LINUX Unplugged - 372: Distro Triforce

Episode Date: September 23, 2020

What would it really take to get you to switch Linux distributions? We debate the practical reasons more and more people are sticking with the big three. Plus Carl from System76 stops by to surprise us with some firmware news. Chapters: 0:00 Pre-Show 2:22 Intro 2:36 SPONSOR: A Cloud Guru 4:24 USB Booting the Pi 4 10:10 System76 Open Firmware Update 23:14 SPONSOR: Linode 25:28 OpenPOWER Summit 2020 29:23 EndeavourOS ARM 30:14 Housekeeping 30:53 SPONSOR: Unplugged Core Contributors 32:59 It's Really Just a Three Distro World 46:37 Feedback: systemd Skepticism 50:50 Feedback: EmacsConf2020 51:40 Picks 52:12 Pick: Cloud Hypervisor 53:51 Pick: SongRec 54:45 Pick: tmpmail 55:55 Pick: MyPaas 57:16 Outro 59:11 Post-Show Special Guests: Brent Gervais, Carl Richell, and Neal Gompa.

How about this for a little bit of troubleshooting? A secondhand television wiped out the broadband for an entire village. And this went on for months and even impacted neighboring communities at time. And poor broadband was just a way of life starting at 7 a.m. despite repeat visits by engineers to try to find and fix the fault. They replaced equipment, large sections of cable even, but the problems remained. Local open reach engineer Michael Jones was determined to find the Abrahosian broadband smoking gun. And I should note there, I don't know if I'm saying that town right,

because my Welsh is a little rusty. That's strange. Well, they say on their blog post about this, as a last resort, we decided to bring in a crack squad of engineers from the chief engineer's office who were based in other parts of the UK to investigate. Accommodation was understandably hard to find due to COVID-19 lockdown. But we eventually found a field

where we camped and made the 55 mile journey back to the town the next morning. They exhausted all avenues. So they wanted to take a final test to see if maybe a phenomenon known as shine, single high-level impulse noise, was causing this interference.

I love this, right? It's sort of a moonshot, like, okay, we've tried everything we know. What about this weird random thing? Well, by using a device called a spectrum analyzer, we walked up and down the village in the torrential rain at 6 a.m. Uphill both ways, I'm sure, to see if we could find an electrical noise to support our theory.

And at 7 a.m., like clockwork, it happened. Our device picked up a large burst of electrical interference in the village. The source of the electrical noise was traced to a property in the village, The source of the electrical noise was traced to a property in the village, and it turned out that at 7 a.m. every morning, the occupant would switch on their old TV, which would in turn knock out broadband for the entire village. As you can imagine, when we pointed this out to the resident, they were mortified that their old second-hand TV

was the cause of an entire village's broadband problems, and they immediately agreed to switch it off and not use it again. And since then, there has been no reported issues with the network. Hello, friends, and welcome into your weekly Linux talk show. My name is Chris. My name is Wes. Well, hello, Wes Payne. You are a gentleman that looks good in uniform.

It's looking good on you today. I thought you might like this cap. I do. You know me too well. Today's episode is brought to you by a cloud guru, the leader in hands-on learning. The only way to learn a new skill is by doing. That's why ACG provides hands-on labs in cloud Linux servers and much more. Get your hands cloudy at acloudguru.com. Well, here we are with episode

372, and we are going to ask a question that has been evolving in the back of our minds, I think, for a couple of years. We ask it with all sincerity. is there really just no room for new boutique distributions? Is it essentially a three distribution world with Ubuntu derivatives, Arch, Fedora, Red Hat? There's some Debian flavors. It's a small world of Linux distributions these days. And Wes and I recently tried out some really great distributions and wonder why more people aren't using them. And so we'll talk about that. But of course, we have a bunch of community news to get into. And we have a special guest joining us today, the CEO of System76, an all-around great guy. Carl is here. Hello, Carl. Welcome back to the Unplugged program. It's been a while.

Hey, Chris. Thanks for having me. Carl, we're going to get into a news story from System76 here in a little bit. So feel free to chime in on everything. But of course, we'll go to the source for that one. But I wanted to start with some news that I'm super ecstatic about. So to help us analyze all of the community news this week, we have to bring in our team of detectives, our crew of captains. Why, yes, it is our virtual lug. Hello, Mumble Room. Hello. Hello.

Bienvenido. Absolutely. Good to see. Good to see see we got 21 in there we're shy a little bit today of a few folks i guess you know they didn't know you're wearing your uniform today or maybe you scared them off wes i look a little too authoritarian today i think but don't worry we're friendly here at linux unplugged as long as you comply with our orders that's right that thank you gotta make that disclaimer all right, some news I'm actually super excited about. I haven't got to mess with it yet. In fact, I'm not even sure if it's fully shipping anywhere yet. I just saw this on GitHub and wanted to talk about it immediately. It appears that if not right now, very soon,

the Raspberry Pi 4 will officially be able to directly boot from USB. This sounds pretty cool, because this is kind of like the big performance limitation in my Pi setup. Yeah, it was. You know, I think this change is basically that the support had been in beta, and now they're finally saying, yeah, all right. Everyone can use this now. You don't have to jump through even more hoops. Although I still think updating the firmware on a Raspberry Pi is a little non-standard,

and so that that part can be a little confusing. You've got to do like a full upgrade and then use the firmware on a Raspberry Pi is a little non-standard. And so that part can be a little confusing. You've got to do like a full upgrade and then use the firmware specific commands that OSs like Raspbian have to make sure that you've got the right firmware and the right configuration so that you can actually use these handy new features. Tricky, tricky firmware. I think we'll talk more about that in a moment. But something else I noticed is it, and I don't know if you saw this, more about that in a moment. But something else I noticed is it, and I don't know if you saw this, but it looks like they're playing around with treating the USB disks or block devices more like SCSI devices, which could also improve performance. Have you seen this?

