Linus Torvalds still wants Linux to take over the desktop
cio.com
cio.com
The commands for git are an example of that -> they don't make sense. They are very poorly designed and counterintuitive.
Linus I think has this belief that 'technology & complexity = power' when really, it's 'products that do useful things' that matter.
Making seamless experiences matters. Consumers really don't care how many techies think something is cool, they don't care about drivers, or open-source, they just want it to work.
Linux could absolutely provide the foundation for the desktop, but whatever is on top of it will have to be equally innovative and provide a lot of value, so much so that I'd argue that the high level OS layer will be 'the hard part' not the kernel.
Here's an example: "system, orangeFS, bringing the number of supported file systems to 35. (Remind me, how many file systems do Mac OS X and Windows support?)"
Really? Who cares. You only need to support ONE, the right one, and support it well. Maybe there is an advantage in supporting a few others, but not really. This is a perfect example of 'engineering' that doesn't necessarily ad up to any end-user value.
"Pretty is not my primary thing." and here it is. Linux conflates 'aesthetics' with 'design'. A very common failure. They are not the same thing, not remotely.
Is it conflating the two? Or is it saying "I don't care about aesthetics as long as it's well designed (= comfortable to use)?" Linus is famous for rants about how the Linux Desktop requires to much technical detail to know for the average user.
> Linux could absolutely provide the foundation for the desktop, but whatever is on top of it will have to be equally innovative and provide a lot of value, so much so that I'd argue that the high level OS layer will be 'the hard part' not the kernel.
Did anyone argue that the hard part would be the kernel?
I understand your frustrations, but to me it seems you are picking an odd target for them. Linus opinion and knowledge about UX doesn't matter very much, because he doesn't work on anything that an end-user should have to see.
I think it's maybe better to think of the kernel as something like a 'chipset' in terms of it's relevance to the desktop.
Linux? BSD? It's like AMD vs. Intel. The user-based OS is so abstracted it almost doesn't matter. It's a technical detail.
huh? Only thing i recall was him ranting at the Gnome/Freedesktop people over having to enter the root password to connect to a wifi hotspot (network manager and polkit not getting along).
[1] https://bugs.launchpad.net/ubuntu/+bug/1573454 [2] https://bugs.launchpad.net/ubuntu/+source/gnome-power-manage...
On the other hand, people get used to living with issues such as: 'High CPU use by taskhost.exe when Windows 8.1 user name contains "user"' [0]. Possibly because we've become so accustomed (slaves?) to a certain computing environment. Moreover, you can't even contribute a bug fix.
Ubuntu on Windows hasn't arrived as a service to Windows users. This is an acknowledgement of a growing threat (at least in a limited developer community).
--
That said, get rid of Unity and Ubuntu would suddenly so much better. Canonical might have damaged Linux lesser if it made Openbox the default WM.
I do not understand from where comes all this constant Unity bashing. I think that this is actually what creates the real damage.
I think that Unity is great user interface that takes the maximum out of the current distorted screens and I have used it for years as my main (and only) desktop.
It's not just Unity; anything even remotely disliked gets bashed (Gnome Shell has a history of getting criticised at every bit of news about it). I agree that it is poisonous behaviour that benefits no one, but I wonder what causes it.
1. forced reduced-functionality or questionable UI choices on a project 2. imported the buggy, unexpected behavior of previous Windows releases into the Linux world.
I also think it is a conflict inside a changing culture surrounding Linux.
Yes, the interface is good but it sucks the resources out of the system. Have you noticed how much CPU and memory does `compiz` consume?
> I have used it for years as my main (and only) desktop
Only if you tried using anything else!
Yes, Unity has it problems, for example some processes start to run amok time to time, but I do not notice that compiz is one of them. Considering that I usually ran my Windows installations to the ground, Ubuntu has stand the test of time for me.
It loads faster, but it lacks about half the keyboard shortcuts I use in Ubuntu.
ElementaryOS lacks even more. No menus in the applications (like Chrome) with lots of wasted screen state, and a broken launcher.
