LINUX Unplugged - Episode 12: Debating Debian Decisions | LINUX Unplugged 12

Episode Date: October 30, 2013

Upstart or systemd which will Debian choose? We’ll discuss the inherent benefits and disadvantages of both, and the larger ramification Debian’s decision will have on the Linux ecosystem.

This is Linux Unplugged, Jupiter Broadcasting's weekly talk show that believes sometimes the esoteric aspects of your operating system can be the most important aspects of your operating system. My name is Chris. My name is Matt. Hey there, Matt. Episode 12, buddy, and we're chugging right along into pretty technical territory this week. Oh, it's pretty deep stuff. Let's just put it that way. I'll be curious to see if we can pull this episode off because, first of all, it's hard to be experts in either one of these topics we're talking about today because they're moving fast. And they are very esoteric.

They're very low-level stuff. And a lot of people either don't know what they are, don't know why they should matter, or feel like it'll never impact them. But I hope after this week's episode, we change that line of thought. That'd be great. Kind of a big goal, but we'll see. So this week, Debian's been asked to make a choice to pick between two different startup systems. One created by Canonical that's used in Ubuntu, Red Hat Enterprise Linux 6, Chrome OS called Upstart. Obviously, Ubuntu is a major derivative of Debian, so it's a pretty important decision for those guys.

And then SystemD, created primarily by Red Hat, Intel, and others, that is shipping in Arch, and it's going to be shipping to Red Hat Enterprise Intel, and others, that is shipping in Arch, and it's going to be shipping to Red Hat Enterprise Linux 7, SUSE. A lot of different distros are picking up SystemD. And they all do different things. SystemD does a lot of things, and that's one of the actual criticisms against it. But people are really now watching Debian to see which way they're going to go,

because whichever way Debian goes, probably the rest of the market's going to go in whole. Because if Debian goes with SystemD, then it's almost de facto standardized across all of Linux, except with the exception of the large distribution that some of you might have heard of called Ubuntu. I think it will be interesting. Yeah, because if they go with Upstart, then there's two very large distributions that many other distributions are based off that will not be using systemd. And it will sort of introduce sort of this long-term differentiator between the two sides. So it's kind of all eyes are on Debian at the moment. And this is a topic that has been being debated within Debian.

I remember reading a post on LWN.net, which after my comments in last week's episode i went over and purchased a year subscription to lwn.net which is once you get in behind the paywall it is a totally different experience so they should tease some of that out a little more because it is actually really great uh anyways i read an article at lwn back on july 27th of 2011 saying that debbie was debating if they would switch to systemd so this has been going on for a really long time, but it's getting serious this week, Matt. It's getting serious. You said 2007?

Oh, I'm sorry, 2011. It was July 27th, 2011. I was going to say, my goodness. Yeah. It feels like that. It feels like that, honestly. And, you know, it's funny because I actually remember covering Upstart. Upstart landed in edgy.

What was that, like Ubuntu 6.10, I think. And, of course, they weren't fully converted over from System 5 in its scripts until, I think, like version 9-something of Ubuntu. But I remember, like, talking for a very, very long time saying, man, Ubuntu's main feature is just it boots faster, it boots faster. Well, during that period of time was when they were transitioning to Upstart and uh meanwhile debbie's just been hanging back been like ah you know what we'll let the kids figure it out and we wrote we rolled debbie and stable here we rolled debbie and stable so we let the kids figure it out and then we package it all up once they've got it all sorted out now of course uh system d and upstart have both been available in debbie and testing for a while so we're going to talk about that but before we talk this week, I wanted to get into our follow-up, because we like

to continue the threads from previous week's episodes, and we want to do that right up here at the top of the show. And we talked last week about sort of the sad state of Linux news, and how it was so personality-driven, so cult of personality-driven, and you talked a lot about it from a writer's standpoint. I did, because you've got to understand that a lot of times what i put out there is assigned and then i have the option of saying i would like to eat this week or i would not like to eat this week and i basically and usually it's a pretty straightforward decision for me that i enjoy

food and it's actually not as not as evil as that because that kind of makes it sound evil but it really what it is it's like well if my editor or I don't know who's telling you, but if the editor or whoever it is that gives you your assignment, aren't they basing that on what they perceive to be interest from their audience? It's not the editor's fault. It all rolls downhill from way above, and basically it comes down to eyeballs, eyeballs, eyeballs. That's really what it is. So he's no more at fault than I am, but he does – he is the one that basically is the gatekeeper there as to what I'm able to write. The one that should know better. He is the one that basically is the gatekeeper there as to what I'm able to write. The one that should know better.

One that should know better. But again, he's in a fairly stable position, and he's not necessarily in a position that he wants to rock that boat because he is managing things that aren't Linux-related as well. It's funny, though. Yeah, so he's not an expert. But it's funny because in this episode, this is probably not the topic you pick to get a bunch of people to listen to your podcast. System D and Upstart, right? I mean people – so you could see how an editor would never want you to write an article about that, because that's not a guaranteed win, right? But something about, you know, top 10 lists

for something or whatever, top 77 lists, like I saw today, a top 77, okay, those drive clicks. And that's just one example. It's one example. And also the fact that a lot of the sites, because as you're well aware of, Tux machines being sold off and all this, the landscape is changing. A lot of people are finding, oh, wow, this is so not a money-making thing. It's really not. Yeah. So at least from the writing perspective.

So I think that folks are beginning to try for more of an SEO grab. Sometimes they're trying for a title grab. And neither are necessarily working super awesome. I found even on my own site, it's been affected. And sometimes it's this horse race that you run into where people try different things to grab headlines and those quit

working and they have to switch to something else. And this is sort of the point that Matt, not you, but a different Matt who wrote into the show was making. He said, you know, dear Chris and Matt, by the way, I'm a JB lurker, regular listener to Unplugged and a regular viewer of Last. I also enjoy TechSnap and FoShow from time to time. But he points out he just finished listening to Linux Unplugged 11 where we talked about this. And he thinks that, you know, this whole driven by personalities, talking about scandal instead of honest news.

This is really just sort of a slice of what's going on on a much larger scale, say on Fox News and MSNBC and everything in between. Profit-driven news outlets are only interested in viewership. Yeah, that's factual. And he said, anyway, it's just my two cents. Thanks for the great shows and for reporting Linux news that is important, if not always sexy. That's what we're going to try to do today. That is true. And I think I would counter that, too, in saying he's absolutely correct.

A profit-driven news center definitely has to take that into account, but I also think that even if you go nonprofit, you do have to have some kind of funding source because you have fixed expenses. Oh, yeah. Even if everybody was volunteering, which isn't always plausible. So all those things come together. There's a lot of things under the hood that people have to really stop and think about. So Andreas wrote in – oh, by the way, I think we probably got maybe 30, 35 emails just on that news topic alone. We got so much email on that. Linux Unplugged this week generated more email feedback than last did when we did the Synology review. And now Cinnamon has last reclaimed the crown with the Cinnamon topic.