Yeah, it seems like people are taking it seriously that you might actually use these things, well, I guess like you do, Chris, and host your whole home infrastructure. And if a whole bunch of the audience is running from USB, it makes sense to maybe invest in whatever technology you can find to make that work a little more efficiently. It's not just, you know, a drive there that's going to have big files from your NAS that you never access. It might be your whole OS. Wes, anyone can have a whole home off-grid internet hot cache in their house as long

as they're willing to have like four Raspberry Pis. That's all it takes, four. Is it really only four? Actually, at the moment, it's only three. Once I got on the ButterFS sauce and started combining my volumes and reloaded everything onto LTS Ubuntu 2004, I'm like rocking with less Pis than I ever had to. It's kind of ridiculous, actually, how great these Raspberry Pis are.

And now I can say I have driven them into the summer of Texas and back where the booth got up to 104 degrees where these things were running and they kept on delivering. There was a couple of times where thermal throttling kicked in and the Plex like stream cut out or something like that. But I mean, really very impressed they're still running right now and i have like no reason to replace them like i fantasized about an x86 system in there just one x86 box with with proxmox that runs everything in vms and then you know just keep it super easy to do snapshots and swap it to a new box but the reality is is, is not only have they been problem free, but their low power draw is great for somebody who's trying to run on solar.

So it's kind of just been a win win. I think they're really cool. And so I'm glad to see the community, you know, picking up around them and and making them even even more serious boxes in a way. Now it just has all those little architecture problems to work out. Yeah. I mean, it's still probably not an ideal home server, but with faster disk assets coming online, maybe support for trim, things like that, it's getting closer. I was talking to Alex in the old self-hosted podcast about a future where I netboot these

things and maybe one Pi has all the storage and the rest all just netboot. And in that netboot these things. And maybe one Pi has all the storage and the rest all just netboot. And in that netboot environment, it could just be one cable because you can get a PoE hat for Raspberry Pis and you could power them with the Ethernet cable you're using to supply them networking. And so I could run all these Pis

with just a single clean cable that goes back to the switch. Oh, that sounds perfect. And the Raspberry Pi sounds like a great candidate for being a little net booting server, right? I mean, maybe. We'll see. I'll see how far I take it. But the possibilities seem like they're pretty appealing with the Raspberry Pi platform because it's low power, because you could run it off PoE, and now you can net boot or USB boot. For the kind of setup I'm using,

it just sort of seems to be a perfect little device. I don't know if I'd recommend it for everybody. And such a low cost investment, right? I mean, you don't have to go all the way to configuring a smaller x86 system or something like a NUC and investing in all that where, you know, you're thinking a lot about, oh, I got to buy extra hardware. What memory and SSD am I putting in here? You're just assembling these pies, you know, as you've got a few extra bucks here and there because the price point's so low. Pretty low risk. And in fact, the last thing I'll say about it is because it's not going to break the bank, even to buy an eight gigabyte one, when you're comparing it to the cost of a traditional x86 computer. And if you just need,

like in my case, a test one, I have, I know it sounds silly, but I basically have a bench pie that I just use for testing things before I roll it out on my home setup. And it's just sort of one of the things you can do when the thing just costs 40 bucks. Yes. Right. And I mean, in your, you know, your old life as a professional admin, you would have spare machines, you have backups, you might have, you know, hot standbys. That's also something that's a little more accessible in the home lab when your base is a Raspberry Pi and not a big old x86 server. Yep. And actually, to that point, I carry a spare Pi in the RV fully ready to go in the case. All I have to do is just pop in an SD card or a USB drive and it's functional again. And the idea

was that when I went down to Texas, if I burned out a pie, because it seemed very possible. Literally burned out. Yeah. I would pull that pie, pull its storage, and then pop in the new pie, which are literally Velcroed to the inside of this booth. So it's literally you pull it off and you slide in the new one to the Velcro and just reconnect all of the accessories. And it just picks up and takes right over. And because it's a $40 device, for me, it's worth it's like an insurance policy to just have one in the cupboard ready to go. So there's advantages to it. But

there is those tricky things like the firmware. Yes. And that's why it's the perfect opportunity to chat with Carl Rochelle from System76, who, I got to say, has really delivered on a promise that didn't seem didn't seem all that deliverable. How often do people say, buy this product from us, and it's running this type of firmware today, but we promise one day we'll make it possible for you to replace that firmware. That's just often one of those promises that kind of gets missed. And I think System76 has delivered here because if I'm not mistaken, not only has the open firmware project at System76 gotten continued development, which is awesome, but it looks like it's been widened

to previous machines that have already shipped. Is that right, Carl? Well, you're hitting on, I think, what is the most notable part of this work. I mean, first, it's quite an achievement to get open firmware on machines on day zero, meaning when they're available from Intel or AMD, we have open firmware. Well, AMD coming soon. But Intel, we have available on day zero when the chipsets are available

and when the CPUs are available. And that was quite a lift. But once you start building the infrastructure and the knowledge and the know-how to do it, things start moving faster. And so being who we are, we like to try to push it further to more things that will further liberate our customers and their computers. And so we went into the OpenEC, which to me is the most exciting part about all the

firmware because the OpenEC is what controls the things that you feel about a machine, like the fan curves, the hotkeys, and the interaction between the... And we're talking about the embedded controller here, right? Right, the embedded controller. Yeah, the interaction between the computer, the hardware, and the operating system, those things that you really see as a user, that happens at the EC. So with the Lemur, when that came out,

that was our first model that had an open source EC that we wrote all the firmware for. As you alluded to, even though that's super exciting, something that we were able to do that I don't think most companies would do is invest in bringing this technology to existing customers for products that we're already shipping.

And we're working on bringing it to customers whose products we aren't shipping any longer. And we're working on bringing it to customers whose products we aren't shipping any longer. So you're going to get those advantages from that work that we've done. And we really are starting at a chipset level. So we can say the chipsets in the current Gazelle and Adder and Serval were the same chipsets as the previous models, although there were some changes. But that means the bulk of the work was done. So we can port to those previous models, run through the QC team, and ship our customers liberated firmware. Well, congratulations.

So I have a question about that. But just an aside, Carl, has Intel's protracted development cycle enabled this in any way? No, not really. What enabled it was Intel reaching out to us and asking us to partner on making open source firmware. Oh. Right, yeah. Okay, wait a minute. Hold on, hold on a second. That's not something I understood about this before. Intel is aware of what you are doing and copacetic with it? Oh, not only that, we have a Slack channel where we chat with their firmware engineers. How is this a thing? Yeah, well, that's what got it all rolling.

