Here's the point you're missing. The overwhelming majority Linux users don't give a flying hoot about the average user or mass adoption. They just want a system that gets out of their way and let's them do their technical tasks. Linux succeeds and excels at this because it is precisely designed for that niche and curated by its end-users.

There are however people who do care about mass adoption, i.e. Canonical, "Linux Advocates". Generally they're either emotionally invested in Linux for some reason or they get a paycheck for their work. These people are the minority of Linux users.

Saying the culture doesn't care about design is absurd, it's the non-technical user we don't care about. Let's not conflate the two. Looking at a piece of software like XMonad or TMux it's hard to not appreciate the elegance in it's design. It's just different form of design that say iTunes or Photoshop.

Actually, I'd say the same of Windows, if not more so. I far prefer the Windows 7 interface to OSX, it's just stuff like the terminal/command prompt that really lets Windows down.

Design is an engineering discipline.

Design isn't engineering discipline. It is certainly a discipline and deserves respect but the reason software engineers fail at design so often is that it is different discipline with a different mindset.

Apple's work looks pretty-- because it is designed to function well.

Here I simply disagree. Apple's look pretty because they are designed well. I don't find that they function well in terms of usability - that I have to use the mouse for nearly everything is a usability nightmare. The way they've hidden file system is also a usability nightmare. But enough prettiness and responsiveness, Apple can still feel enjoyable to use without being more useful or even as useful as Linux or Windows.

In fact, you can see it in the rejection of apple's patents. This is why they think that apple patents are not original is because they reject that any engineering went into them. But that's just one example.

Wow! Holy leaps of logic, here we've got the disingenuous conflation part of the rant. For the sake of argument, the particular way iOS or OSX is designed might indeed be a work of genius but the patents Apple is hurling Samsung are not for that unique approach but rather are a disgusting grab of the right for anything vaguely like iOS or OSX, which would include really badly designed things.

Jeesh...

I thought it was designed well. So what does that make me then?

"if you want to run a linux desktop and don't value or care about design"

But I do care about design, it just sounds like you and I like different kinds of design.

Linux copied the Windows interaction paradigm (and everything else too - fvwm95/afterstep/olvwm/cde/etc) because the goal wasn't to create a new method of interaction and suffer from two adoption battles, but to appease people complaining about how Linux wasn't intuitive (after they'd already gone through one UI learning curve). Ubuntu will never ship with a default tiling window manager (certainly not the "best" design, but it's a novel design).

> based on their lack of design perspective that would let them see why things should work that way.

If a user desires for their operating environment to function in some way, they should be able to develop that customization, full stop. I use 24 virtual desktops with windows generally maximized, switching between them with [Alt+]F1-F12. Is this good design, especially for the average user? Hell no. But it fits my desires and is quick enough that I can't understand why people desire multiple monitors. The idea of someone saying my setup is 'wrong' from a design perspective is laughable, and only serves to illustrate the bounds of the field.

> This is why they think that apple patents are not original is because they reject that any engineering went into them.

Sigh, back to this again? Nope, what's being rejected is the idea that engineering an implementation of a concept should get you the rights to the entire concept through a "concept plus computer" patent. Where's the novel signal processing algorithms to clean up capacitive touchscreen input? Instead, we get descriptions of straightforward processes in terms of vague programming idioms.

The process of getting linux to boot, and then run usably -- with wireless networking and sound -- was too hard. If you got as far as being able to actually get work done, you claimed victory and used your machine as a continuing monument to your success.

In 2001, when I switched from Linux to OS X, the mere fact that I could go buy a machine, turn it on, open a terminal, and type 'ls' was enough to make me switch. It still took hours to get a portable machine working reliably with linux -- and it took 2 machines if you wanted wifi to work at the time. One to get wifi working, and another to look up the documentation and download whatever you needed.

Even as I'm back on Linux, I don't think of buying anything other than a ThinkPad, for the fear of returning to that worthlessly hellish place.

I believe that the lack of "strong-D" Design you refer to is because Linux's desktop is pretty much the ultimate committee-driven project. It's very democratic: great for preventing abuse, but rather bad when it comes to vision.

Most of us here know how difficult it is for one person in a company to drive an idea through without it being diluted; countless product companies have gone bust over the years because of committee-induced paralysis. It's much worse when you have a distributed group of volunteers who rarely, if ever, get to meet.

