Canonical turned a profit, back above 500 employees | Hacker News Reader

Canonical turned a profit, back above 500 employees

phoronix.com

143 pointsmarcodiego60 comments

I miss the days of exciting Ubuntu releases... the 6.06 release with Compiz... the 12.04 release, finally with mature unity... It is now full of boring releases, desktop software often is nowadays, and that is a good thing mostly because of the stability it brings and because it is finally maturing.

Most recently increases are probably mainly due to cloud services, WSL and partnerships like the Dell and Lenovo ones, but the feeling of dedicated 'contributors for a cause' is still there. WSL may convince a set of users to stay on windows instead of converting, but it makes Linux more known and more used and that will only make the number of users increase at first. Long term, it is hard to predict.

Linux on the desktop is now around 1-2% by some estimates. I don't think it will ever reach the market share of windows, I don't think it will even reach the market share of the 'previous windows release', but it may grow still. If it reaches something similar to MacOS, maybe 10% 15%, hardware vendors and service providers won't be able to ignore it anymore and that will bring a new cycle of adoption. After that, it will stabilize, but at what point is also not easy to predict.

As whole, FLOSS is getting good enough in most aspects and needs. It doesn't needs to be 'the best' or 'the favorite among professionals', amateurs doing something outnumber professionals in most areas. Inkscape got good enough, audacity (or what ever fork replaces it) got good enough, krita and blender got good enough even among professionals, Libreoffice is good enough if you don't have to handle MSO-born documents. The situation continues to improve and gimp 3.0 will be better, 3.2 will likely be good enough even for professional photographers. The landscape seems for video amateurs, animators musicians...

Around a decade ago, Linux on the desktop was adrift. Now, there are routes which will take some time to mature but seems like good bets: PipeWire will bring good enough latency like jack with the same features of PulseAudio for audio AND video; snaps took centralized software distribution, Flatpak took decentralized software distribution and AppImages took direct distribution... that alone made software distribution on linux much better; Wayland will fix remaining Xorg problems... Oomd will fix lockups when out of memory... Wherever there were serious problems, now there is a route.

So, things are going in the direction of 'good enough' soon for the general case and most niche cases too. For me, it has been good enough for more than a decade.

> Around a decade ago, Linux on the desktop was adrift.

If you wanted to ship a commercial Linux desktop application today (ie something that can print, takes advantage of hardware video acceleration, handles multi-monitor / multi-dpi setups, can be remotely managed across a company, is distro agnostic, erc...) could you do it?

I'm specifically thinking of native applications and not Electron or Java.

Why not?

I have a variety of pieces of commercial software on my debian workstation, including MS Teams. Most of them are snaps.

What is the specific problem you think might prevent that?

Teams is Electron, is it not?

No. Edge Webview

Not yet. WebView2 is not available on Mac+Linux yet (but that’s coming).

That's basically the same thing. It's still a web app (which makes sense).

Teams 1 is electron, teams 2 will apparently be based on edge webview...

Err ... looks like it is for now, though they are dropping it apparently

Most of the commercial apps I use appear to be electron or java, or at least have components in those (zoom, spotify, jetbrains tools, teams). Regardless of OS - I use them across MacOS, Linux and Windows. So I guess these seem to be the easiest ways to target multiple platforms. These programs can use acceleration, print, whatever AFAICT. Zoom happily uses multiple screens.

Steam also distributes native games which GPU-accelerate, as well as using stuff like proton for windows compatibility.

Again, what is it you think might be the particular problem?

Printing and 3D acceleration are supported by the OS through stable and cross-platform APIs (CUPS, OpenGL). I‘ve had the smoothest multi-DPI experience on Linux compared with macOS and Windows (using Wayland, things are apparently a bit more complicated on X) and don‘t think the apps need to do anything that special, as even the most niche apps using Gtk and Qt just work.

Remote management would either have to be integrated into the application or a third party management solution (not that different from any other OS, maybe except Windows with its group policies).

Conclusion: Yeah. Fairly easily.

> Printing and 3D acceleration are supported by the OS through stable and cross-platform APIs (CUPS, OpenGL).

Is CUPS (and OpenGL) really available on Windows without bundling additional libraries and/or services?

Well Canonical provide commercial support for Blender on Ubuntu.

Not sure there's any reason other commercial software companies can't support their own software or partner up with Canonical to offer similar support.

You didn't mention this but playing games is also in the"good enough" zone now. Steams Proton is absolutely amazing and you can play almost any single player games with zero problems and most multiplayer games with no anti cheat!

This and the recent Steam Deck news makes me think we might have a few million Linux gamers soon. That might even be enough to tempt software support from the traditional vendors like Adobe.

I kinda doubt many Deck owners are planning to do creative work on it, though then again it does have the power of a mid-level laptop …

For me personally, the fact that it's basically a portable gaming handheld raspberry pi is enough for me to consider buying it.

