that you hear thrown around in the community a lot, so I thought it'd be fun to read on the show. Sigflup writes in, he says, I find it's a common misconception that compiling programs makes them run faster than if you were to install the same program via a binary package. I personally don't think this is the case because compiling something is often just a mirror of how the package was built

and you produce the same binary for more or less. Do you think I'm wrong? Maybe try to change my mind. I'd like to hear a discussion about compiling things from source and its advantages. Thanks, Sigflut. So what do you say, Wes? Let's take a first stab at this.

Is there a difference between pulling down a software package and building it yourself versus the binary package? I think it all depends on what options you use. So perhaps depending on your system, there may be some tunables that they can't apply in the general case, and you could have a little bit of a speed up. And in many cases, there are not. But I think actually, yes, Mr. Martin Wimpress,

who's been stuck over in linked library hell, probably has a lot of thoughts about this. Yeah, I think you're dead right. That, you know, the performance optimizations that, you know, build from source distributions tout are usually because each of the architectures that you run those distros on have tuned compiler flags specifically for that

architecture and you can get you know for clever use of each of the options you can get a small gain in performance but then by you know aggregation of small margins you can get you know a measurable improvement in performance and this is what something like clear linux does improvement in performance and this is what something like clear linux does it's a single architecture focused distro it's x86 64-bit so consequently they can tune that specifically to intel cpus and that's why it runs so well on that architecture but when you're making a multi-architecture distro then you have to be more conservative in the compiler directives that you set as defaults to make sure that stuff works reliably across all of your architectures. And the gains are modest, but certainly the rices of Gen 2 past, it's not quite like it used to be anymore. Hey, that's me.

Yeah, when I ran Gen 2, it was really to get access to things that weren't quite mainstream yet, that now every distro does, and it's not a big deal, and there's no reason that I needed to build it myself. And it's not a big deal. And there's no reason that I needed to build it myself. But traditionally for me, I've always built software because I either needed a version that was newer than the distro shipped. Yeah, that's a good reason. Or it's the only way it's distributed.

Or in the past, and when I say past, it's like the far distant past now. Because I needed something that maybe distributions weren't building by default. Maybe they weren't optimizing for that particular architecture. Because, you know, back then people were still shipping packages that were compatible with i386. Like, it was a long time ago. And it just seemed like problem solved as time moved on. That's been my biggest go-to. to. I think there's really no performance advantage inherently to just grabbing the same exact packages in your repo versus grabbing something off of GitHub and building it. If it's the same

version, same features, you're probably, even if you could come up with a few ways to optimize it in the build, they're probably imperceptible in the performance of the application. So there's really no point from a performance standpoint. Just save yourself the time and don't build it if you don't need to.

There's only one other advantage I'd say is sometimes there's some, for whatever reason, there's some types of software that I like to have in like a bin subfolder in my home directory. And I like to have just certain programs that are just there. They're not, my package manager isn't aware of them.

It's not trying to update them. You know, it's not reliant on libraries throughout my system. It's like it lives in this folder. And sometimes that might be a reason why I might build something from source if it's simple enough. Right, yeah, you might want to have a whole separate little thing,

maybe not necessarily all the way into a container or cheroot, but just a separate little dependency chain. How often do you build things from source manually these days? So not like from the AUR or anything like that where you actually think, right, here's the thing, I need to go and download a tarball and compile it from source. The last thing I tried to build from source was butt.

B-U-T-T, the broadcast utility that does ice cast streams. Yeah, I think audio stuff ends up making us both do it a little bit more than otherwise. Didn't you just build something recently too? Yeah, totally. Yeah, so there's some weird esoteric tools here and there. But otherwise, I would say since the invention of snaps.

Yes, right. Yeah, I mean, not more distribution. It's just gotten a lot better. Snaps, PPAs, people just shipping debbs. Yeah, what about you, Whippy? I mean, it must be different for you. No, no um there is uh so i've just upgraded from 1810 to 1904 today so the first thing i noticed is i only had one ppa on my system so you know there was none of that nonsense to be dealing with post install and also the ppa in question had a disco release and it automatically transitioned i'm sorry did you say sorry i had to yeah um but i had one i have got one bit of software that i compile from

source on my system right now that i had to rebuild post upgrade and that is green with envy yep which which i love by the way that is that is one of my favorite bits of software right now. It is brilliant. Remind me what Green with Envy does. For Windows users, this is MSI Afterburner for NVIDIA GPUs for Linux. It's so good. And I've used this to really aggressively overclock my 2080 Ti. It's amazing. And that's distributed as source right now? Well, it's available as a flat pack. They're working on a snap and it's available as source. But because I'd, you might remember some months ago,

I talked about writing scripts to do all of this overclocking stuff. I wanted to poke around inside it. So I've been having a play and there's a feature I'm working on that I want to send back upstream. Okay, Brent, so when's the last time you, or if ever, have had to build software on Linux? Would that be interesting? Have you ever been forced to try to give it a go before? Oh, great question. So I actually used to do it a lot. Well, I say a lot compared to some of you. I used to do it sometimes lot. Well, I say a lot compared to some of you. I used to do it sometimes.