But for one week, Linux Unplugged pulled ahead for episode 11 with the amount of feedback volume that we got. Goodness. So Andreas wrote in. He says, I need to get feedback on the question you asked in Linux Unplugged. Which distro will take the lead in a few years? When SteamOS is out and Ubuntu continued to fail on the desktop. I don't think we said that, but okay, we'll roll with it. He says, I'm pretty sure it will be SteamOS.

Almost every Linux gamer might have at least a dual boot partition with SteamOS. Okay, there will be ones who refuse to use it because of the Steam DRM stuff, but there might be a lot of Steam boxes sold and self-built out at this time point. If I got the question wrong, please ignore this feedback, but I think this might be it. I think Andres, you know what? At least as far as desktop deployment, it's going to be hard to beat Ubuntu and Red Hat on the server

deployments in Suice Enterprise Linux, and even Debian. I kind of disagree. I don't know. I mean, so for me, it's like I think it depends on the location and actual the task. Usually, most people use a computer to actually accomplish a set of tasks. If it's SteamOS, you're trying

to accomplish getting some games going. If you're using Ubuntu, you're probably trying to get real work done. You're trying to actually do something from a desktop experience. I don't see SteamOS being really popular at your local non-profit or at a school that's actually trying to get real work done. I mean, the kids would love you for it, but I don't see that happening.

I see Ubuntu having a much bigger impact. Alright, fair point, fair point. So I was just looking total consoles sold total Xbox 360 consoles sold to date looks like 59 million I don't know if, is that right? And the original Xbox, 24 million

there's 31 million Xbox subscribers. So almost 60 million Xbox 360s sold to date, okay? And then, you you know 24 million original xboxes that is an s ton of xboxes don't get me wrong but to put that into some perspective there's been 170 something a million ipads sold right and there's been um and and which which works out to be i can't remember the number now because i got it third i got it third hand but it works out to be something like represents six months worth of desktop PC sales.

So you put it in perspective of desktop PC sales, it is not even half the year worth of sales. So it is, yeah, I think SteamOS could get some good traction on the console, but I don't know if it's really going to change the tide at this point in any one direction. I think it's just going to be a strong contender. Definitely a high number, but I don't think it's going to be number one. I tend to agree with that. I think that you're trying to undo time and undo progress that's already taken place. And so if that does happen, it's going to be a five-year plan. If it doesn't happen, they'll maintain a real comfortable standing, but it may not

ever become the big thing. If it did, that'd be pretty interesting times. It'd be awesome. I'm totally rooting for them. I want to see them absolutely just mop the floor with everybody. I think that'd be great, because what an exciting thing to watch. People are really passionate about it.

We have to wait and see how things play out. Now, we got, like I said, so much email, and I'm going to be totally up front with you guys. I didn't get a chance to read it all. So if you didn't get a response from me, I watched a lot of videos on SystemD and Upstart in the last couple days. I'm not even kidding. So I actually didn't get too much of my email. I used my email time to do that because I just kind of fell down the rabbit hole of these things because it's actually pretty interesting stuff. So if you want,

you have my permission to resend. And also, I encourage you to go to linuxactionshow.reddit.com and while you're over there, start a thread if you think it's something maybe the broader community could chime in on. Now, Matt, before we get into our main topic today, I want to thank our sponsor, and that is Ting.com. Matt? Yes. Ting is mobile that makes sense. They're my mobile service provider. Now, Matt, before we get into our main topic today, I want to thank our sponsor, and that is Ting.com.

Matt? Yes. Ting is mobile that makes sense. They're my mobile service provider. They're Matt's mobile service provider. And I'll tell you, Matt and I, we are men who prefer to be free agents. That's right.

I don't like being in a contract, Matt. I don't like early termination fees. What is this, some kind of scam? Oh, totally, yeah. I like having control over how I use my phone. Exactly. And you know what's great about ting is it's is so not only is there no or the termination fee not only is there no contract i mean that's a great start i'd take that right there oh you also only pay for what you use so you take they take your minutes your megabytes and your messages at the end of the month they'll just add it up whatever bucket

you fall into and they've got some great examples on the ting website um my bill right now for example and it uh it is it is up in four days the ting dashboard tells you they have an amazing dashboard over at ting it's it's got to be um it's not just like it's not just like good from like an interface standpoint but it's also good from like ting knows the internet like they are one of the first companies that are on board with Mozilla Personas, which is a very, very great authentication system that sort of is an alternative to OpenID and other things like that. Ting was on board right away. Ting knows about this kind of stuff. So not only do they have a very awesome dashboard, but they actually get web standards. They really get this kind of stuff. And you go right in, they say, yeah, you got this much time left on

your account. You're currently at this and they have nice like gas gauges that tell you exactly where your usage is at. And oh, by the way, this is your billing period. It'll be up in three days. So when I go in there right now, it says my billing period has 41 or four days left. I'm at 21 bucks. I'm at 21 bucks. And man, I, I, every time I'm in the car, I usually have an audio stream going and I usually have waysaze going. I don't pretty much go anywhere without Waze going all the time and even still. Now, I'm not a big voice user because I'll use Skype and things like that, but that's what's so awesome. As the holidays come up, I'm going to start using my voice minutes a lot more. Now, I'm not paying into some sort of plan, so it's like, okay, well, I get 200 minutes this month and if I use more than 200 minutes, I have to pay extra.

And then, of course, when I'm not using the phone very much, i.e. not the holidays, I don't pay for that phone usage. It's just a flat $6 a month. I love the fact that they have a fluid bucket that will adjust to my month-to-month needs because they do fluctuate. I mean, it really does fluctuate for me, so I love that. Absolutely. So go get started over at linux.ting.com. That'll take $25 off your first device or $25 off your first month if you're going to bring your own device. They've got a bunch of compatible devices.

Also, right now, if you actually think this is sticking around for a while, go take advantage of this. Go to ting.com slash ETF. They have an early termination relief program. They're going to give you up to $75 off per line that you have to get terminated if you're going to move over to Ting. And trust me, you're going to want to move over to Ting. If you just go to linux.ting.com, they got a savings calculator that'll show you the awesome amounts of money you'll save over the long run. And by the way, that savings will include hotspot and tethering right with the plan.

No extra package you have to get into. And it'll also include real personal phone support. When you call a Ting agent, they answer the phone and they see your problem all the way through to the end. It's like a real Canadian. I love that. I love the one-on-one connection you can make with these folks because instead of talking to a machine, you're actually talking to someone that not only gives a rip but can actually solve your problem. They're also, as far as telcos go, the most transparent you're ever going to find. So go over to their blog and just check out some of the posts over there.

They've just done an introduction to one of their team members. You get a little bit of behind-the-scenes info about Ting. You know, and as somebody who considers themselves a researcher of things before I buy, I really appreciate that they expose that kind of information to me because it helps me make an informed decision. So go over to linux.ting.com

and thanks to Ting for sponsoring. Linux Unplugged. Ish. Okay, man. So let's get down and dirty in the plumbing of Linux this week. And I want to talk about SystemD and Upstart and this whole Debian's choice. So I'm going to read you something. And it's so funny because you're going to find the same blurb all over the entire internet.