Maybe collaborative development is incapable of producing something as complex and difficult as a clean, well-designed desktop environment and API? Forking, competing spinoffs, factionalism are all useful things in some ways, but they're death to this kind of project.

If you want to make a system popular, start with the people who are actually using it. Linux is used by technical people, so why are they putting us off with silly animations and glossy layers that will break in the next upgrade? OSX is not easier to use either, it has many nonintuitive quirks but they stick to their design and improve it instead of overhauling it.

Windows is nothing but mature. Billions of hours running, automated bug reporting, millions of developers, nth generation tools.

The problem that linux seems to have, is that everyone wants everything and so everything gets thrown in with configuration options to change absolutely everything.

I don't think, however, that lack of good design - or even seeing the need for it, is isolated to linux - in a very real sense, I think most of the population is usually oblivious to it - they just know whether something is easy to use or not.

This "design," argument is a floppy, magical, hand-waving fish. I know what design is. I design software all the damn time. I am sensitive to interface and expectation. I want as little friction between my tool and the thing I am creating.

In that regard, as a software developer, OS X is a piss-poor experience. Using emacs on the thing is ripping years out of the cartilage in my fingers due to the masochistic meta-key layout that I've had to accustom myself to. It's terrible.

The eco-system of package management systems on OS X is infuriatingly shoddy. Between macports, homebrew, and fink -- none of them can seem to get it quite right. They have to compete with the user-land eco-system that Apple ships. Half the packages are broken. And I can never bloody well find any configuration files anywhere once it has installed something. It's ludicrous how much time I've wasted trying to figure out how in the seven hells you're supposed to get MySQL or Postgres to run on the damn thing (there are half a dozen ways at least by my reckoning and google's).

Their /design/ isn't the best design there is. They are designing for someone, but it surely isn't me. If you happen to fall into their mold then I'm sure you're as happy as a clam but don't think the entire world of people are just like you.

To you. Not to everyone.

> Look at the UIs of Linux... they didn't design one, they just copied windows.

Yes, because ratpoison looks just like Windows 95. Just like.

I'm not sure he is missing it, even if he doesn't name it explicitly on the post.

For one, he was always pro good design (and trying to get the good design parts of other systems), but also, when he laments the anarchy and lack of long term planning on the Linux side, he is essentially lamenting the lack of a central design authority.

> The second dimension to the problem is that no two Linux distributions agreed on which core components the system should use.

Linux on the desktop suffered from a lack of coherent, strategic vision, consistency and philosophy. Every engineer I know likes to do things a particular way. They also have a distorted view on the level of customization that people want and need.

I like OSX. Out of the box it's fine. That's what I want. I don't want to dick around with Windows managers or the like. Some do and that's fine but almost no one really does.

Whereas Windows and OSX can (and do) dictate a topdown vision for the desktop experience, Linux can't do this. Or maybe there's been no one with the drive, conviction and gravitas to pull it off? Who knows? Whatever the case, this really matters for a desktop experience.

I have two monitors on my Linux desktop. A month ago full screen on video stopped working. Or I guess I should say it moved to the center of the two screens so is unusable. I have no idea why. It could be an update gone awry. It could be corp-specific modifications. It could be anything. But the point is: I don't care what the problem is, I just want it to work. In this regard, both Windows and OSX just work. In many others too.

I can't describe to you how much torture it always seems to be to get anything desktop-related to work on Linux. I loathe it with a passion. I've long since given up any idea that Linux will ever get anywhere on the desktop. It won't. That takes a topdown approach, the kind that anarchies can't solve.

And clearly OS X is an extremely polished Unix and is going to appeal to the more UI-focused of the hacker set. And Miquel is definitely among the most UI-focused of the hacker set. He's also an inconsolate "platform fan". Much of his early work was chasing Microsoft products and technologies, of course; now he's an iPhone nut apparently, and that doesn't really surprise me.

But at the same time the Linux desktop was never really in the game. I use it (Gnome 3 currently) and prefer it. Lots of others do. For many, it really does just work better. But in a world where super-polished products are the norm, a hacker-focused suite of software isn't ever going to amount to more than a curiosity. (And again, I say this as someone who will likely never work in a Windows or OS X desktop.)