Wow, that's amazing. Yeah, they reached out to us and said, hey, we want to do open firmware. And then their team came out. They had, it was kind of wild. Their firmware team came out and They had, it was quite, it was kind of wild. Their firmware team came out and they had this long presentation that after eight hours, they're like, okay, let's come back tomorrow. We have some more to show you. So, so we, so we cut for the day. We come back the next morning.

They have like a few more slides to go through and they were back chatting and they're like, they say, Hey, well, so you guys partnered with us on firmware. And we essentially said, you, you had us at open source. That was enough. You had us about 12 hours ago. Well, that's, that's really neat. I didn't quite realize that.

Yeah. We have access to the documentation that's necessary to enable hardware, things like Thunderbolt that we just couldn't do before that now we're able to do and offer on products with open firmware. And now the same thing is happening with AMD. We're working with Google and AMD on open source firmware for AMD platforms. How is that developing? Is that also going pretty smoothly? It's going pretty well. We were working, well, we've just been working through legal for a moment, but I think we're past the... Sure, I bet there's a lot of that. Yeah, there's a bit of that.

But I think we're past the legal as of this week. And that means, it didn't slow us down too much, but it does mean certain, just want to be on good sound footing when you're doing any kind of open source work. Sure, absolutely. Yeah, we'll be able to move more quickly now that that's past us and onwards and upwards with open firmware. Well, so here was my actual original question before that aside is, practically speaking, as an end user, what is the benefit for me? I mean, I like the idea of the world having a better understanding of what makes my computer run, but what is an actual

practical benefit that I get? Right, because this is like a big investment for you guys, right? You're putting engineering hours, investing and making all these changes. What does that mean to a customer? Right, well, the first thing is that we can very rapidly respond to our customers in ways that we couldn't before. If there's a problem or an enhancement that we can do through the firmware, something simple like, for example, when you close the Lemur lid,

a customer reported to us, it's a little scary to do, but if you put enough pressure on the lid, you can touch one of the keys underneath on the keyboard and turn on the machine. So now we can very quickly work on a fix and roll that out to our customers. And that keyboard, when the lid is shut, will no longer activate or resume from suspend. So our response, our ability to respond is a lot faster. And the really cool stuff you're going to start seeing coming over the next few months,

that's things like battery thresholds that are built in to the embedded controller, can be controlled in the embedded controller in the operating system. So in Pop! OS, you can do things to extend your battery life or the life of your battery the life of your battery is that so are you implying like charge control settings charge control right yeah battery thresholds and charge control so oh that's great yeah in gnome settings you'll see options for max you'll charge it all the way up or charge it 80% or maximum battery longevity, 60%. Oh, that's fantastic. It's, you know, again, those are things

that you can't do without control of the firmware. And then one that I'm really excited about, I like altering my keyboards. I like to change the key map around to things that I'm more efficient with. So I like to do that. Are we talking Dvorak here or something else? No, mostly I like to move around the things down by the spacebar.

Oh, sure. Yeah. See, I like to use my thumbs when I can. And having alt next to my spacebar is basically useless. And then don't do anything with it. So with this, with open source firmware and the embedded controller, we're working on an application that will ship as a utility with our products that you can move the key maps around.

You can create function layers. You can alter. Is this firmware level changes? So it would be persistent or is this? Right, right. And it's persistent. So when you save them, it saves it to the EC and those changes are persistent. All right.

Well, I mean, I bet you have a whole other list, but you've got me convinced. So my question is, what is the risk do you feel like as a company for reflashing these firmwares on machines that are in production? They're out in the wild doing who knows what connected to who knows what. And you're updating a pretty core part. That must have been at least some point of heartburn and consternation for a bit to figure that out. Well, that is challenging, and it hasn't gone without some hiccups. Usually what we like to do is roll it out into the factory first. So that means the factory is going to see a lot of different configurations that

might not be in the wild or that are in the wild, but we're on proprietary firmware. And through that process, we can catch a lot of problems that might happen. But yeah, it is risky. And so we also enabled rolling back so that if there's something that we haven't caught, a customer can roll back to a proprietary firmware. We can examine what might be the cause and get it fixed. Oh, okay. So there is an escape hatch. So that's always nice. That's the best you can really go for when you're doing these kinds of things, really.

Wow. Well, this is pretty exciting. So what customers starting now can get this that didn't have this before? It looks like it's the Adder workstation and the Gazelle. Is that right? The Adder and the Gazelle. We're also looking at the previous generation Adder, previous generation Oryx. And I think that's going to probably fill out the product line for where our starting point is. And from there on out, we will be working with open firmware. Well, I guess my last obligatory question is, do I have to be on Pop! OS to get these new firmwares? Or is there a means for people outside of Pop! OS to get it too? No, you can get it with any Linux distribution or Windows.

You just use a USB stick and you boot to the USB stick and run it. It's like a mini operating system, really, that will update the firmware for you. Could you do it from within Pop! OS, though? You can do it from within Pop! OS. And you can do it within, I think, any Ubuntu base would be essentially the same.

The features that we're working on to improve the operating system experience or the hardware experience through the operating system with this firmware are things that we're going to have. Some of them will be add-on applications like the keyboard mapping and that utility. But we're working upstream with GNOME

on battery thresholds. We just filed some issues up there and we're working on design. So we hope that that's something that will be usable not only by SystemDX 6 customers, but there are standard interfaces, kernel interfaces for this. So that means if we can get this work done,

other laptops that do have this type of functionality in their firmware will be able to take advantage of it too. Carl, don't you know how capitalism is supposed to work? You're supposed to keep all this stuff for yourself. You know, I don't think that's the way to do it anymore. You might be onto something. You might be. You know, I don't think that's the way to do it anymore. You might be on to something.