And it's even all over on the Ubuntu wiki on their SystemD page. It's on Arch's wiki. It's even on the main SystemD page, and it's on the freedesktop.org. It is the explanation of what systemd is. It might not help, but I'm going to give it to you, and then maybe we'll parse it a little bit. So what is systemd? Systemd is a system and session manager for Linux compatible with System 5 and LSB init scripts. Systemd provides an aggressive parallelization capabilities. It uses sockets and dbus activation for starting of services, offers on-demand starting of daemons,

keeps track of processes using Linux C groups, supports snapshotting and restoring of the system states, maintains mount and auto-mount points, and implements an elaborate transactional dependency-based service control logic. It can work as a drop-in replacement for System 5 and it. Now, to sort of parse some of this for you,

so it uses socket and dbus activation for starting of services. What socket and dbus? Well, let's focus on the socket part. So, for example, if you have a very lightly used Apache web server, System D can actually open up the port 80 socket and listen for incoming connections on 80 and leave Apache as completely closed. So it's not taking any CPU or memory until that very first connection comes in. SystemD will answer that connection on the port that the daemon uses,

and then it will go start up Apache and then serve the page for the person that they need. And the user never notices any kind of interruption. Additionally, it allows you to restart services without breaking the connection because since system D is managing that socket, you can have the underlying daemon restart without interrupting it. So that's what socket-based activation for starting services is. D-Bus would be message-based. Event-based driven, which is also Upstart, is things like network state changes or hardware state changes or user command changes. Those are sort of the more event side of things. contributors from all over, including a large group from Intel. It is almost a new layer of the operating system, to just give you an abstract.

It takes care of a lot of these things like mounting of file systems, logging. A lot of multi-purpose stuff under the hood, it sounds like. Yeah, it can replace syslog. They have a journal that it replaces, which some people don't like. We'll get into that. So there's a lot to it. And the core question that now Debian is facing is do we go with Upstart or do we go with SystemD? And I want to play a clip for you, Matt,

because we played this on the pre-show and I think you and I both agreed it was worth playing even though it's pretty meaty in the show. We're going to play this and we'll kind of parse it. This is Lernart Pottering, the primary developer behind SystemD. And he was at a Debian conference in August and they held two talks at this conference. The first talk was a SystemD talk, and the talk was why Debian should or should not

switch to SystemD. And I have that linked in the show notes. And the second talk the next day was why Debian needs Upstart. Obviously, one was from Lenart, one was from the canonical rep. And we have both of those linked. They're about an hour long if you want to watch them. And, you know, as we talk about this, keep in mind that this is from Lenart. He says, whether Debian chooses SystemD or Upstart, it has major implications for the future. So you shouldn't only look at what is now, but keep in mind what will come next. And there are at least two areas where opting for Upstart

will mean that you shut Debian out of major changes. So in this clip I'm about to play, it's from this 55-minute talk, it's about three minutes long, he gives some examples of where he feels, now this is Lenard's opinion, the creator of SystemD, where he feels that upstart falls short to SystemD. Of course, if you compare it to the other contender, which is upstart, we believe that, I mean, the reason we initially created System.ly actually is because we believe that the Upstart design was wrong. A little bit about the history there. We started, like I played around with a little project I called BabyKit.

You know, it was back at the time where everything had to be a kit. This BabyKit thing was supposed to be like an experiment how we think that it should work. And then Upstart came along and then I put that aside because we actually believed that Upstart would be the big and great future. However, over a year or so, and after talking, seeing Scott a couple of times. And I actually, I buy Lenart here where they actually bought into Upstart as a solution because they shipped it with Red Hat Enterprise Linux 6. Like their premier product, they shipped it. They actually backed it up by putting it to work, sure.

The Plumbers Conf, we've eventually figured out that Scott's probably not giving us what we want. And we believe that inherently it had the wrong design. The reason what exactly is broken there is basically whatStat does is you have a language how you can express how events happen and what's supposed to be done that way. What Systemly does instead is you express relations between things. Now,

one is a lot more flexible and powerful than the other, because in one thing you basically have the administrator figure out what should happen, or the developer, what should happen in which case. You write that down, and then AppStart will just execute it. However, in the system design, where you have

the dependency tree, basically you just say these are the dependencies, and then the system will figure out what to do at what place. So what Lenart is considering a feature, some of the critics of system D actually say is a disadvantage. And I'll give you my old crotchety sysadmin, everything should be conservative. I don't like a lot of changes perspective because, boy, do I have it sometimes.

Upstart – and you picked up on this on the pre-show. Upstart relies a lot more on the sysadmin or the developer giving it a good set of chain of events to follow, rules if you were, right? Whereas system D, you say okay well i want to start my sql and my sql is going to need network connectivity and network connectivity is going to need that these lower level things start up so here's what i want i want all these things the end result is i want my sql and then supposedly there's some there's some there's some black magic that happens there where system d can actually figure out dependencies in some cases and it is so there's advantages to that because you know systemsd can actually figure out dependencies in some cases and it is

so there's advantages to that because you know systems nowadays can be a lot of different things at different times you know in the case of a laptop it can be a wi-fi device and then that runs on battery and then 20 minutes later it can be hardwired into a gigabit connection with a you know ac power line so there's there are like and with maybe a second monitor attached like there's a lot of different things that can happen so you could see why maybe a more flexible system could be advantageous as computing starts to change, but I just wanted to

kind of draw attention to the fact here that these are major differences here in how these two things are structured. This has a couple of effects like for example in system D, if you want to have a minimal boot you want to just start that and that and that, it's a relatively easy thing to do,

because you simply look at the dependency graph, figure out what you want, pull in all the dependencies, and run it in the right order. This, of course, is systematically more difficult with Upstart, because the after rules in the first place have to express what actually happens on the system. This eventually becomes even more problematic. Like, for example, something very recent

is the rearrangement of C groups. I'm not sure if you guys have followed that. C groups are this thing how you can attach a label to a set of processes and then optionally set a couple of resource values to it. Like, for example, you can say Apache gets as much CPU time as MySQL regardless how many processes Apache starts and how few

MySQL will start. Cgroups, I think, are probably, in terms of server deployment, one of the most important things that Debian has to keep in consideration when they're making the choice. Not that they can't take advantage of our mumble room.

Not that they can't take advantage of C groups if they're not using Upstart. But what C groups are, as Lenard just said, they allow you to apply labeling to processes. And the advantage is if systemd starts a process, it can label that, say, Apache from start to finish. So the system always knows this group should have these rules applied to it. And if Apache has some CGI workers, and those spawn 30 sub threads, and you know, you normally in some cases would completely lose track of what's attached to what. But systemd, because it's been involved in all of that, will keep it all coherent, will keep it all labeled. Now this has big ramifications for like, say, I only want my SQL to use X amount of CPU so that way it can't clobber my whole system. Or I want Wayland to be able to do a whole bunch of stuff, but I need to make sure I keep track of all of its processes because if somebody ever compromises Wayland, I want to follow where that goes.