So in that light, I think the idea that the Linux desktop got "killed" is sort of missing the point. It's no more moribund now than it was before. It's more fractured in a sense, as the "Gnome" side of the previous desktop war has split into 3+ camps (Unity, Gnome 3 and Gnome2/Xfce, though there are other spliter camps like Mint/Cinnamon too). But it's here and it works, and it's not going anywhere. Try it!

(And, there are of course Linux-based systems that were built by someone controlling the whole experience, and those work really well. Android and ChromeOS come to mind, though those aren't really desktops per se.)

The other day, someone here was complaining about udev. It has ruined Linux forever, or something. I have a different experience: udev has made my life very easy. I have a rule for each device I care about, and that device is automatically made available at a fixed location when it is plugged in. For example, I have a rule that detects a microcontroller that is waiting to be programmed with avrdude in avr109 mode that symlinks the raw device (/dev/ttyUSB<whatever>) to /dev/avr109. I then have a script that waits for inotify to detect the symlink, and then call avrdude to program the microcontroller. A few lines of shell scripting (actually, it's in my Makefile), and I can just plug in a microcontroller, press the programming button on it, and everything just works. No screwing around with figuring out which device address it's assigned to. How do you do that in Windows?

http://www.jwz.org/doc/cadt.html

Also, part of what killed the Linux desktop was Miguel and his total lack of understanding of the unix philosophy which drove him to create abominations like BONOBO. D-Bus is not much better either.

That he fell in love with an iPhone goes to show he didn't fully appreciate the value of open source either.

We were just yesterday commenting with some friends in #cat-v how Evolution is one of the worst pieces of software ever created, and Evolution is supposedly considered by Miguel and co to be the epitome of the Linux desktop.

When i'm not on Linux I run OSX everywhere else (and IOS) because its unix-like (is) and because it works so well. I am sure Windows 7 and 8 are great, but I doubt they have gotten rid of c: or \ as path delimiter or any of the other nonsense that DOS introduced (copied from PIP) back in the dark ages. why should they, MSFT still runs DOS apps so they aren't going to change and choosing between OSX and Linux on a non-work desktop is a no-brainer, Netflix, Photoshop etc etc etc...

Let me give an example: a few months ago, a new version of Skype was announced for Linux. I was excited, since I used Skype 2 for Linux but then it stopped working for me and I couldn't be bothered to fix it. But if you go to the Skype for Linux download page, you will find a few downloads for specific distros, then some tar files which are, statistically speaking, guaranteed not to work.

Long story short(er), I still don't have Skype working on my desktop, because my distro isn't in the list, I can't get one of the other distro packages to work on my system, and of course none of the statically-linked binaries work.

(I could almost certainly get it to work if I was willing to install 32-bit binary support. But it's 2012. If your app requires me to install 32-bit binary support, I don't need your app that badly.)

Steam for Linux, recently announced by Valve, will run into the same problem. I suspect it will actually be Steam for Ubuntu and Debian, possibly with a version for Fedora, assuming you have the proper libraries installed and are using the right sound daemon and graphical environment.

But if big-name software comes out for Linux, hopefully distros will get in line. Do you want to be that distro which can't run Steam? Doesn't really matter if you think that OSSv4 is superior to ALSA and PulseAudio...if Steam requires the latter, you will toe the freaking line, or disappear into obsolescence.

One of the big thrusts of the Linux desktop wasn't simply dominance itself, but for it to simply not matter what you were using on the desktop. The Linux desktop fought to produce the first cracks in Windows hegemony a decade ago, but the final push came from the rebirth of Apple and the rise of the smartphone.

Today people happily do their normal productive or recreational tasks from a variety of computing environments: Windows, GNOME, Unity, KDE, OS X, iOS, Android, et al. Probably the majority of (Western) web users use at least one non-Windows internet device.

During the golden age of the Linux desktop everything seemed predicated on reaching exactly this point -- that you wouldn't need Windows, and then, by virtue of competing on a leveler playing field, the Linux desktop would ascend.

But the Linux desktops didn't "scate where the puck is going" -- or their attempts at such missed the mark. By the time we reached the era post-Windows dominance, the Linux desktops weren't positioned to take advantage of the new playing field dynamics. The rest of the industry isn't even all that concerned with the desktop wars anymore. It stopped mattering very much -- and ironically, that came around to bite the projects in the ass that first got the ball rolling.