Maybe. I think it will solve the bootstrapping problem of all games being Linux compatible. The demand for proper software might not be far behind.

The Steam Deck is as much Linux as Android is. The reason it will be a better gaming platform is because the users that will buy it will be rich, whereas that is not true generally of Android. This is just my perspective, but it should also inform you why it has no impact on software vendors to make professional software for Android, which makes hundreds of millions of Linux gamers, because an iPad too is a much more expensive device despite lower market share and hence worthwhile to target.

Android may be a poor analogy, as it runs the Linux kernel, but uses a very different userland. The Steamdeck, on the other hand, runs mainline Arch with (I believe) a KDE-based DE that comes pre-configured to boot directly to Steam.

I would challenge the assertion that Android hasn't had an impact on software support for Linux.

For example, most game engines and hundreds of pieces of middleware are generally supporting Linux now (site note: middleware was IME the worst part about porting games to Linux, only second to the graphics stack), even if just for Android games. Vulcan rendering is widely supported for a similar reason.

You are right about userland though, and what excites me is this might be the first mainstream device with a traditional Linux desktop setup. Even if people aren't going to be running Photoshop, they will run Discord, and they will likely be hungry for other software after that.

No? The Steam Deck runs an Arch derivative with a KDE plasma desktop. It basically just defaults to Steam Big Picture mode. You can alt-tab out of a game and open the Linux version of Chrome in your KDE desktop environment and you can see just that in the IGN hands on video.

"almost"

Looked at two games I play, AC Valhalla and Hitman 3, they don't work.

Proton "works" but does not replace Windows and never will.

Never? Proton has already replaced Windows for some and I don‘t think it’s going to become any weaker. Valve keeps pushing and making pretty good progress: Even games with anti cheat software are expected to work some time soon.

The Steam console was too early, but the Steam Deck might give Linux gaming a good push (even though I expect many users will install Windows, ha) and shows how committed Valve is to gaming on Linux.

> Even games with anti cheat software are expected to work some time soon.

I'm extremely excited about this, but there is a very sad reality check: anti-cheat vendors have pointed out that a hackable kernel breaks their model. They will most likely aim to detect and block WINE execution.

I mean of course not all games work! But for me personally literally every single game I wanted to play works! Everyone's experience is different. I mostly play single player games.

https://www.protondb.com/

If you change it to single player you can see 81% of the top 1000 games are compatible and 90% of the top 100 are as well.

If you don't do single player its still around 76% for top 1000. Top 10 is 50%

Also just noticed that hitman 2 (Three isn't listed for some reason) does work!

I think the problem is that Hitman 3 is only on the epic store, while Hitman 1 and 2 are on steam.

I love the ProtonDb comments but I am disgusted by the misleading tagging. Like a title marked Platinum will not work out of the box (need various workarounds like copying dlls around, custom flags) , Gold titles will often require custom Proton versions so is IMO an advanced feature, so who the hell is pushing this misleading labels and numbers is IMO making us the Linux users a bunch of dudes that we trick people to believe that things just work.

If you play AAA games that are pretty new you're experience will definitely suffer. If you play anything older than 2015-ish and doesn't require another software center to play a game (ubisoft or origin) then it's usually pretty good.

There's a reason they don't work: Both use incredibly invasive always online DRM. (Despite being single player games!) Even on Windows people have trouble getting both working.

I played Hitman 2 on Proton without issue, I highly doubt that Hitman 3 doesn't work for some reason.

Your speculation is less informative then someone's evidence.

Here's some evidence then:

https://github.com/derrod/legendary/wiki/Hitman-3-Guide

https://software.kaminata.net/linux-wine-games/hitman-3-linu...

https://youtu.be/luo_xMfOtAU

Weird, AC Valhalla works better for me on Linux than Windows l. Performance is better.

>games with no anti cheat!

Games that don't require kernel-level anti-cheat software.

Not sure which problem needs fixing there. Shudders

Well, cheaters arguably reduce the enjoyment of people who don't cheat. So it seems something that would be nice to fix.

And it's basically impossible to fix perfectly. Having kernel co-operate with the anti-cheating tool seems to make the problem tractable.

What is your alternative solution to this problem?

It will be interesting to see how Valve will address this issue, given they've stated interest in solving this in co-operation with the anti-cheat vendors—and it will be surprising if the solution won't involve kernel-level code.

Linux kernel license doesn't have the anti-tivoization clause and if they "just" provide a secure way from boot to anti-cheat code with hardware-assisted ability to hold keys, they could have a solution even if the system is completely open.

Obviously this is a very complex issue with many different (all valid) views. I feel the main reason against kernel level anti cheat is that its just not worth the risk! People will always find new ways to cheat with or without kernel level protection. However, it does seem to be working well for valorant! (I don't play it so no idea how accurate this article is)

https://valorantstrike.com/valorant-cheaters-how-good-is-val...

Nobody will be playing a competitive FPS with joysticks when the user also owns a PC.

Proton has been brilliant.