So there's a lot of different advantages to C groups. Not that you don't get them. You could still have them. You could even make upstart-C groups, which would integrate cgroup capabilities into upstart. But what you lose is systemd is going to sort of provide this API-level access to it. You're just going to kind of get it as a developer if the system uses systemd. Yeah. So with the resource handling with cgroups, you suddenly have this problem that if you have multiple services,

which basically should be handled the same way, you suddenly have this problem that if you have multiple services which are basically should be handled the same way, you need to propagate certain things between these services. For example, if you have Apache and you have MySQL running in the same slice, as we call it, then you need to make sure that either both of them have CPU limits

applied on neither of them. And this gets more complex as an entire dependency tree where you have to walk up the tree to the root and do enable and things like that. Now, with systemd, with the design, this is very, very easy because we have the dependency tree so you can just walk it.

Upstart, again, this dependency tree doesn't exist and you cannot just do that. I don't know. I figure the Upstart people will eventually figure that out themselves. But yeah, the summary of it all is, I can talk about that for a long, long time. I don't know. I figure the upstart people will eventually figure that out themselves. But yeah, the summary of it all is, I can talk about that for a long, long time. I figure if you haven't looked at the details of what upstart and system

do there, you probably can't really deal with much of what I say in this regard. But inherently, we believe the upstart design and the system design before it was inherently flawed at its core. It's not how we should do these things. We should have a dependency tree so that we can figure out all the right things.

And yeah, that's the reason why SystemD was invented and why we believe that it technically is the right thing. And none of the current contenders, other contenders can do that. I want to give one more example of why C-groups are pretty important and why having SystemD provide an APIi level access to them and and also just having it be cgroup aware is really important is there's a project that i have really been talking a lot about especially on quarter radio you and i did an episode about it on the linux action show called docker and docker it allows for a completely

isolated environment on your linux box all of vc you know sort of like bsd jails so for example in the show, I can't remember exactly. I think what I did is I had my arch machine installed and then in on my arch machine, I had a self-contained Ubuntu 1204 environment running inside a Docker instance. That Docker instance is using C groups to label all of the processes in that Docker instance as belonging to that Docker instance. So they are running completely and totally isolated from the rest of my system because they've all been labeled via cgroups.

And that allows for, example, Linode or GoDaddy or DigitalOcean, you could have a lot higher density on your Linux boxes. You could have maybe potentially hundreds of more clients on there because if you get rid of virtualization and you go right back down to the metal and you do process isolation, this allows for you to get completely isolated environments that are individually different operating systems

with different package configs and different versions of stuff all in the same box at metal speeds, right? And then when you combine that with something like SystemD, where it supports socket activation, you could be DigitalOcean and say, I've got 35 users on this server, but these are my low-end users, and only three of these systems are ever active at once on average. Well, in today's modern setup on a Linux box, all of those instances would just be running, taking CPU and memory, waiting for a connection. But with Docker and SystemD,

they could sit idle until that first visitor comes and then they immediately spin up and they serve the page. And then you can also apply resource limitations within there. So you can see in the server space how it could be massively beneficial. And I would be surprised if Debian didn't sort of weigh that in their decision process. Now, there are some criticisms of SystemD.

And one of them would be probably the one I've heard the most. I can't remember. Oh, it was Linux Today. I pulled this one. I love it too because whenever somebody, usually the person criticizing SystemD generally mistypes how you spell it. SystemD is all lowercase.

It's funny because whenever I see somebody criticizing it, they get that wrong. And in this case, the guy got it wrong. He says, systemd is one program that is a garbage dump of a lot of functionality which used to be provided with shell scripts and special tools. He says, and it consumes more memory doing nothing more than

system5 and it. And he says, by the way, the systemd journal is in binary format boom oh now that is true so gone are the days of text files for your logs and a lot of people hate this and i can understand now uh journal ctl supports like a really sweet api and a really sweet like parsing syntax and also supports exporting out to text so all of your existing shell scripts could just bang on that text if you wanted to but it also adds like really nice formatting like it it it it it clearly you know will enunciate like here was here was a reboot

executed by this user and then it'll do a nice break in the log and see all the events that happen from there um so you know it's it's like you lose some stuff, but then you also gain some stuff. And maybe if tools are built to take advantage of that API, it'd be a net benefit in the end. I think that's what's going to be needed. Yeah. Yeah. And maybe Red Hat and companies like that would do that as part of their enterprise offerings. I don't know. I think so. But OK, so his argument, though, that it's kind of a garbage dump of different functionality,

you know, I kind of... I would argue the same as far as going back to the old school way of it. It was a garbage dump of different functionality. I would argue the same as far as going back to the old school way of it. It was a mess. I mean, other platforms used to make fun of us for it. That's true. I mean, they did. And I'm not saying that's accurately – they're not accurately making fun of us or not. I'm just saying that was the perception. And so it's really going to – for me personally, I think systemd has a lot of benefits, and it's compelling.

Upstart has its way of doing this as well. All right, but let me ask you a question. I make this available to the Mumble Room as well. If the Unix philosophy of a lot of small individual tools working in concert has gotten us this far, SystemD is really taking all of this stuff that people didn't want to do. And let's be fair. Some of the stuff they're replacing was sort of abandoned and hasn't been getting worked on.

So they're just integrating that functionality in here. But it depends on things like cgroups, which are only available on Linux, which means if you write something that depends on SystemD and SystemD depends on Linux, then it's not portable to one of the BSDs. So there's a downside, which is probably one of the main downsides for Debian because they want to run on multiple kernels. So there's a downside. The second downside is you take a lot of stuff that used to be individually separated out, like individual componentized, and you bring it all into one conglomerate that manages everything.

Almost like making a lower level OS that sits above the kernel. Okay, so there's another downside to it. Then the third downside I see is this binary file format. That's a bit of a step back, I think, in some ways. Now, so those are sort of the negative aspects of SystemD. Because if you look at Upstart, it's deployed, it's tested. Those are things Debian loves. Okay, that's good stuff.

It's essentially fairly straightforward. You know, I've had some experience with Upstart. I've had some experience with SystemD. Either way, it seems like they both work. I actually kind of prefer the way SystemD works. I do for myself, but I think that from a Debian's perspective, based on the way they like things to work in general from their side of things, I think, and of course with other factors as well, I think they're probably reluctantly going to go with Upstart. I think it'll happen.

So I'm not saying upstart so cheeseburg's asking in the chat room upstart is not usable for bsds what i'm saying is if you build an application or if a project has dependency and there are already projects that have dependency on systemd sure uh like login d like time data d like locality hostname d u dev and more all are now projects that depend on systemd they will no longer work or function on a bsd system because systemd itself is dependent on the linux kernel so there is so that's one downside however so lenard's response to this is well i'm not worried about that i just want to make the best operating system possible

and see and and systemd does have a bit of a home turf advantage. Number one, it's used as an API by more and more projects, like I just mentioned. It takes advantage of C groups, you know, the control groups in Linux kernel, which is pretty important. And they've, maybe to court Debian, I'm not sure, have implemented some Debian designs into SystemD, like how the host name gets set and other things like this, that they looked over at Debian and said, you know what, they do that better than the Fedora project does it. We'll do it that way in systemd. And that might be a way to court it. But the main advantage that systemd has, well, it actually comes out in this clip from the

upstart talk at the recent Debian conference. I'd like to, you know, systemd had their promo talk yesterday. This is upstart's promo talk. So I think there are other things that I want to talk about. The most significant obstacle, it seems to me, from a Debian point of view, is the contributor agreement. I noticed you didn't mention that at all. Is there any possibility of movement on that?