We have tried, and Unity was still superior for doing stuff.
I do think that ubuntu could be better , and that with canonical heavy focus on convergence devices and servers, the desktop side is not improving as much as i hoped.
Can you point to an option that reliably does?
There is almost always a different distro that handles a specific issue better than the distro a given user is troubleshooting. If it were always the same distro, we would be in good shape...
The desktop feels currently like a neglected area. OSX gets more iOSified with every release, whatever Microsoft does gets booed, regardless if it's good or bad.
Much functionality that used to be on the desktop has wandered into the browser. What desktop you're using isn't as important as it used to be.
I wonder if that pendulum ever swings back.
The future of desktops I see is something like:
Chrome OS: 70% of computers for ordinary use
Windows: 20% of computers for gaming
OS X: 5% of computers for video editing
Linux: 5% of computers for development
A bunch of fucking nerds who don't understand consumers are never going to take over the desktop. I've watched Linux for 25 years, basically since the beginning. Stop telling people you're going to take over the desktop and figure out how to get another 1%, then repeat. I'd recommend starting with "putting all your wood behind one arrow".
I agree that that would make it more successfull... but what's the point in being successful if you do so by throwing out your unique benefits and cloning a product that already exists?
https://en.m.wikipedia.org/wiki/Open_Letter_to_Hobbyists
He had a business goal and effectively followed though getting the job done:
http://money.cnn.com/2015/04/05/technology/bill-gates-email-...
Desktop is a different story. It's pretty hard to make the same amount of money on desktop open-source software. So profit from it - and thus desktop software is developed by enthusiasts. Enthusiasm is good, but it results in worse quality of software, even though these are the best developers in the world developing it. As soon as there is way to get profit not from selling software - it can get good open-source implementation - e.g. browsers (incomes from ad) or game engines (UE4 for example is not FOSS, but source code is published because you just can't publish game and not pay). Yet no altherntive for Photoshop, or CAD, or Chessbase.
That was a bit messy, but I hope you got my point. You want to promote Linux on desktops - just invent a method to earn on OSS same amount of money es from selling software - the wave of open sourcing will be enormous.
EDIT: What I mean is that without earning money off of the end-users, developers will develop a product for themselves, not normal users. That's fine. That's FOSS developers's prerogative. But it won't make Linux on the desktop take off.
I am not sure if i agree that something like redox would make more sense than using linux. The most lucrative market for linux desktop right now would be dev/front-end and backend. Going to redox you would lose all that ecosystem. I think microsoft is realizing that which is why they are pushing the linux on windows thing.
But definitly, an desktop os based on something more modern than unix would be awesome
The commercial interest is there.
I don't see what has Redox to do with your point, or if you are simply trying to leverage this thread to promote it, but I guess any downvote comes from that angle.
So much of the kernel source code is driver code. The old approach was always to release the source code for your driver and they'll incorporate it in the major source tree.
WIFI- and GPU-drivers still are very much binary only and don't have source code released.
ADDED: it might also indeed fix one of the biggest problems with Linux on the desk/laptop. For example, in moving through Debian starting with lenny, skipping jessie (systemd) and moving to Ubuntu LTS 14.04, I've had both sound devices and my Wacom tablet stop working and start working again.
This was both severely annoying and expensive in the former case (had to buy a $$$ PCIe sound card to cover for inexpensive and utterly standard USB devices to use Skype (or any other VOIP system) and play music)). Still better than Windows, which I abandoned after XP, and still believe was the right move, but....
That's also the version that broke the Wacom support as I recall, it came back with wheezy or maybe Ubuntu Vivid's kernel.
Pulse Audio is pretty awful, but I've always been able to get it to do my bidding, modulo not being able to get the Ubuntu version I'm using to multiplex outputs (I can probably make that happen, it's not a Pulse Audio issue per se I'm pretty sure, but it's just easier to switch output from stereo system to headphones the few times I use Skype).
And, yeah, I can imagine how much of a zoo USB in the real world is....