You might be. Yeah. You know, I don't know. You know, this does make me curious. You know, you guys have been exploring more, you know, with Athelio, making your own stuff. And there's talk maybe someday of, you know, in-house laptop. Has working at this low level of hardware given you guys any insights there to future designs it has um well in in some ways and that we want the work that we do to be as open source as as is possible given the

resources and supply chain we have so we've learned things from that but where we've learned where i think we're learning more is actually in our keyboard project believe it or not because in in that project we've designed a PCB. There's some secret sauce in the keyboard project that enables us to learn some other things and have great game experience with laptops. But the third thing it does is we're working on precision milling with the keyboard. Oh, sure. So with precision milling, we'll have a better understanding of how to approach laptop chassis design. in purchasing. So it's kind of a win-win. It's kind of clever, Carl. I didn't expect all of the perks that you mentioned with the open firmware. And I got to say, I didn't expect the, I don't

know how I missed that detail about the fact that you guys were working with Intel, but all this stuff looks really good. You know, as long as, as soon as I can get a laptop with a 2K screen, Carl, as soon as I think you might have a new customer. It's just a dream. It's a dream, but it's the panel manufacturers that need to deliver. Have you seen the Adder? I just have to ask. I don't recall. I do know about the Adder, but does it have a dedicated GPU?

It does. Oh, okay. See, I thought it was Intel GPU. It's dedicated, but hybrid graphics. And it has an OLED display that is the absolute most gorgeous display possible. Yeah. It's striking. Really? It has an OLED display that is the absolute most gorgeous display possible. Yeah, it's striking. Really?

I can't even put my laptops next to it because mine look terrible next to it. Okay. Too good. Yeah. Hmm. All right. All right. Okay.

I'll have to give that one a look. Well, Carl, thanks for coming on and chatting about that. You're welcome to hang out for the rest of the show or you are cleared free to navigate as unique. I'm sure you got lots to do, too. But appreciate the update on stuff. He's got keyboard settings to tweak. Come on, Chris. Thanks for having me, guys. Yeah. Thanks, Carl. It was great. Linode.com slash unplugged. Get a $100 60-day credit on a new account. Linode solves the problem of servers, whether it's a big, powerful GPU-powered server or something that just gets a job done in the background that

you never think about. They've got the whole range to choose from. I go with Linode for not only my personal stuff, but for all of our Jupyter Broadcasting infrastructure as well. And Linode costs 30% to 50% less than competitors like AWS. And they have a much better interface and dashboard. But I think more importantly for us, Linux is baked into their culture. You might recall Jeff Dyke, who was an active Linux contributor, had developed a technology called User Mode Linux a decade ago, plus now. Wow. User Mode Linux allowed developers to create virtual machines within Linux computers.

And at the time, single-use applications seemed like the way to go. But some developers saw a different path. And one of those developers was a young technologist named Christopher Acker. He saw an opportunity to use the hypervisor technology, not only to build something like Salesforce or Amazon, but to build something that made cloud computing less complicated, less expensive, and more accessible to every developer regardless of where they're located or what their resources were. He built

a company around that idea, and that company is Linode. They helped pioneer cloud computing. And now, two decades later, Linode is the largest independent open cloud provider in the world with 11 global data centers. The culture and experience in the Linux community is baked into every aspect of the company from their own internal culture, but also the controls they give you over the servers themselves. For my infrastructure, I generally go with a mid-tier system, but for the boxes that do encoding, we went with dedicated CPUs. And for my system that essentially just acts as a jump host, well, that's a $5 a month system. I take advantage of object storage that's backed by native SSDs and enterprise-grade

40 gigabit connections. It's really fast, it's easy to use, and they're part of the Linux community. You get started by going to linode.com slash unplugged, you get that $100 60-day credit on your new account, and you're off to the races. Go see what you can build. Linode.com slash unplugged. I fantasize about an open power, you know, like an open power PC laptop. Who doesn't? And it turns out that the Open Power Summit was this week, and there has been an update on the status of a powerPC laptop. It's not PowerPC like you used to know it. IBM's really been promoting how in they are with how open this thing is.

They also released a bunch of other stuff open source over the week. We'll have a link to all of the video presentations, which cover a ton of the announcements in the show notes. But the Linux Foundation had the Open Power Summit North America edition is what it's called. I guess there's multiple editions. I didn't realize that, but the recordings are online. So we're going to leave most of the homework up to you. But there was one that did catch our attention.

Roberto Innocenti provided an update on the Power64 laptop that is still being worked on via community donations as an open hardware platform now we should know before you get too excited mostly this is r&d right now as opposed to being focused on making or selling a physical product but you have to start somewhere we have this difficult know how to solve this was not simple to find even already built body notebook body but at the end thanks to slimbook we was able to to to find this solution so nowadays we have the possibility we have designed our motherboard especially for slimbooks eclipse this is the the model that is designed our motherboard especially for Slimbook Eclipse.

This is the model that is game-oriented notebook body. Yeah, the Slimbook Eclipse chassis is what they're aiming at, and they think that it's kind of like, you've got to know the chassis

you're building for to really do all the electrical stuff properly. So this is going to great to just get this nailed down. And they're going to include a GPU in there. I don't know where it really stacks up in the grand scheme of things. But this entire project's really been going since 2014. It's slow R&D, but there's been various points along the way where they've done a community fundraiser to get over whatever particular mountain they happen to be climbing at that moment. And they're reaching another one of those moments right now. And I got to say, it's not going very well.

This is getting really close to an actual laptop here, an actual power PC laptop that isn't a known good chassis. But the funding is stalled. They are aiming for five thousand and they've got eight hundred and ten of non-U.S. funny money. And the reality is, is that that trajectory, they're just not going to get there. There's been a couple of two hundred dollars. Well, not dollars. It's funny money. There's been a couple of donations here and there. But I don't think it's going to make it, Wes. Unless this changes or they get the word out. And this has been running since the 7th of September. Unless they get the word out, I don't think it's going to make it. Yeah. I mean, if you're interested, I guess now's the time. We were recently debating, discussing, contemplating the future after the NVIDIA ARM acquisition. Yeah.

And we sort of jokingly threw out open power. But it does turn out that there's a lot going on in this space. And maybe we all need to be paying a little more attention and just investing a little bit more. Maybe it's a pipe dream. Maybe everybody just feels like it's, you know, they're writing it off. It just seems like it's been going since 2014. And by the time they're done, it's not even going to be really a competitive system. It's just going to 2014, and by the time they're done, it's not even going to be really a competitive system. It's just going to be a—

Right. I mean, you're still in those stages of trying to get it to work at all, really, right? To get it all pieced together, get it in the right form factor. And then by the time you're done, well, it's hardly going to be competitive with the latest XPS 13 or System76 offering. Yeah, yeah. Well, okay. Speaking of ARM, you know Endeavor OS OS, which born out of Integros. Well, they're aiming for an ARM launch themselves. So you'll be able to get Endeavor OS on ARM, which since Manjaro has been pretty much the only arch derivative targeting ARM, they've kind of ran away with it. But now Endeavor OS realizes maybe we missed the mark. they've kind of ran away with it.