I’m really keen to see if the Steam Deck can help gain any ground with online multiplayer games or stupid drm stuff that prevents people from running their favourite AAA games on Linux. I’m not hugely optimistic that much progress can be made but I can dream, right?

> Linux on the desktop is now around 1-2% by some estimates. I don't think it will ever reach the market share of windows

20 years ago, Linux was an attractive way to be able to work in a Unix environment on the desktop. With the release of MacOS X 20 years ago, and the release of package managers like brew, people can get a lot of that from MacOS. For personal use, I still primarily use Linux on desktop and laptop - I use System76 Ubuntu, because I usually just want a working environment, and not to futz with it too much.

Tablets and especially phones have been a popular form factor in recent years, and Linux is ubiquitous on both.

I'm quite bullish on system76. But they have a big challenge ahead of them. They effectively have to compete with Apple for hardware and software quality to win the developer market.

But.. Apple has been shooting themselves in the foot on the software side by breaking basic toolchains for software development. And System76 has the ability to progressively grow a small user base without an all or nothing strategy.

As I understand System76 doesn’t actually do much with their own hardware but instead repackages ones made by Clevo, an OEM manufacturer whose models can be found under many different brands (https://en.m.wikipedia.org/wiki/Clevo).

That's a little unfair to S76. For one, they have been actively working with Clevo on the hardware. For two, most laptop OEMs don't own their designs fully, they just design a case and contract out the electronics to a third party in Taiwan.

And they develop their own firmware.

All the Clevo I have seen didn't have a great look and feel compared to most other brands. I can buy the same Clevo laptop from local sellers, for much cheaper.

I know it's such a small thing, but I refuse to buy a S76 laptop until they get rid of those silly numpads. I don't know a lot of developers that use them, and they just make the keyboard feel cramped and both the keyboard and trackpad feel off-center. I would pick up an Oryx Pro sans numpad in a heartbeat.

This is what System76 engineers say. That their motherboards are made by an ODM (Clevo), and Clevo sells similar motherboards to customers other than System76 - but that most other OEMs (other than Apple) have their motherboards made by ODMs, who then go off to sell similar motherboards to other customers.

In other words, most OEMs are selling machines built around "repackaged" ODM motherboards, so it is not a trait unique to System76.

Aye, they are moving to their own in house machines starting with the thelio.

You won’t win the developer market selling repackaged Clevos.

It's actually the other way around. Clevo resells off-label versions of the laptops system76 designs.

https://www.reddit.com/r/linuxhardware/comments/8uhfpo/why_b...

That's not what that says

What does it say then?

This is true. When OS X appeared, many lost interest in Linux. In so many places, only Apple laptops were seen in the laps of developers, for about a decade.

Nowadays, this has changed and Linux is interesting again. In fact, it can be argued it is the opposite. OSX has stagnated and it is no longer interesting.

Happy for them. Back in 2004/2005, Ubuntu was what got me into the Linux wonderland. It was an amazing time. Canonical's work on the desktop towards the end was so exciting for me personally. Unity as a concept even though just partially realized is still my favorite ever DE. To this day I install little tools on my mac that give me some Unity elements like super+{1,2..N} combo. Part of me still wishes they or someone else can come back and finish that vision.

The original Unity design docs were so good but I can't find them anymore. Does anyone have access to them? They detailed things like workspaces, "windicators", HUD etc. It was an exciting time even though there were a lot of flamewars because of Gnome fans not being happy with Canonical doing their own thing.

If I remember right Unity was based on user testing and not on some pretend designer preferences and weird philosophy.

I don't know if it was based on user-testing but yes it did involve a lot of user-group studies where they'd bring in random people with no prior Linux experience and tell them to use Unity

$138M revenue isn't half bad. I didn't see a product breakdown on a quick skim of the financial report. Anyone have an idea which products they are actually making money from?

My complete guess is extended support for otherwise-EOL ubuntu releases.

They get a cut for every Ubuntu image spun up on AWS. I imagine that's the bulk.

It is interesting to consider what happened to Browser as OS market. It is also interesting to consider with what happened to OLPC, one laptop per child, initiative. If all one does is use a device as a web browser then it doesn't matter what that device is. Browser games are gaining traction. As a Linux developer I use Linux as my primary OS and have Windows 10 loaded as a VirtualBox VM but the inverse would be easily doable if my brain fell out of my head and I decided to become a Windows developer. Windows is trying really hard to walk the tight line of making Linux run well for server applications natively in Windows but not for a Linux desktop or web browser. What I'm getting at here is that how users think about these things could easily break a market-changing usage in the next ten years.

I used Ubuntu for years, but switched back to Windows for easy gaming.

If I need a Linux for work, I run it in the cloud, which is usually good enough. Cloud gaming on the other hand is too sluggish for my taste.

I just discovered this and it looks promising : https://news.ycombinator.com/item?id=27870399

Thanks!

Reply on news.ycombinator.com