The larger problem is the number of "winmodem" style devices. That do as little as possible in hardware, and everything else in the driver.
And releasing source code is one thing, but companies like AMD are releasing the interface details that allow others to produce and maintain the driver. The problem there has been third party licenses and patents, particularly as GPUs do more than apply shaders to triangles (video codec acceleration being a big one).
As a software developer, I'm fine with Linux,I think I can I live with a non-Windows system.
Of course the problem is I still play games on my laptop/PC :p
Pretty much sums up why it won't happen until someone starts imitating Steve Jobs and DICTATING how The Desktop should look like. And that means more pretty than functional.
But we have plenty of distros where the people behind it do just that?
Wow, this is such a difficult problem. Let's just get back to bragging about how many file systems Linux supports. "And how many file systems do Windows and Mac OS support?" "Yuck, yuck, yuck"
I used OSX around 10.4 for about four years, because it seemed like the best of both worlds: nice UX with an unix backend. That is, indeed, why OSX appealed to most geeks, right?
I became tired of OSX after a couple of years. What apple dictates might look good, but it's far from functional. It's eye candy with little substance. On the surface is pretty, but Finder under OSX was essentially a crapshot. Under Linux I could have virtual desktops, but not under OSX (that only arrived years later). Tweaking, even slightly, any default behavior is hard. That is also the reason why Gnome doesn't appeal to me. The basic OSX desktop doesn't offer anything beyond a current Gnome release. Things change if you need commercial programs, but the desktop itself is not inferior in both behavior, looks and general interaction. Gnome is, performance wise, slower, but the people that use gnome do not seem to notice.
Not to mention that what every developer does on OSX is install an external package manager, and by and large replace every single userland utility and library with an up-to-date variant from Brew, MacPorts, etc. This was a massive burden. Any linux distribution is far superior in that regard.
I don't think that Gnome today is far away from OSX. Sometimes OSX is smoother, sometimes Gnome actually is.
But, my desktop today is nowhere even remotely comparable to what gnome offers. I'm using a tiling window manager with pretty much any program customized to the point of being unrecognizable. It's incredibly efficient compared to a pretty desktop. I'm aggressively disabling eye candy and animations everywhere. I'm often using CLI versions of GUI programs because the actual user interface is superior. Some TUI programs have no rivals in GUI form.
The fact is, is that if you use your computer professionally and switch to linux, you start to understand what user interface efficiency is. At some point, eye candy becomes a second, third, fourth priority. This doesn't mean the program is any less usable though. I still value presentation. I'll take a nice UI any day over a crappy one, given the interaction is the same.
I've witnessed the same happening to many people that I've suggested linux. Many of those still use windows or mac, but with completely different habits.
> Tweaking, even slightly, any default behavior is hard.
I understand why they want to restrict tweaking - it makes support easier. Hide the tweaks in config files, make it clear that you're not going to support tweaked systems. But don't ban tweaks.
That highly depends on what you do with your computer professionally though. So it might be a fact for some, but completely false for others.
I'm using a tiling window manager with pretty much any program customized to the point of being unrecognizable.
I played that game for years but gave up afterwards. Maybe I was doing it wrong but I spent just too much time to my liking on it. Now I'm more in a mindest of 'if you can't build me software which I can use almost as-is, I'll look for something else'. Not always ideal either and sometimes there are no alternatives so you'r stuck with crap after all. And yeah maybe I'm not using my computer as efficiently as possible (looking at some collegues though, I'm a wizard compared to their 'let's use a mouse for everything and double-clik to make sure it hits'-style), but I'm getting work done.
I have my own list here, but I'm curious to hear yours to potentially add to my list. Thanks for the great post.
Very true. Before starting my current position, i was a Visual studio die hard.While i still think it's the best single IDE, one thing thing i notice using linux CLI and vim in particular how much faster is it to do thing. And i seem to remember better how to do stuff instead of being lost in menus .