But now Endeavor OS realizes maybe we missed the mark. Now, we probably should note that Endeavor OS at the moment is still a small team, so they haven't tested it on all possible ARM devices. That's just crazy. But if you're running like an Odroid or, of course, a Raspberry Pi 4, well, then, yeah, you're in luck. Yeah. Let's file Endeavor OS away for talking about later on in the show today, because I think that kind of comes up in the area that we're going to be talking about.

Yes. In the meantime, let's do a little housekeeping. Now, this here program is live on a Tuesday at noon Pacific over at JBLive.tv. We were off last week, but we're back. I mean, if you get the download, we were never away. You won't notice a thing. But we're back in our regular live time, and we'd love to have you join us in the Mumble Room. Also, if you're around on Sunday at the same time, just a different day, the Luplug is going. You can get in there and just chat Linux with like-minded folks in our Mumble Room lobby. And then, you know your Mumble's all set up,

so when you're playing hooky from work, you could jump in there and join us on the show. That's great. You're already set up. Why not do it? And also, last but not least, thank you to our members. Our core contributors are helping not only this show run independently, but right now, they're helping do the production. Like, it just wouldn't be possible without them. It's fantastic.

If you'd like to be a Core contributor, you can go to unpluggedcore.com. You get two feeds to choose from. The live version, which is a lot longer of a show. So if you have a long commute or you just want more Linux Unplugged. You want to hear all of our mistakes. Yeah. Or, you know, welcoming guests on as they join us and we get all set up and all of our friends join the mumble us and we get all set up and all

of our friends join the mumble room and we all catch up with each other. That's what's going on there. And then, of course, in the post show, we're usually fired up about something from the show. We usually get into something and it just doesn't make it into the show because we also want to have a tight edited version available for everybody. And if you remember, that's where that second feed comes in. You get a version of that show with limited ads. There's some that are contractually obligated to be in there, but most aren't. And we cut them out. You still get the same great production, just no ads. And those are available for our members at unpluggedcore.com. Unpluggedcore.com. Thank you, everybody who joined up. We really appreciate it. And Mr. Payne with

that. That's the housekeeping. Still keeping it pretty tidy these days. Yeah, it looks good in here. I got the natural light, a few more houseplants. Ship shape. You got houseplants? Really? Do you really have houseplants? Of course. Huh. I don't know. I just, I didn't picture you as, not that you wouldn't have houseplants because because I mean, I'm sure you have a lovely place. But what I didn't picture is you being the kind of guy who takes an afternoon to go out and buy plants, you know, because it's not like you just it's not like a precision thing where you just run out and get a plant. You got a plant shop. You got to figure out what plant place you're going to go to. You got to get plant maintenance supplies. And that's where I just didn't really see you being into it. You got to get plant maintenance supplies.

And that's where I just didn't really see you being into it. Well, it helps when you have too many friends who are also into plants and just offering you cuttings all the time. I see. So they come over and they go, oh, Wes. Plant pushers. Wes, we got to do something about this. We got to liven it up in here. There's no green.

These poor dogs. Well, OK, so let's talk about some poor distributions that despite how interesting and compelling they might be, I don't know if they're ever going to get traction. It might be my bias. You may have noticed in this here program, we don't do a lot of distro reviews anymore. Used to do a lot of distro reviews. But now we kind of just do the big heavy hitters because I think and it's just been an unspoken bias that may have developed, I can't really see me coming on this show and telling you about how great Deepin20 is and getting more than 5% of you from actually switching

from your actual daily driver Linux. It just doesn't seem like it's going to happen. Yeah, you know, this really started quite the discussion between us because we were, you know, Deepin20 came out and it was on our radar. We were interested in it. I mean, there's a lot about it to like, seemingly, at least from the bit that I've played with it. But yeah, who are we recommending it for? And where does that fit in where Linux is going and where desktop Linux fits in 2020? And here's a scenario for you. Okay, so we're not going to be able to convince a diehard arch user or a person who believes in the elementary way of designing an operating system

or somebody who has been an Ubuntu user and used it for work for years. We're not going to convince them that Deepin 20 is something that's even worth them really trying because they've got work to do. They're set in their ways. You have people who like to hobby hop and they'll try it out, but you're not going to get actually any real traction. Like the distribution isn't going to see millions in growth. And so then I thought, well, maybe it's for new users, new users who are already not set in their ways. You know, they're open to the possibilities that the many, many distros provide. But then I thought, wouldn't that be a disservice to that end user if you set them up on a niche distribution and then expect them to just go about their computing life as if they can just get everything done that they could if they are on Windows or Mac or Ubuntu?

No, of course not. Now that they're on this niche distribution, it's much harder for them to do any research and find support, especially if their Google foo isn't tuned to years of Linux searching. That's just it. Sure, Deepin has a Debian base, but if you don't really understand how all of that works and you're Googling specifically for the OS that your system tells you you're running, that might be more trouble. Well, and how would they even be able to deduce what the base OS is, right? I mean, that's just, they're just not thinking like that. That's not how a consumer that isn't actively selecting and choosing an operating system like this conceives of the construction of the computer. That's just not that's not where their head is at. Right. So Deepin isn't great for new users then necessarily, unless it's a managed environment, perhaps, where there is an administrator who can guide and help users. But that's a pretty niche case for a pretty niche distribution all of a sudden.

and help users. But that's a pretty niche case for a pretty niche distribution all of a sudden. And so as you whittle it down, you start to realize we really have no shot of convincing an Arch user to move off Arch

or a Fedora user to switch to Ubuntu. I mean, they may elect on their own based on some of the features we talk about, but I think I just started realizing there's almost no point to it other than just for

the technical exploration, which I enjoy and I think is worth it on its own, but it doesn't really seem to go beyond that. And so I can't really fathom a reality where a distribution like Deepin or a distribution like Endeavor OS really has a shot of ever going beyond a few hundred thousand users. You know, Minimac over in the IRC raises the question, if we're not going to be spreading the word, who will? Yeah. I guess the question, Brent, is what is their definition of success?