This is completely missing the point - a statically compiled end-user binary should be compatible across all distributions of Linux, using the same version of the kernel or any newer version.

The only caveats to that are (a) hardware and (b) poorly-packaged software.

(A) is the fault of hardware manufacturers and is increasingly not an issue these days anyway; driver issues are becoming increasingly rare.

(B) is easy to solve for any open-source software, as it is the responsibility of the community for that distribution to provide the appropriate packaging. They prefer to do it themselves. And they're good at it - it gets done!

If you want to ship a closed-source binary on Linux, just make sure you don't dynamically link it against any libraries that you don't also ship with the binary. Problem solved.

Honestly, I can't remember one single instance ever where I have run into end-user software that will run on one distribution of Linux and not another, as long as that principle was followed.

A lot of applications break on newer versions of Mac OS X. That's why there are websites like http://roaringapps.com/apps:table

Also, there are a lot of "transitions" that Apple loves doing: PowerPC -> Intel. Java -> Objective-C. Carbon -> Cocoa. 32-bit > 64-bit. Access everything -> Sandbox.

See also Cocoa docs: "method X introduced in 10.5. Deprecated in 10.6".

I have a few devices that don't work in 10.8.

Basically, what I'm saying is that OS X is a bad example for backward compatibility. Windows is much better at this. Open source software is much better at this.

Now take the mobile world for example, Linux on mobile had been around for a decade but it never really took off until a huge company like Google decided to throw its billions of dollars and its great ingenuity at the task. Getting an OS to be popular is just incredibly difficult and it needs way more than just good driver support and/or good software. It needs marketing, talking to manufacturers, dedicated and well payed devs, designers, UI and UX professionals, sales, R&D and so on and so forth.

Focusing on the technicality of drivers and API is typical of us devs, but it has nothing to do with why Linux didn't take off on the desktop, sure Linux did fail because it couldn't do any or some of that well, but why couldn't it do any or some of that? Because it didn't have a huge and focused company pushing for it. How many popular desktop OS are there? Only 2, I think that's enough to show that it's incredibly hard to get into that market and that only a huge company can make it. Also, let's not forget that Windows was good enough and there was not much Linux could do to attract users, in fact this is still true and probably why even OS X is still at 5%: Windows is good enough and it's the de facto standard used by +90%. Having the best UI and UX in the world like OS X doesn't help that much either.

When I go home, I'll be using my personal laptop running linux. My wife and kids run a netbook with a linux desktop.

The linux desktop may be dead to Miguel, but it works just fine for me, a lot of other people in my life, and a lot of people in the world.

<shrug>

What killed the linux desktop? Drivers. Mostly graphics drivers but some others as well. Who cares if the UI isn't ideal if the damn thing can't sleep and wake up properly, or if it spazs out every time I plug in an external monitor.

However, no way in hell anyone will get me to switch to Mac OS. I am simply too enamored with having an environment that I can hack on if it strikes my fancy, as well as an environment that I can customize how I want it. Despite all its flaws, it still does focus follows mouse pretty well, and not having that would drive me batty.

Also, Apple is an 800 pound gorilla that has always been about Being In Control. The Samsung lawsuit wasn't anything new:

http://en.wikipedia.org/wiki/Apple_Computer,_Inc._v._Microso...

I just don't want to be part of that kind of walled garden.

I don't really see how the Linux desktop is dead. I've been running the same OS on this same laptop since 2007. The only upgrade I've added is an SSD and an extra gig of memory. It's still pretty speedy and I've never had any problems.

I use a Macbook Pro with OS X at work because that's just what I was issued by default. I hate it. I hate the over-reliance on the mouse, on gestures, the abundant and tedious animations; I hate the crappy ecosystem of package repositories and how most of the packages are broken or completely mess with the system; I hate never being able to find where any of the configuration files are or where something is installed; I hate the plethora of ways you can start and stop services; the confusing GUI; the masochistic meta-key layout; the awful full-screen support; and the complete lack of customization options.

I've had much better experiences with the Linux desktop for 95% of the things I do.