Because that's going to be a big problem for Debian. I think I'm speaking for many people here. Wow. of movement on that because that's going to be a big problem for Debian. I think I'm speaking for many people here. Wow. Always about the licenses with you, Ian. What is it with Debian and licenses? I don't understand. You bloody well ought to. Sorry? You bloody well ought to by now.

I didn't see you on Debian Legal helping out. That's not the place the decisions get made, is it? No. So to make sure everybody understands the status of Upstart, Canonical has a policy of requiring a contributor license agreement, which is not the same thing as copyright assignment, but we do require, in terms of upstream contributions, anything that's going to be contributed upstream to Upstart

does have to be granted. A contributor license agreement has to be signed, giving Canonical, as the copyright holder of the overall work, the right to, for instance, relicense. It provides for certain capabilities of making sure they can deal with any litigation that might come up, and various other aspects. It's not something that is actually negotiable in terms of whether Upstart, as an upstream, is going to change that license agreement.

This is not actually altogether different from lots of other projects that Debian ships. Of course, we can compare and contrast with the FSF's position, particularly in that FSF is, of course, a nonprofit dedicated to free software, whereas Canonical is a corporation. And so I recognize that people are not necessarily okay with this, but I will point out there's lots of other people in the ecosystem doing this. The FSFs promise not to take it non-free. I'm sorry, but we are running out of time, so I suggest we continue this discussion in the hallway track.

The FSFs contributor agreement, the copyright license, the FSF legally, bindingly promise not to take anything that you sign over to them in that way non-free. Canonical specifically intend to be able to do that. We've heard Mark Shutterworth tell us that that's the purpose and that this is a thing that companies like Canonical should be allowed to do. like Canonical should be allowed to do. So I think he was speaking in a general philosophical sense that the COLA, that companies, that it is something that companies should be allowed to do. Well, it's a thing that the Canonical Contributor Agreement allows Canonical to do.

It does. And when Mark, who is in charge of Canonical, tells us he thinks companies should be allowed to do it, when the FSF tell us, not only do they not think it's a good idea, they also promise in a legally binding, not to. So that's a very clear distinction. Okay, so submit your patches for upstart.

Don't sign the CLA and we'll see what happens. You try. Well, submit them to the Debian maintainer. So we're out of time. Thanks. Please thank Steve and James for their presentation,

and we'll continue the discussion later. I'm pretty sure of it. Yeah, I'm pretty sure of it too. So the CLA is probably going to be a pretty big sticking point for Debbie now. I don't know what can be done there, actually. That's a great point, yeah. All right. For reference, the people who were talking there, the guy in the audience was, I think, Ian Jackson.

Well, wait, wait. For reference, who are you? So I'm Popey, Alan Pope. I work for Canonical as an engineering manager, not a developer. Right. And I've seen some of both of those videos. And I think the guy who was asking the question from the audience is Ian Jackson jackson who's an ex canonical employee and is a debian developer the guy answering the questions was steve langercheck who is one of the upstart maintainers and is a canonical employee and

has been a debian developer longer than he's been an ubuntu contributor or canonical employee poppy do you know is he on the technical committee of debbing because that's been one of the criticisms is that a couple of canonical employees are sitting on this technical steering committee. And if they vote, I think it's like three of them are canonical employees. Two, I think. Two? Okay, two, yeah.

So Steve Langecek and Colin Watson, both of which. I think Colin's actually the longest, after Mark Shuttler, he's the longest serving employee of canonical now. And before that, he was a Debian developer and he still is a Debian developer, contributes heavily to Debian, as does Steve Langecek. And yes, they're both on the technical committee.

They both have a vote as to which way this goes. And yes, it would be incredibly surprising if either of them decided to vote for SystemD. However, that said, I know if you read the mailing list thread on Debbie and Devil, which is very long, Colin has actually said, you know, if I thought that I was going to get any pressure from my employer, I would pull myself out. Recuse myself, yeah. Exactly. So I don't question either of those guys' integrity at all.

So here we sit now, though, kind of at a spot where you can see there's multiple sides to the argument. It's not quite a black and white situation. And Debian finds themselves in kind of a sticky spot where they have Red Hat and a lot of the other Linux on one side. But they also have some good arguments for Upstart, for one. I mean, not just because of their connections to the Ubuntu distribution, but also because they want to support K free BSD and Hurd and all these kinds of things, right? Like, they just want to be packaging all this stuff up and don't necessarily want to be tied to just the Linux kernel.

But my question about that is, like, really? I mean, at this point, it seems like that doesn't, I don't know, does that really hold water to you? No, my response to that, if I could interject for a moment, is why don't they use systemd for the Linux kernel

and then stick to what they have for everything else? Yeah. I don't know how hard, that would probably be a huge pain in the ass though. Yeah. That'd be way too much maintenance for them. What they want to do is completely,

like, it is quite a big task to do. To switch? Yes, to switch. I think that, well, me, personally, I'd prefer them to do the work to support SystemD because they don't, no one personally, I'd prefer them to do the work to support SystemD because they don't... No one knows.

People are looking at it from an init point of view. Only an init point of view. SystemD is an abstract scheduler with before and pre and post dependency management. It is a scheduler in its core, which init is a simple instance of a scheduling problem, or cron is a scheduling problem. That's what systemd is. Well, Android still does use initd, and it just uses bash scripts. I understand it needs to be changed, but I don't know why it needs to be changed. Right, Gen 2 has OpenRC.

Yeah, it's the same thing. They all do their init however they want. But that's why this SystemD is so invasive, because it does cover a lot of ground. It can be very powerful and useful. I agree. I don't know. But okay, so invasive or appropriate? Because first of all, I think what Lenart and his team have tackled are honestly some plumbing stuff that until Upstart came along, nobody was addressing.

And SystemD went beyond what Upstart accomplished. So I think, first of all, I feel like it's a modernization of the Linux core, and I do find it to be a little foreign. It feels a little foreign, but at the same time, it also seems like without some of these features, we lose a certain competitive edge. seems like without some of these features we lose a certain competitive edge i think it's fair to say a few things um have in modern times uh come across as foreign like for example pulse audio which is also a lennart offering which you know i think i think we can all pretty much agree is a

fantastic thing now and at the time i think ubuntu was one of the first distros to to put it in and you know we we put it in the distro and uh put our put our backing to it and yeah we got some flack for that and uh and so did leonard um oh yeah i mean back when it came out it was a pain in the butt to use or configure but some of these things are painful decisions painful switches. And that's why Debian are deliberating about it so hard. When we switched Ubuntu to Upstart, I remember you mentioned earlier about the work done to reduce the boot time on Ubuntu.