If it's perhaps to scratch their own itch, then, you know, check. But if it's to attract a rich community, then there needs to be a way of spreading that regularly. Right. And maybe there's a fatigue there, fatigue for finding new, interesting stuff. Yeah, I could definitely see that. And I think there's also the question of relevancy, the operating system as a whole. If you were to just take the base temperature of the computing industry, the importance of the desktop has rescinded a bit. And the importance of the web browser and the importance of mobile has ascended quite a bit. And I think you have to

consider that has to have some sort of perceptive knock-on effect on the Linux desktop as well. It has to. You could imply from that that, well, if the desktop itself has become less important, then the operating system it runs has become less important. And then the distro has become less important. And it's really, can I run Firefox and Chrome? And then there's the essential question of what's the base aim of the distro, Neil? Like what market is it going after? And to that, that seems like that's where it appeals to each individual user. There's usually two different aspects. What is the novel aspect of the distribution?

And what is the goal of the distribution? So, for example, if I take my preferred distribution, If I take my preferred distribution, Fedora, the novel part of Fedora is that it's a community that attempts to integrate all the software and make it work together while simultaneously working with the upstream projects. So, like, things fit together, but not in a way where nobody else benefits from it. Now, then, you know, that fits the novel and the aim because it's both about bringing the latest stuff in, which is part of the aim of it, to support bringing the latest stuff and making it available to people. And, you know, what is the novel aspect of it

is that instead of just doing it on its own in its own island, it aims to contribute and benefit the larger Linux community. Now, if you look at, let's say, a Pop!OS, the System76 Pop!OS platform provides them an opportunity to experiment and develop a more integrated experience with the hardware and software. That doesn't necessarily imply that it needs to be an Ubuntu base or a Debian base or a Fedora base or an OpenSUSE base,

but having a platform in that sense that they can ship with their hardware allows them to demonstrate in the best possible light the possibility of using their hardware. their hardware. I agree with your point to an extent, except for I would argue that the Ubuntu base was not only a safe bet for System76 and an encouraging bet for their existing customers at the point in time had been using Ubuntu that shipped with the laptops, but it also was a signal to enterprises and customers that if your software works on Ubuntu, it'll work on Pop!OS. Just like Carl mentioned in the interview, if you're on any Ubuntu base, you can use the System76 updater. That, I think, is a contributing factor to its success. But it is also an outlier here because it is so highly coupled with the OEM's overall ambitions and goals that it clearly serves a purpose in multiple respects. It's a distribution appealing to end users. I have it running myself, and I think it's pretty good. And it's obviously serving some of their longer-term goals. So I think Pop! OS is kind

of an exception here. Well, actually, it's not an exception because PC Maker, Boutique PC Maker, that actually does this. Tuxedo Computers has Tuxedo OS. Yeah, but have you looked at Tuxedo OS? I think you can make this argument with Manjaro. I think Manjaro demonstrates it was a niche distribution that now actually has some significant momentum and hardware deals. So it's clearly possible Elementary OS has seen their Pantheon desktop environment, trying to show an integrated design-centric user experience that's very tightly controlled. And with that, you don't really have a lot of value. Deepin attempts to showcase the Deepin desktop. And maybe they actually do ship on hardware in China. We don't know because we don't really go down that road. as well as some kind of avenue in which it satisfies or fulfills a particular market demand. Yeah, we were talking about this, like package managers, desktop environments, installation,

thoughts to how you treat the whole end user experience. When we talk about distros, those all get lumped in together. So Brent, let me see if I've got this right. You have if it's serving the right audience and if it's built in the right philosophy, using the right tools with the distribution, which reflects itself in the way packages are handled and installed. But then you also have the support aspect of it, which I think for a lot of us in this conversation is trickier to think of because we are often our own support. Being your own support also means finding the answers. And so I know for my own trajectory into the Linux world, I've had to work hard at, you know, typing into the keyboard, into the search engine to find a bunch of answers. And it seems to me that, you know, I'm not a crazy super user like some of our wonderful guests here, but I've been around Linux for a long time.

And so my ability to find answers about the little tiny things that go wrong on any system has been a big decision in my choice of which distribution to run and my hesitation to jump to some of the lesser knowns. And it's too bad that that's a reality for me personally, and it may be for many others, but I, I, you know, I, I just want to get work done most of the time. Sometimes I play. Yeah. Um, but that, that's, that's a hesitation for me for sure. I have to say, I mean, you know, there are some resources that I always go back to and it's often the best resources I find are, and I'd say in this order, Arch, because of the Arch Wiki, roll your eyes, then Ubuntu, because you have the Ask Ubuntu community and all of the related blogs, and then Fedora, which not only has a big community, but also has really good

documentation. But I think this is for us, Wes. And, you know, a couple episodes ago, we had Wimpy on and he got us thinking about the younger generation, like, you know, my kid's age and a little bit older and a little bit older than that, that are going to be coming over to Linux and open source. The next generation of Linux. Yeah. Yeah. And you got to wonder what's where are they going to land? Yeah, you do have to think about that. Right. Where are they going to land? Yeah, you do have to think about that, right? And what makes sense in a world where we're trying to, you know, seduce them over to the light side of things?

And, I mean, it's probably not going to be a niche distro, right? It's going to be one of the big ones. Or maybe not. Okay, how about this for a theory? The niche distro pulls in the technologically curious out on the edges. Like if you look at it as a sales funnel, the niche distros are on the edge of the funnel. And then they get people in because they look flashy and do something new and cool. And then as people try them and they get more experience and more practical, they move inwards into the funnel.

And by the end, they're coming out the other end running Ubuntu. That's tough. I mean, you're right. I think that works for the very curious or the folks who are willing to do more troubleshooting on themselves. But the advantage to just starting with whatever the latest Ubuntu is, is, I mean, you've got all the Stack Overflow answers, right? You can Google any random error that you come up to when you're trying to first try out this strange new Linux thing, and it's just an easier upload path. Yeah, and then you obviously see some specializations getting a lot of success.