Now before some OS X fan-person decides to point out how woefully misguided and ignorant I am, my point is that there are different folks out there who want different things from their desktop experience. Apple gets to decide top-down what that experience is all the way down to the hardware. I prefer a little more flexibility. I like being able to swap out my own battery or adding a new memory module when I need one. I like being able to switch from a GUI desktop to a tiled window manager. Some folks don't -- there are Linux distros that hide as much of that as possible. Either way there are plenty of options and I think that's a good thing. Competition breeds innovation and even though I don't particularly like Unity I am glad to see people trying new things.

The Linux desktop isn't dead. It may just smell funny. You may switch to OS X and wonder why anyone could possibly want anything else. I just gave you a bunch of answers.

As well as Linux's presence in the data centre, witness the success of 'embedded' Linux: many TVs, routers, set top boxes and other bits of sealed-box electronics all run on it. It's broad in its scope because of the large team of divergent interests working on it, and it's able to support those systems because it's been well made as a direct result of that team's philosophy. Is it really so bad that the average Facebooker does't want to use it?

It really is very, very hard indeed to be all things to all men and no single system around today can make that claim. Linux has its place in the world of computing, just like Android, Windows, OSX and everything else.

Did you know that X11 is why we have shared libs (the UNIX version of "dll hell")? If not for having to run X11, shared libs really would not have been needed.

There are many window managers. Maybe too many. Too much choice for a noob. That selection or the pre-selections Linux distribution people make does not equate to "the" Linux Desktop. It equates someone else's configurations and choice of applications. It equates to having to fiddle with X11, whether you are just configuring it or developing programs to run in it. And that has always been extremely frustrating for too many people- constant tweaking; it never ends. This is like a brick wall to people who might want to try Linux, coming from Windows. You are inheriting a system that's been configured to someone else's preferences. (Same is true with Apple, but they have a knack for making things easy.)

I skipped Linux altogther and went from using Windows to using BSD. I've also been a Mac user. And BSD is way better than OSX, or any of the previous MacOS's for doing most everyday things: email, internet and secure web (ramdisk). Moreover it's flexible - you can shape into what you want - without this being an overwhelming task of undoing someone else's settings.

If you want a citation for the shared libs thing I will track it down, but honestly anyone can do it on their own. The historical research will do you good. Educate yourself.

All of the article's criticism of mainstream workstation distributions is accurate, of course. But it's important to note that those represent nowhere near the sum total of the linux user experience these days.

This developer culture DEFINES Linux. A fruit is either an apple or an orange. I couldn't have an OS with wonderful package management, developer tools, endless configurability AND a desktop Miguel de Icaza dreams of.

I'm on Linux now (GNU/Linux, maybe lump BSD in there too, I'm using "Linux"). I know plenty of users on Linux. I know plenty of users of Windows and OS X who run virtual Linux Desktop distributions for testing/development/security. I'm sure some of HN are running Linux.

Does Linux have the potential to enter the market as a third core option for desktop usage - not really. But why does it matter?

The problem with Linux is that there are too many choices. People who like technical choices and options trend toward Linux (needs citation).

John Q. ComputerUser isn't going to use Linux unless his geeky son or nephew installs it for him AND provides support. He can't get support anywhere else - because there are too many possibilities for it to be fiscally effective.

If/When something gets confusing or broken on Windows/OS X, you call JoeBob's SuperDuperPuter, and say it's broken. JoeBob asks, "What Windows version?" While he might need to poke and pry a bit to get the user to tell him he's running Millenium edition, once he gets that data, it's a pretty straightforward troubleshooting effort and fix.

If you call some mythical Computer Service group that actually supports Linux, and say your machine is broken, they would need to know a LOT more about your system just to figure what they need to do to start.

Distribution? Parent Distribution? Shell? Window Manager? Hardware? ...

I find generic computer service companies to be extremely expensive. To be able to provide even basic service for Linux in general, your techs need to be very familiar with more operating systems (emerge, apt, yum, zypper, pacman), and more core applications. Each service effort inherently takes longer. These factors pile up and everything becomes necessarily more expensive. It's downright impractical to support Linux generically. The support costs for one or two issues on Linux would far outweigh the cost of an upfront OS license and cheaper support for the end user.