I remember being on a flight from London to America and Scott James Remnant was working on reducing boot time and he was sat there for most of the flight on a Dell Mini 9 and every so often he'd get up and come back further down the back of the plane and talk to us about how he'd like shaved off a few seconds here or a few milliseconds there or something and and you know this was leading edge stuff at the time um and we made that big switch and that was a painful switch to make you know and everyone keeps on us saying oh ubuntu is the only distro that uses uh upstart which you know is chroma west also uses it chroma west also uses it yeah i know that's the point i was just about to make and um thank you uh we're not clearly the only distro that uses it,

because every derivative distro of Ubuntu uses it as well, as does Chrome OS and RHEL 6, as Leonard said. And it is sort of also similar to the boot system that OS X uses, so it's sort of a cousin of it, so that's also widely deployed. I guess it feels like... Well, it clearly takes inspiration from OS X a lot with LaunchD, you know, SystemD, and ZeroConf, Avahi, you know.

Yeah, but I mean, why not, right? I mean, why not? Oh, yeah, I'm not complaining. It's inspired by good design. Yeah, and so I feel like that is kind of like one of the funny, like you can, I guess maybe part of what it comes from good design yeah and so i feel like that is kind of like one of the funny like you can i um uh i guess maybe part of what it comes from is upstart relies on a library that's titled literally

titled not invented here so maybe that's part of why people get on him for it because like it you know that is kind of a funny thing but what people forget is upstart actually was the first person well the first project to derive from like, I mean, like system D came afterwards. Right. So they're actually working on the principles of what Upstart started. And this is the, this is the basis after watching all of this, you know, I look at it and I think maybe system D is the superior answer because it is based on experiences with, with Upstart, which was good, right?

So it's like, okay, we went this far, now we want to go further. And I also think there is a big advantage to the Linux community as a whole if we standardize on some of this stuff. You know, you look at – go ahead. So it's okay if someone takes an existing project like, for example, Upstart, and makes something brand new to make it better, but it's not okay if someone takes

something existing like, I don't know, Wayland and makes something new like MIR. Right, but... Listen to that! I got a monkey suit riding on the whole thing, so let's go. All right, hold on, hold on. I could use three different

distros with three different init systems and still use the applications I want. The display manager, if that causes compatibility issues, where am I going to be left then? This is the big thing, right? Especially if Debian buys in on SystemD, then it is essentially all Linux will be – well, with the exceptions of Slackware and Gen2 and other things. But all Linux will be using well, with exceptions of Slackware and Gen 2 and other things, but all Linux will be using Systemd, right? So it's sort of, but in the case of Wayland and Mirror, I think it's apples and oranges there because it's… Yeah, I mean, I could use Ubuntu with Upstart. I could use OpenSUSE with Systemd.

I could use Debian with System 5 and still use all the applications that I need. But if the display servers were different and had different callings, what am I going to be stuck with? It was a somewhat facetious comparison. It got brought up. I have to mention this, and this is going to be very, very quick. The thing is, Linux in general

has less than 1% of the desktop market. When companies like AMD, NVIDIA, Intel, when they have to make drivers to make this software or these display drivers work with a hardware it costs money they're not gonna at the moment they do it because there's only one display drive and it's xorg and that's it but when it comes to when it's when when it's decision between Wayland and MIR, they're not going to fund a project to support both display drivers. Dragging it back on topic to SystemD, saying that the switch to MIR will cost companies money, do you think the switch to SystemD will cost nobody any money?

Right. That is actually a good point because for companies that have massive deployments and with custom init scripts and things like that, this is going to be a big transition. I'm actually curious, actually. I mean, does like a, you know, initd or system or upstart or systemd, does that have as much effect as like a display driver? Not necessarily because there are compatibility layers which allow you to run the old init scripts in a lot of cases. It's basically backwards compatible with initd? It can be, but that's questionable. So the official word is that systemd is quote-unquote 95% compatible with system 5 init scripts. I doubt that.

No, that's about true. Really? I doubt that. No, that's about true. Really? And I'm kind of arguing this from a user end perspective where when it comes to the display server discussion, I'm worried whether my applications are going to work or not. I don't usually have that worry when it comes to the init script. Well, the packager will take care of that. So the package – so, okay.

So I want to zoom out for a second because this is another benefit of system B being widely deployed is this makes it a lot easier for packagers, right? Because let's take BitTorrent Sync, right? They just make a BitTorrent Sync, here's your basic systemd startup file, and then the Arch user repo guy and the guy that's making the PPA for Ubuntu, or well, I guess Fedora, the guy's making the package for Fedora, they just have to tweak it specific for the details of their specific distro, but otherwise they can just take directly from upstream the init script for systemd. So it kind of like, it sort of makes that process easier

for those guys. Not only that, but from that standpoint, systemd is a whole lot better documented than upstart when it comes to that type of thing. Oh yes. Well, okay. I think you guys are shortchanging the benefits of deployment in the field and how many people that end up working with that and then document stuff about it and blog about it and know how to work with it. Because the advantages to that in IT are huge. I mean, that's why Windows gets such a big deployment.

So Upstart definitely has the leg up in terms of on the ground, already out deployed, people working with it and know how to use it. And I can't wait. So does SystemDd doesn't it systemd will get that advantage once it's rolled out into newer versions most things are transitioning to systemd as well yeah that's true well fine fine fine there's also one thing for the

encoding for the binary logging everything is encoded anyway you have a character set which is utf-8 or iso whatnot or little indian and big indian when it comes to the cpu it's all just the encodings even your logs have to be deciphered by their encoding okay they have they can export it out to yes exactly text exactly just yeah go ahead just say something real quick? Yeah, go ahead. I just want to say that I've really had no problems with Upstart. I run Ubuntu on my server.

Upstart's been pretty good. It's really easy to just type in the commands, but it's also easy to do that with SystemD. The thing that I think is going to happen is Upstart's kind of going to be left in the dust because SystemD is more of the – I believe it came out of the Fedora project. Is that correct?

Yeah, Fedora, yeah. Yeah, yeah. So usually like Pulse Audio came out of there as well as I think, right? A lot of things, Network Manager, LibVert, a lot of things that can matter, yeah. as I think, right? A lot of things, network manager, libvert, a lot of things that can matter, yeah. So I think that overall this is going to become the standard in the long term. I think Upstart is just what people are using right now,

and like I said, I don't have a problem with it. It's just that I just don't know, you know, once SystemD evolves and becomes more stable and all this stuff and gets more features and everything like that, down the road, I think that people are going to be switching to SystemD simply because it's going to be where everyone's going. That could well happen. I mean, for us, we've got 1204 LTS out, and we've got however many years left on that because it's got five years of support. And then there's all the servers out there that have Ubuntu on, all the cloud deployments that are Ubuntu-based that have Upstart on them. I would hate to see those bit rot and not get any love,

but the fact is that with the next release of ubuntu is going to be another lts so there's going to be another five years of maintenance of upstart i think we should be able to choose our poison i don't know how hard it is on the technical side well i mean you know what i'm saying i think it's pretty hard to swap those out uh yeah well the interesting thing about upstart like uh if you're on about like uh like drive derivatives so we're talking about like a linux mint and uh zon and all these other ones i mean they ship with upstart but the thing is if like uh the default desktop environment is actually gnome free um has a parent because it's part of dependencies. It does actually use system D.