I think that's a clear point. Brent, you have some examples. Yeah, what came to mind right away was the Kallies and the Cubes who just kind of do one thing and do it really well. And maybe that's the Unix philosophy coming in there. But we also see some really popular ones like Raspbian and the likes that are just targeting a small selection of the problems to solve and and they're seeing a lot of success so maybe what we're going to see uh moving forward is just a lot of this sort of you know if you want

to play in this specific playground you know as raspberry pies and such then you have a distro of choice or two or three um but expecting them to be your daily driver, I think we've pretty much sorted those out for most people. Kali is a great example. That's for sure. Well, I don't know really where to go with this other than now I kind of feel good about niche distros when I thought of the whole funnel thing, which kind of made me feel better.

And I also have to say, go good, just look at Deepin 20. Yeah, you should. You really, I mean, it looks like a lovely release and I don't want to come off too harsh here in that I think that some of the wonderfulness of Linux is just this, you know, you have all this free and open source software that you can experiment with. And we have a rich ecosystem of different contributors all playing around with things.

And all right, you might not use Deepin, but down the road, maybe you use their desktop environment on your favorite distro of choice. I'd also love to get the audience's perspective on the future of niche distributions and this overall topic. So go to linuxunplugged.com slash contact and drop us a line of feedback and we'll cover it on a future episode. Nathan wrote in, I was listening to episode 227. Okay, so this is 372 right now. Okay. It was about Devon, which is how I like to say it, and System D, which is how I like to say it.

He says, I just wanted to add my two cents to it. I'm not as strong as some are against it, but, he says, I do not want it on my system personally. Not all alternative inits are dead, such as OpenRC, which is the closest de facto replacement. Personally, I use S6, which is fast and also maintained. Interesting. But he goes on to say, what I have against SystemD is that it feels like it intends to replace everything, in which now they have replaced my home with it.

At this rate, it wouldn't surprise me if it eventually replaces the kernel in the future. This may be a bit dramatic, but I think you get my idea. I just like simpler systems myself. You know, I can certainly appreciate that. And that's one of the reasons I might be drawn to an Archbase distro, just to talk about our last topic a little bit more there. But really, Nathan, I just appreciate that you're still, you know, you're listening, you're giving us feedback, even if you aren't exactly caught up. Yeah, I agree with him.

And I also recognize the trouble it causes. So since that episode, I think I've come to the conclusion that Dev1 is, it's a good outlet for that idea of a system. And what I kind of think would have been better is if that whole shenanigans had just left the Debian project with that. And that's not a disparaging remark. I just simply mean,

let each project have its own clear path. There was a news item we were debating coverage that is just some silly technical arguments going on once again in development in Debian land around systemd. And it's like one of these super simple maintainer versus somebody submitting the patch thing, and maintainer has a hardline position, and a bunch of other people would like this thing to support both systemd and non-systemd systems. other people would like this thing to support both systemd and non-systemd systems. And it's just turned into a kind of a nasty debate that's now making circulation on social media and kind of painted individuals in a bad light. And it just comes down to this core conflict around systemd

and Debian. Well, and the issues that, you know, the trickiness of managing the support when so much of upstream has just totally, you know, gone the system D way? And then to what extent are downstreams obligated to continue, you know, keeping support for these other init systems, which, to be clear, I think we should support. I mean, it's nice that we can have that diversity, folks that want a simpler system or just don't buy into the whole system D mess, I'm fine with it. But if you're not, that's cool. It's interesting to watch what that means that like what what buy in do you have from your users? What is worth it from a maintainer's perspective in ongoing maintenance to keep making sure that that works, especially if

maybe at this point, none of the systems that you actually have access to test on are using any of those old in it. Yeah. And what should their burden be? Probably worth addressing this system D home folder stuff. Carl, you've experimented a little bit with it. My understanding is you install your distribution and it immediately takes over your home folder, removes all your data and encrypts it with a password that it doesn't share with you. Is that how it works? Of course. Take over the world, right? over the world, right? No, that's obviously you're joking. That doesn't work at all like that. HomeD is a hundred percent optional and I really don't get the consternation about it in the community. It's, if you don't want to like it, don't use it. It's just, it's a thing that's there for you if you want it. Yeah, absolutely. And it's one of those things too, where you look at it and you go, oh yeah, okay. If we were making an operating system today, that's just absolutely how we would

do home directories. I mean, it makes total sense. We've covered it in the show before and why it's cool technology, so you can always check our back catalog. But yeah, it's still early days, too. There's a lot of integration to go with PAM and SC Linux and other things on various distros. So it's a tool. If you don't want to use it, you can completely ignore it. Yeah, but I do appreciate him sending that feedback in. And I guess we're just still in this phase. And there are resources out there for everybody. It's a big Linux world. Steve writes in about EmacsConf.

Speaking of a big world, yes, EmacsConf. Have you heard about the recent announcement for EmacsConf 2020? It's a free two-day online Emacs conference scheduled for the weekend of November 28th and 29th, 2020. Those of us on the Pacific Coast have to get up early to catch the first few sessions, but recording should be available online. The conference is currently asking for proposals at emacsconf.org slash 2020. That's amazing. I mean, when you're finally over Nano, Chris,

it might be just in time to join and attend this conference. You know, one or two things is going to happen. I'm eventually going to win you over somehow with Nano, or we're all just going to be using VS Code on the command line. Wait for it. I think the latter is more likely. Okay, so how did we end up with this huge mess of picks? I'm looking at this like in two ways.

Like we could save some of these and we could just coast for a while. Or we could like toss out a whole batch of them, which is kind of the one I'm starting to lean towards just because we have so many. But I look at this list. There's two obvious ones I think we should do. Yeah, okay. I think I know which ones you mean, but let's find out. Oh, yeah. Okay.