Linux has (and will likely continue to have) a comfortable hold on the technically-capable DIY market. It may not be on track to step beyond that market in the desktop arena - but that certainly doesn't indicate it's time for a toe tag.

Having been a small-scale Mac developer for many years, that really made me chuckle. Not since OS X 10.2 did Apple release a major upgrade that didn't break my apps and make me struggle to push an update out as quickly as possible to fix all the things that Apple broke. Apple has heard of deprecation, but they don't seem to really grok the concept.

If I had been developing for Linux, I could have simply tested on pre-release versions of the distros I wanted to support and would have been ready when the new versions were released. On OS X I would have had to have paid a prohibitive fee for that privilege.

In any case, this article made me happy. You see, for so many years, I used a Mac, and everybody said "Apple is on its last legs; the Mac will be dead in a few years". Apple had to scramble to compete, and that drove them to provide such a good product. But I knew that situation might not last forever, and I was right. After seeing the turn that Apple had taken over the last few years, I switched to an Ubuntu laptop six months ago.

It's refreshing, once again, to be using an OS that people are calling "dead".

ADDED: jrockaway's comment, added while I was writing this, hits it just right: "I think the issue is that getting everything working requires a deep understanding of each component and the system as a whole." Which is what makes it so frustrating, even to very intelligent people who have other interests than computers in and of themselves.

I've been using Linux for the last decade and every year it gets better, more polished, more integrated, featuring a better design; I hear more & more people talking about it and using it. Linux is more alive than ever on the desktop!

Depending on your needs, Linux can make an exceptional desktop. Yes, true, it is not for _everyone_, but then again neither are Windows or MacosX.

Personally, I've been primarily a Mac user since the Mississippian superperiod, but I used an X-11 Windows(™) environment (on top of FreeBSD) for years at work. I don't miss it, even one iota, but I know plenty of smart people who prefer that sort of thing. De gustibus non disputandum est and all that.

I strongly disagree. It is losing exactly the sort of person that the author is: developers who, all else equal, would rather use Linux. But who eventually get tired of the BS and just want something that works and you can actually get software for. I have a lot of sympathy for that.

I write this from a Linux laptop, but that's more out of mulish stubbornness and 15 years of accumulated irritation with Steve Jobs and his dickish business practices.

The last time I got a new laptop I knew I didn't have time to screw around for days with X configuration files. And so I paid a vendor several hundred dollars over list to give me a laptop that JFW. And despite that the sound is still way too quiet. After a few time-boxed 2-hour excursions into whatever sound system they're using this week, I still can't fix it. I've given up.

The only legitimate reason I have for staying on desktop Linux is that I code for Linux servers, and I think it's impossible to really understand system performance if you're not running the same OS. But even that seems shaky to me; hardware keeps getting cheaper and developers keep getting more expensive, so it just doesn't matter as much.

One day some bright Linux spark is going to "innovate" again in a way that I'm expected to put up with their rough edges for 6 months (hello, Unity!) and I'm going to say fuck it and buy a Mac because I just don't have time to screw around right then. Or maybe I'll just want to watch a Netflix movie without hassle, or play the video game my pals are all talking about. And maybe by then it will be a fancy dock for my 8-core Android phone.

Overall, I agree with his point, except that I don't think Linux on the desktop is so much dying as cutting its own throat.

Oh, but it is. Because it has lost a lot of momentum that it had, momentum that was coming from the "we're gonna overtake MS and win over the Desktop" feeling prevalent at the time.

Heck, the guy behind GTK complained recently that he is just one man taking care of the project. The full GUI foundation for Gnome, and one that is far from feature complete at that, and it only has one guy working on it.

It has also lost a lot of people and companies associated with it at the time betting on this possibility [of it winning the desktop]. Most companies nowadays support Linux development only for the server stuff, but it wasn't always so. Miguel moved on, Ximian moved on, Hazel moved on, Rasterman moved on, etc etc. Even Adobe quit developing Flash for it.

And it also lost a lot of "alpha geeks" to OS X, which wasn't even commercially available at the time (1997-2001). This "desktop UNIX" come out of nowhere and it was it that did what Linux was supposed (and expected) to do, ie eat into MS market share. Well, even OS X didn't eat that much, but 15% is still a lot.