Kind of the same reason as you, Chris. Whenever I'd find some tool that I really felt that I wanted to try but wasn't available, either in the repo or something like that, I thought, well, this is worth the effort of jumping down a tunnel or a rabbit hole that I don't completely understand to try to make it go. And I'd say 50% of understand to try to make it go. And I'd say 50% of the time I'd get it going and 50% of the time I wouldn't, but it'd be many, many hours.

You never know how that dot configure is going to go. Yeah, these days I just sort of cringe whenever I see I need to do that. I just look away and try something else. But I would say that for the tools that I really need most of the time, I never have to do it anymore or even think about it. So we've come

a long way. Yeah. I do the math. Like, I'll be honest, I've probably just been like, I've noped out now a few times. I've probably noped out a couple of times in the last year, but it's only been a couple of times. I do think sometimes, at least like audio stuff stands out, but anytime you have that thing, it's kind of

like a superpower. You know, you perform the incantation and then at the end, the software works. Well, maybe. I think that's why people like to build their own kernels. That's like the ultimate. Yeah, you get to customize options. It can be a good way to start learning with stuff, right? There is that.

Yeah, but that's a pretty niche case. You know, that's not, no, it's probably not, not really. I mean, everybody could probably, if you don't do it very often, why not try it? Why not see it? It's probably something we could do more of. We have such ubiquitous access to all of these great, from complex to simple open source applications,

and that's such an easy way to learn that we kind of just avoid now. Even if you don't intend to use that as a way to manage software, why not give it a go just to learn it? And our operating systems and the culture behind it make it easy to do so, right? You don't have to go set up some crazy Windows development environment with proprietary compilers. It's all there and just a regular package away.

I'll tell you, I went from kind of like, you know, I understood the concept of kernel modules and proprietary modules that link to the kernel. Like, I got that. But, you know, it's nothing like actually building your own kernel where you're choosing which modules are built versus not. And, like, truly understanding, like, really how this works and how you can have stuff just built into the kernel because it's free software. So you can just build it right in. just built into the kernel because it's free software so you can just build it right in. Going through the process of building your own kernel

and getting it to boot and creating an image out of it and all of that and updating Grub, it's a very educational process and there's so much documentation on how to do it that it's probably worth doing in a controlled environment, maybe like in a VM. Or a spare system

you don't need. Yeah, that's the way to do it. I used to do it on my main computer and I'm just crazy. It's like car maintenance but where you can afford a practice car. Yeah, that's probably true.

It probably is probably the same advice you could give to maintaining your own vehicle. All right, well, you know what? I think that's probably

a good way to leave it right there. If you want to get your question into the show and have us discuss it, go to linuxunplugged.com slash contact and send it in. Let us know. We'd love to hear it and kick it around

on a future episode of the Unplugged program. That's linuxunplugged.com slash contact. You'll also find links to everything we talked about today at slash 295 as well as in your podcast catcher, and you can subscribe to future episodes of the Unplugged program on that page as well.

We've got links all over that site. Everywhere. You know, I've got to say, I really enjoyed the last TechSnap. Aw, thank you. You guys, killing it. Killing it. Ways that people could attack you in the supply chain. Things I never even thought about. Basically everywhere.

Yeah, it was a little hard to sleep that night. Yeah, don't listen if you're overly scared, but do go check it out, techsnap.systems. Also, the Ubuntu podcast, rumor has it, may be returning soon. Oh boy. I don't know. I don't know. I would just make sure I'm subscribed to the Ubuntu podcast right now. That's what I'm saying. Maybe constantly refreshing. I'd make sure

you do that this week. Yeah. Yeah. I'd just say that's a good idea. Just make sure you're subscribed to Ubuntu podcast. Thank you for joining us Wimpy. Now, thank you all for making it. We appreciate it very much. I encourage you to join us live over at jblive.tv. We do it on a Tuesday. Go to jubilabroadcasting.com slash calendar for your time zone.

See you then. next tuesday Next Tuesday! I mean, I don't want to disappoint. All right, now we've got to pick our title, jbtitles.com. Thank you, guys. Great show. Really enjoyed chatting with everybody. Look at that huge mumble room still there. Amazing.

You guys even stuck around for the whole show. And the post-show. Yeah. So what should we call this thing jbtitles.com everybody go boat in the meantime um mr bacon has a little community news to share with the class oh yeah yeah um so let me jump back here uh mr lonnie webb uh one of our community members is uh has going right now called Killing Wryth 1. And I can drop that link in the IRC here shortly, or we'll also put it in the show notes.