Well, that's a clear clue. That's right. We got two Rust picks. Now, the first one's for you server types. It's the Cloud Hypervisor. It's a Rust virtual machine manager based on Intel's open source project, if I recall correctly, called Cloud Hypervisor. And it's a tool written in Rust

because, well, it's the language of choice now. Well, so I think this is kind of interesting. It's early days for the Cloud Hypervisor project. But, you know, we've talked a bit about Firecracker. Google's got their own sorts of, you know, tiny little virtual machine monitors that are implemented. And those are really focused on running these like super lightweight containery workloads, right? But Cloud Hypervisor is a little more general. It wants to handle more of your day-to-day, like, you know, you want to run the regular Ubuntu cloud image on this thing,

but you want to self-host it. Cloud Hypervisor might be the virtual machine monitor for you. And the point about Rust is mostly that, you know, Firecracker and some of the Google work, they're all powered by this Rust VMM crate. So it's a shared upstream library that these are all based on contributing to. So it seems like Rust is becoming something of a good place to do this kind of development. Yeah, it's KVM based. So it checks that box. They are going for minimal emulation and lowest latency possible, low complexity, small attack

surface. Check, check, check. Not only are they targeting the x86-64 platform, but they're also going for the ARM64 architecture as well with obvious functionality differences, which they've documented. And you're right. It actually does look legitimately like it could be a great way to self-host VMs. You know, I was mentioning Proxmox before, but, you know, there could be a future where it's something like that instead.

Now, on the other end of the Rust picks, we have SongRec, which is an open source Shazam client for Linux, which, as you guessed, is written in Rust. as you guessed, is written in Rust. What's kind of notable about this particular Rust app is that this is a GUI. It has a command line version as well, but this is a graphical desktop Rust application. Is it like Python Rust? What is this? Yeah, we were sort of debating, like,

how common is that? Because most of the picks that we've been finding these days, well, they're all neat, but they're all command line apps. Yeah. All right. So there you go. There's a couple now. Here's a couple more that have nothing to do with Rust at all. These are pretty cool that we've been sitting on for a little bit. So now Crazy Chris has given away picks. It's more than two picks. It's going to be probably four picks. This is one I think is probably going to improve your quality of life. Temp mail. TMP mail. A temporary email right in your frickin' terminal.

Right? I mean, the reason to use this is obvious. What else do I need to say? It's a command line utility that allows you to create a temporary email address and receive emails to a temporary email address. It's using, was it OneSecMail and their API, I think? Oh, they've got a few. They've tied in with ProtonMail with some of the other, you know, popular sort of just accounts that let you easily set up a pseudo anonymous mailbox. But you don't have to go log into, you know, use a set of different GUIs on the web. You've just got one command line interface.

That's so cool. And really minimal dependencies, W3M, Curl, JQ, and Auken. You're off to the temporary email races, and that's pretty neat. All right, I'm looking at the mix here, Wes. Do you want to pick our final crazy giveaway? Oh, more picks. Let's do it. All right. Well, we picks. Let's do it. All right, well...

We have too many good options here. How to choose? Well, how about for those self-hoster fans out there, what about MyPass? My P-A-A-S, or Platform as a Service. It's basically a system to enable, you know, if you've used something like Heroku, you're already running things with Docker Compose and Docker tooling in your local dev environment, and you want your own self-hosted cloud platform that's as easy as Docker Push. Well, maybe give MyPass a try. It's powered by Docker,

Traffic, another one of our favorites, and it really espousesPass a try. It's powered by Docker, Traffic, another one of our favorites, and it really espouses its excellent analytics. It's got some dashboards built in for you, so you can actually see what the heck's going on on your new platform. Yeah, it does. I love these. This is really cool. So this is one I did not open the tab to, and then I opened it up as you're talking about it,

and I had not seen the screenshot with all the stats and stuff. I love this stuff. That's really cool. So we basically gave people two picks to just turn their network into their own AWS is really what we just did. You know what? You're welcome, everybody. How about that? So, yeah, this is just sitting on top of Docker and traffic.

It utilizes Let's Encrypt to get your SSL TLS certs. And MyPass, which I didn't know really much about it, but looks pretty cool. Nice pick. That was a good one, Wes. That was absolutely, I'm going to say, right ho. If you've got something you'd like to share with the community, head over to

linuxunplugged.com slash contact or link me up on social media. I'm at Chris Lass. What about you, Wes? I'm at Wes Payne. Our sponsor, a cloud guru. You can find them at a slash cloud guru at YouTube, Twitter or Facebook. Go get them on social media. The show is at Linux Unplugged. I don't think it's Linux Unplugged show. Some of them are on the Twitter. And then, of course, the network. That one I do know for sure at Jupiter Signal.

Links to everything we talked about today are Linux Unuxunplugged.com slash 372, including links to our guests and all of that good stuff. And for our members, remember, you have special feeds, limited ad and full live stream feeds. And you also have that Fedora 33 Bugathon, which is available as an exclusive download to our core contributors. Exclusive! Thanks to our members.

And also thanks to everyone for tuning in to this week's episode. Even if you're not a member, we sincerely appreciate you listening or taking advantage of our sponsor or sharing the show. It means a lot to us, and we never forget that. We talk about the livestream and we talk about the members, but in our hearts, it's you downloaders. See you next Tuesday! Thank you. So we have a complicated hardware situation. We have one machine that's doing great, and that's the least important machine. And then we have the recording machine that has bad blocks on the recording hard drive.

And it's also just sort of shoddily put together. And now after this weekend, the Reaper machine seems to be in a bad state. We had a power outage. We lost one of our audio interfaces and Jack audio has been wonky ever since that audio interface died. The machine has what I can only describe as like these blink out sessions where the entire system sort of locks up and the screen flashes. Initially, we thought it was Gnome Shell. So we switched to Plasma and it went away for about a year. And now it's back in Plasma and it just creates such a headache. So I got two machines right now that are in a super bad shape and no money to replace them. I'm trying to get clever.

I'm trying to think if we were to. So this this Reaper machine is also the machine we use to do mumble. It's the machine we do remote interviews and hosts with and recordings. How beefy are your individual machines? Oh, they're all in like the I seven, you know, six core area, but nothing, nothing incredibly powerful because we're not doing video. How old are they? Ah, that's, that's the problem. It's been probably, I think we built them in 2016, 2017. All right, I've just got one question for you, Chris.

Yeah. How many pies have you ordered? Yeah, just a fleet of pies to replace them, Wes. It's going to take 62 pies by my math, but it'll be totally worth it.