So upstart kicks in and then it starts up system D. Well, it's not very efficient, but that's like, so if you want to use a different decimal environment that's not Unity, and all these other decimal environments are actually using system D, basically, in principle, it starts up, uses upstart, and then you need to use a parent. Yeah, good point.

Good point about the GNOME point, too, because I think GNOME, like, I think the official desktop of Debian is still GNOME, right? And systemd is becoming a dependency. GNOME is becoming dependent on systemd, and that's only continuing, right? That's starting in 3.10 and not impossible to avoid now. But in starting in 3.12, it looks like it could be. Literally, a dependency of GNOME 3.12 could be System D. Yeah, well, if you use the PPA to install GNOME 3.10, it actually installs dependencies for System D. being source dependencies for SystemD.

Now, if I could just make a prediction, I think Debian, because they tend to go along with the standards, will probably eventually go with SystemD. Yeah, that would be interesting to see. I'm not sure. A lot of people are always going with Upstart just because they didn't have an option back when it was created. I'm going to do a poll. SystemD wasn't around yet.

They didn't have a choice. Let's do a poll, and I'll embed it in the show notes too, so the people on the download audience. Yeah, I'll do a straw poll. Which way, let's see, Debian, let's see. Debian's choice, I'll call it Debian's choice, okay? And there's two options, Upstart or System D, and I'll let everybody vote.

I wanted to sort of pick up on something that Mollusk was talking about. This was sort of the point I was going to wrap up this topic, so then we'll read some email and then we'll run. But I think, you know, it's unavoidable that not only does SystemD have multiple distro momentum now, but Lenart and Remy and other people involved are extremely passionate about it.

And it is their main focus. It is all they are working on in a grand scheme of things. And it's like, whereas Upstart is a component of the Ubuntu distribution. It is a feature. It is an important feature. But it is only an aspect of that distribution, where SystemD is these guys' life. And they go out to every conference and they promote it.

They answer questions. They make posts about it. They live and breathe it. They are continually evolving it. They do frequent releases. They engage outside input and make changes based on that. Not always. Sometimes they don't. But it is sort of like what my argument has been against GNOME and Unity in a sense, too, is that Unity is awesome and it's a great feature. It's one of my favorite Linux desktops. But the reality is the GNOME projects and the KDE projects, all they do are their desktops. That's all they do. That is their only job where it's not just a feature or a package in their distribution. And when you have that kind of passion and momentum and focus and multi-distro and quote-unquote industry buy-in on something, it seems pretty risky to go against the tide.

Yeah, I think system D is definitely the future. I think upstart was an easy answer early on. But just like everything else in the Linux community over the years, we all make patches, we all make remedies, we all make things to make all make remedies. We all make, uh, we, we all make things to, to make the short term better. But I think, um, overall system D is good for Linux. Um, I, I really enjoy in its scripts and, and being able to make, make specific scripts to what I need it to do. Right. But overall, I think that systemD is going to help a lot

because it makes things a lot easier. And, you know, we're still in early days here. We're still in early days. Yeah, totally. Now, it's funny because I just said... Go ahead. I mean, that's the same. One thing I come across when I install,

I go for a very basic Debian installation, or if I want to install Arch, isn't one of those things when you're doing a very basic installation is during the installation you have the choice to set things up like that where you could choose the init manager or whatever? You can do that with Gen 2, but Gen 2

almost forces OpenRC on you. So I don't think it matters. So I guess at the end of the day if we move forward and we look at where things are going to take us, SystemD is still – it's not as fully mature as Upstart is. And I'll give you what I mean in that sense is in an Ubuntu box, if you really screw the pooch and you need to reboot into a rescue mode, that rescue mode you go into is – it's Upstart aware. That rescue mode you go into, it's upstart aware. It can look at your system, communicate to the upstart system, and pull out bits of information it needs to rescue your system.

That is a very nice, complete, user-end feature that works great for users and system admins and developers. It is a very handy feature. And it is the result of something that has multiple years of production where you build stuff around it. SystemD is not there yet. There's not tools that fully take advantage of the API. There's not things that expose certain functionality. Like SystemD has the ability to fully log everything that happens on your system if it's an EFI box. It doesn't require it, but if you have like an EFI BIOS and it wants to talk to SystemD, it can literally time how long it took for you to boot at the BIOS screen all the way up from the bootloader to the kernel to your desktop.

It can pull metrics and diagnose all of that. Like it is a full-end system, right? And that's really cool and fancy, but there's nothing yet really fully built to expose any of that to the end user. And so maybe… Well, the thing is system MD is widely used within the community. It's used by pretty much every distro. There's only two distros that don't use it. I mean, it's Ubuntu-based distros and Chrome OS.

Saying Ubuntu-based distros doesn't account for much. It accounts for a significant chunk of all the people who happen to run Linux. It's a massive number of people. Yeah, yeah. I think at the end of the day, it might... Yeah, but Red Hat's worth a few billion, so if they think that Upstart's not worth using...

Well, my dad's bigger than your dad. I don't think that really makes a difference. No, I know. I think the legal arguments can only go so far. There's a 1.3 billion company, so... It doesn't matter about the price. I think what makes a fair point, though, is that it is hard sometimes to totally wrap our head around the fact that it is way bigger than I think any of us fully realize.

Because at least my metrics tell me that. And they tell me that Ubuntu still has more market share than the other three distros next to it combined. I just wanted to point out that at the know, at the end of the day, especially with Debian considering this, I think it's really just going to come down to two choices. Either they're going to find a way to work out the CLA and go with Ubuntu's thing,

or they're just going to go and go to their next choice down and just look at SystemD. At the end of the day i still want init scripts now my question is if they were to go with uh system d what are they going to do about let's say their variance based on the bsd or the herd kernels right that's the problem this is what no one's really touching on everybody's attaching everybody's going at this from an emotional point of view and that's that's wrong all this passionate nonsense aside, factually speaking, they want something established and stable. Which one is that? Upstairs has a leg there because Upstairs is deployed and it supports

multiple kernels. And it also adheres a little more closely to that overall Unix philosophy. Yep, exactly. And so that's the whole point of it is that when you vacuum pack all the emotional hot air out of this and really look at it logically for what Debian's goals are, not what your goals are, but what Debian's goals are, what – as a project, I think that they're definitely, definitely going to go with what makes sense from a stability point of view. Now, five years down the road, they may completely 180 on that. But at this point, it will will be it will not be system d for debbie and i promise you and you also have to you have to even you know you even understandably got to figure maybe there's a little uh pressure indirectly or directly probably indirectly to have at least a very serious consideration as to how it affects ubuntu since it is they are a huge yeah you know uh they're based on bend it and based on – And that's just one quarter or even an eighth of the pie.