I think the current trend in UX is toward making them more intuitive (which i think linux/CLI could read do a be job at), but sadly we are sacrificing efficiency in the process.
Nope. I left the Linux desktop for OSX not because of aesthetics but because of the freedom from frustration that OS X gave me. Design isn't about making things more pretty, it is about improving the way a system works where the human part of it meets the machine part of it.
Linux distros just haven't put in the man-hours to make it a functional desktop for many.
I am not sure the desktop needs conquering, we have more than one option and Linux won't make it unless it can get apps that people just have to have
Which is hard for it to get. Developing an app for desktop Linux means targeting a small group of people who are willing to endure frustration in order to run code whose source is available. Unless you are have some other revenue stream besides your customer, you can't make it more than a hobby.
The main problem, I'd say, is that there really isn't much of a reason to switch. People who buy a computer don't have any incentive to go through the with hassle.
If computers came with Ubuntu pre-installed I'm sure people would use those without a second thought.
However, the package management still needs to improve. One realizes that upon upgrading from one Ubuntu release to the other particularly if you have a bunch of software installed from PPAs. It would be unreasonable to expect a typical user to be able to fix the problems that it leads to.
While `apt-get` is great, you realize the limitations when you need to resort to `dpkg` when the package management is left in a broken state due to the occasional mess created upon removing a package.
I wish somebody from Canonical were looking at this thread.
Not really. I mean if you think about it, the original Macintosh looked like shit even compared to its predecessors; what was it, something like a nine inch screen in black and white, even greyscale it could only do with migraine-inducing dithering. But it could do desktop publishing when pretty much nothing else could, and Macintosh programs mostly used a consistent interface so you could figure out how to use them, and these functional advantages got people to buy the computer.
Linux Mint looks as pretty as Windows or Mac OS, and Ubuntu would if they'd change the weird purple color scheme, but what happens when a client sends you an Excel spreadsheet? What happens when you want to plug in a second monitor? What happens when you want to play a triple-A video game? What's the battery life like on a laptop? Those are functional issues that need to be solved.
WPS Office covers me on that area. I use spreadsheets every week.
> What happens when you want to plug in a second monitor?
It works. Beautifully. In fact, I got a Powerpoint presentation, opened it with WPS, and the presentation was shown in the second display, while showing the notes and the rest of the 'WPS Presentation' in the laptop display.
> What happens when you want to play a triple-A video game?
Now this depends more on how many triple-A video games get sold in Steam for Linux than anything else. The technology is already there. The market is still small.
> What's the battery life like on a laptop?
This is a serious issue, and it needs to be solved.
You can have aesthetically pleasing and functional at the same time - they aren't mutually exclusive. Making the interface attractive doesn't mean 'dumbing down' the interface either.
Ubuntu is arguably the most popular Linux Desktop distribution and their model of UX development is to have a dedicated UX team designing the Ubuntu desktop experience. Presumably they read feedback from the community of Ubuntu users but ultimately, they make the final call on UX decisions.
The model for open source code contributions just doesn't work for visual and interaction design. (Or if it does, I've yet to see a successful example that wasn't an exception rather than the rule).
When there are too many participants in the visual and interaction design of a program or OS, you end up with a project pulled in every direction and pleasing to no-one. But if you go the opposite route and limit design decisions to a dedicated UX team or sole designer, you end up generating resentment from contributors or users who feel their input is being ignored.
UX/UI work is very different that base dev. work, and maybe should be approach differently.
My other though would that maybe the talent pool of UX designer is just too small in the Open Source world, and that most OSS tend not to treat UX design as a first class requirement.
A: bulk discounts.
B: bundling
C: embrace extend extinguish
All these combine to shore up the Windows presence on the desktop.
2. Steam is still a proprietary platform, though I can see how their support would benefit all distros' gaming capabilities.
I am thankful that SteamOS is nothing more than a lab experiment. If that became popular, Valve will not pay any attention to us, regular Linux users.
I've also been using the Linux on the desktop since 1994 when Yggdrasil and Slackware got started, everyday as a developer.