Right now, basically, he's doing a comic. His goal is $750. He's at 59. He has 28 days to go. So if we could go over there and give him some love, I'm sure he'd appreciate it. And his stuff is actually really cool, man. I like it.

It's the little image that he has, the video that he has there. I don't know. It reminds me of just this kind of Orwellian cyber, pre-cyberpunk kind of look. It's really cool. I think everyone should go take a look at it. Is he using free software to make this comic?

He does not say, but I'm sure, you know, I could reach out to him and ask him. We're going to say he does. Let's just, let's just pretend. Well,

I know that a lot of it is done, you know, it appears to be that he sketches a lot of it out first. So my guess is that he goes back in and inks it later. But, I mean, I just love his style. I think it's super cool. I think a lot of people would enjoy it themselves. So take a look and give him a few bucks if you can.

Boom. Nice. Hopefully it works out. It's nice to see community members working on stuff like that. All right. We've got to pick this title, and then we've got to go. Yeah. You thought your work was done when you suggested a title? Oh, no, my friends.

No, no, no. This is the hardest part because there are some really quality titles today. This is interesting to see YouTube getting a lot of crap today. A piece over at Bloomberg, YouTube executives ignored internal warnings from staff. Concerns that were brought up about toxic videos were ignored. And proposals to change recommendations were dismissed in favor of all-out chasing engagement. And they have documents as recently as 2017, it looks like, showing briefings to upper management in the Alphabet Corporation about how they need to pursue the engagement strategy.

It's just not looking good for YouTube these days, is it? No. Boy. Meanwhile, today is the day to uninstall the Google Plus app from your phone because we are just hours away from Google Plus being taken offline. Right. I had a search result that had a Google Plus answer, and I clicked it, and I got a page that said Google Plus is going away.

So I was like, well, remove it from the search result, Google. You're the one that gave me the link. Look, it's hard when you cancel that many products. I know. It really is kind of hitting home right now because inbox is going away, which was actually kind of nice. I'll say, like, there was, like, one bill I didn't know about that Inbox gave me a push notification.

It was like, hey, Holmes, you better pay this thing. I was like, oh, my God. The robots are good at reading your email. Yeah. So Inbox just got shut down. Google Plus just got shut down. Meanwhile, we have multiple ways to listen to music. You have YouTube music and Google Play music.

You've got multiple ways to do text chatting. You've got multiple ways to do video calling. Oh, Hangouts is ending later this year, too. Yeah, it's weird what remains, and it's weird what they do decide to kill. And it is becoming part of their brand. They launch fast, and they kill fast. And so you don't know if Stadia is going to go away fast and so you don't know like if stadia

is going to go away and now you you don't know about any of this kind of stuff so well that's just dumb fresh meat entering the abattoir right now isn't it what stadia yeah you think so you think it's uh never gonna really stick no i'm being facetious i think i think this is one of those moon shoots that will probably pay off i i think when we looked back at um how crazy um the chrome uh chromium uh chromebooks were you know when they started that initiative because it really everything wasn't aligned for that to be successful at the time and i think that stadia certainly the technologies are not aligned right now for that to be a success, but it will take somebody like Google to make it successful. And I've got mixed feelings about how successful I want Stadia to be.

Yeah. And I would imagine from a game developer standpoint, I would really, I'd have to, I'd have to really think about how much development time do I want to spend to customize my game for a platform that is controlled by Google versus open platforms? I guess they do that for consoles. So I suppose you could look at it. It's almost like another console, but in this case it could be... I mean, you already have a whole bunch

of proprietary platforms you target. It would be interesting to see what you update. At least when you make a console, it's out there for a while, right? This they can turn off any time. But the flip side is when the new console comes out, you can't play the old games. But with streaming, in theory, you could play everything from DOS games

to the most recent crazy games. Right. And what I'm looking forward to seeing is those game studios in the past that have been absolutely ardent. We're never doingux to see what their position on stadia is going to be because whether or not you get a desktop linux version of a game because you've targeted it at stadia that's a separate point but you have to target linux in order for it to run within the Stadia infrastructure. So it would be fascinating to see whether anyone,

Epic, for example, changes their opinions about porting to Linux. And in a way, it only has to be somewhat successful, perhaps, for that to happen. Because it's a whole retooling that has to happen on the dev shops part. So if Stadia is even semi-successful, you may see shops begin to retool around Linux, which could have long-term. Well, then they announced some pretty good, like, Real and Unity both had, or are getting support. And you could see others falling in line, which could have good long-term ramifications.

It's just so, it could go either way. It's a fascinating story in that regard.