That's not even all of it, but it's a big chunk of the pie. Now, that being said, personally, on my own machines, I do prefer SystemD. I like what they're doing. I like where they're going. I think it's awesome. Yeah, I do too. I like it too. But again, taking my own emotion out of it and looking at it logically, there's no question Debian is not going to go with SystemD.

Do you think so? They're going to go with Upstart. I know that for a fact my i'm i i'm wrong on some things but i'm right on others i i my track record on the show has been reasonable so i'm going to say i'm excluding my uh uh monkey suit thing but but that put aside so right now the audience is voting in 82 percent matt think it's going to go system oh yeah no they're gonna they're they're they're voting with their parts with their heart matt With their heart, Matt, with their heart. Yeah, I don't have one. So I win, right?

I will embed this into the show notes. So those of you on the podcast, if you want to go vote, go to episode 12 of the Unplugged program and scroll down the show notes. We'll vote in there and we'll cover it on the follow-up. We'll see what happens. All right, Matt. Well, I thought maybe before we ran ran we could solve a problem for jim so uh jim wrote into the show and he's he's he's hunting for a replacement for ubuntu 1004 yeah 1004 he says i was wondering if you could give me some advice on choosing a distro to replace my ubuntu 1004

installation since it's no longer supported i figured i should upgrade however i'm not a big fan of unity i wonder how he knows that uh He says, I've been fat, dumb, and very happy with Ubuntu 10.04. I love that he's just been hanging out forever on this old version. He says, it's for me the perfect distro. I am hooked on having a top panel to which I add a ton of launchers. I love Nautilus. I like the long-term support from Canonical and a community support. I also like the wide range of software available for Ubuntu. Did I try? Okay. He said he tried Unity and I really don't like it. All right, so he tried it.

Okay, that's all we say. That's all we ask. It does not increase my productivity. I am using a desktop, not a smartphone or a Mac Wattabee. Wow, I didn't read this far. I didn't read this far. Maybe I need to give it more time. This is unfair to Unity.

After all, Unity is actually an impressive software, but why use it on a desktop? I have also tried Ubuntu 12.04 in fallback mode. However, a few things I use, a lot seems broken. Mint 13 Mate seems like the best fit for me. My hesitation is that Mint installs with a lot of extra stuff I don't want or need. Anyway, I'll likely go with Mint 13 Mate unless you guys give me a good suggestion. He said, by the way, he left Windows when Vista came out.

Started with Firefox, switched to Thunderbird OpenOffice, and then he made the jump to Linux after he'd made the transition to all that stuff. That's the recipe. That's how you do it, Jim. Good work, Jim. So, you know what? Mint Mate is probably a pretty good idea. That's where I would go. I'd say Mint Mate is fine. Even Cinnamon is another option.

What do you think about XFCE? What do you think about it? I would say probably Mate is going to be more his speed. XFCE is awesome, but it is different. So while visually it has a lot of similarities, there are some core differences. So I think that Mate is probably going to be more what he's looking for coming from where he's coming from. As far as the extra stuff that's installed, I, really, unless you have a really small hard drive, I don't think you care. It's not like anything's running in the background. He was a little more worried about future security vulnerabilities. But as long as you stick with the updates, what about – okay, let me throw this one at you.

Just a curveball because we haven't talked about it much. Sure. There's that – it's sort of been rebooted months ago, the Ubuntu GNOME edition, right? And the thing is, is starting in GNOME 3.8, they have the classic mode. I don't think they call it that. But it is a two-panel, like GNOME 2.0-ish setup, but it's based on GTK3. It takes advantage of the newer GNOME applications and things like that. So it's one foot in the new stuff, and it's like one foot in maybe your wang into the old stuff.

So it might – that Nome Ubuntu edition could be. But I think his instinct with Mint and Mate edition, like Matt said, is probably the way to go. I think it's the way to go. He's found – he's got a butt groove in his chair, and he's really happy with that groove. And he's not necessarily wanting the latest and greatest. He's just kind of forced into an upgrade. So I think that's where he's going to be happiest. So one last thing. Cheeseburg emailed into the show on the whole topic. He was one of the people that chimed in about the state of Linux news. And then

now he also joined us in the IRC. So he reminded me today. He was traveling. I think he was in Japan. And he saw a good old-fashioned print magazine called Point Linux. And he heard about it over there. I can't remember the details of it. Somebody emailed in about Point Linux. He heard about it over there. I can't remember the details of it. Somebody emailed him about Point Linux. Anyways, Cheeseburger reminds me of Point Linux. It's a distribution that aims to combine the power of Debian with the productivity of MATE or the GNOME 2 desktop. So it is essentially, it's Debian under the hood with MATE on top of it. And now the reason why I think Point Linux could be a good,

one nice thing about it is, what was he on? Ubuntu 10? 10.04, I think. 10.04. All right, so obviously Jim is not a man who likes to reinstall very often. And the problem with Mint is you've got to reload to go to the next release every time. True, true, true, true, true. Whereas Point Linux being based on Debian, you're going to get updates and you're just going to go to the next release every time. True, true, true, true, true. Whereas Point Linux being based on Debian,

you're going to get updates, and you're just going to roll with that for a long time. He just rolls. That's a really good point. I like that. Because of the fact that he likes a comfortable place to be, he has more control over those updates versus having to go full upgrade.

That's a good point. I haven't tried this one myself, but both Cheeseburg and we got an email about somebody who was traveling in Japan. I don't know if that was Cheeseburg, who also saw it deployed there and thought it was awesome. So there you have it. There you go. I think it'd be worth grabbing a pen drive and just DD that sucker up in there and just see how that runs.

Totally. All right, Matt. Well, as we prepare to – whoa. Hi there. Hi. Hello. As we prepare to wrap up, the music just kicks in. It's like I guess it's playing us out like we're on the Oscars here, Matt.

Right, yeah. Get your award. It's like, I guess it's playing us out like we're on the Oscars here, man. Get your award and get off stage, clown. Listen, we'd love to have you join us live for Linux Unplugged. You can go over to jblive.tv. We're live at 2 p.m. Pacific. And, of course, go to jupiterbroadcasting.com slash calendar because the time zone change, that daylight savings thing is about to hit us. So prepare yourselves over at jupiterbroadcasting.com slash calendar. While you're visiting us live, you can also get info to join our Mumble Room

so you can hang out and participate in the conversation. A big thanks to everybody who joined us live in the IRC, and a big thanks to the folks in the Mumble Room. We'd love to see you live here next week, but you can also email us. Just hit the contact link at the top of the Jupiter Broadcasting website. Hey, Matt, I got an OK from the lead dev of OwnCloud to join us on Sunday to talk about OwnCloud 6. Sweet. Good stuff.

I know, so I'll see you then, All right, buddy? All right. Sounds good. All right, everyone. Thank you so much for tuning this week's episode of Linux Unplugged. We'll see you right back here next week.