>He's also an inconsolate "platform fan". Much of his early work was chasing Microsoft products and technologies, of course; now he's an iPhone nut apparently, and that doesn't really surprise me.

You're saying it like it's a bad thing. Gnome (or Linux for that matter) are also platforms.

And to be frank, his early work was not "chasing Microsoft products and technologies". His early world was Midnight Commander, Gnome, Gnumeric, and Evolution.

He indeed like the component model (not the Window platform itself) when it was presented to him at a Microsoft visit, though (the story of this visit should be up there somewhere). But what's not to like about it? A good, clean, component model of sorts was also needed for FOSS, maybe still is.

The phase you describe, IIRC, was several years _later_, when he saw the .NET platform and got hooked.

Want to log in from a terminal and start X? Well, too bad there's some new infrastructure this month that makes sure you'll only get access to the sound device when you log in from a properly configured gdm. Want to modify the keyboard layout? Whoops, the xmodmap format that had been stable for a couple of decades now changed for the third time in a year. Want to add a tmpfs on /tmp to fstab? Well, too bad. Some implicit and undebuggable circular dependencies in systemd will make the system unbootable.

And the sad parts are that the problems this new infrastructure is supposed to solve never actually existed. "Great, now audio doesn't work at all. But if it worked, it would have full support for network transparency". It's easy to understand why the problem exists - creating something new tends to be more rewarding than working on the old stuff. It's much harder to see how to fix this madness.

Requiring everyone to be an expert will prevent Linux from being a dominant OS, because most people do not care enough to gain a deep understanding of any OS, so they'll pick one that's easier to learn the easy bits.

Agreed. That is two desktops and four laptops worth in a decade.

"I think the issue is that getting everything working requires a deep understanding of each component and the system as a whole. "

Not so sure. I'm an end user and tend to just shove the CD-ROM in and cross my fingers. Most 'stuff' works. My stuff is simpler than yours however (audio interface, cameras, keyboard (musical) controller)

[0] Last time I used ChromeOS it was still mostly in line with ordinary Linux (thought it might have been partially Gentoo based?); but a rather impractical one as it's less than simple to run [things that aren't Chrome]. Heard about Xorg getting replaced at some point, I'm unclear on if that actually happened.

But that's exactly one of the major problems he's complaining about!

Stupid jokes aside, I always had a feeling that the Linux desktop was actually gaining users in the last couple of years, especially since Ubuntu came along and truth be told it's still the best operating system I've been using but that question is largely open to taste anyway.

For the last couple o' weeks I had to use Windows 7 at work and oh my god, what a pain in the bottom it is! I've had a couple of looks at OSX and never liked it either. Then again my linux desktop isn't exactly standard either - but that's what I like about it: customizability. I'm the kind of guy who likes to tinker with the system until it's just about right for me. The last time I set up my laptop it took me 3 or 4 days to finish but it absolutely paid off. The result is something I find aesthetically pleasing and extremely usable.

Linux has a couple of show-off features and facts 'n' figures like running most of the Top 500 and second-to-none package management but it's real power is that it can very well be all things to all people. It's been said a million times but Linux gives the user a choice, and I honestly value that a million times more than any nice OSXesque UX or Windows-esque games support.

The catch is that I'm definitely not the average user who wants things to "just work" (although they rarely do).

That seems a weird way to phrase it. This is not a new thing, and generally drivers have gotten better over the years. Saying that drivers inhibited the growth of Linux makes sense, saying that they killed it -- not so much.

I don't actually see any problems with Linux. It's just not meant to be.

Your comment looks about as asinine to me as the as the parent comment does to you.

That really has nothing to do with the OS you're running. The OS isn't driving the DisplayPort<->HDMI protocol conversion in any way (your 'no-name' brand adapter is probably relying on DP++ and just passing through the HDMI signal anyway).

This is like saying your VGA cable is compatible with Macs with a VGA port. Of course it is.

The problem is that hardware often doesn't really conform to standards, even those it claims to.

One of the reasons device support in Linux (and other OSes which target a wide range of platforms) is difficult, is that it's necessary in practice to cope with and have workarounds for buggy and out-of-spec hardware and firmware. Just "coding to the standard" isn't good enough.

Such buggy hardware/firmware is rarely documented as such, and finding these problems and the appropriate way to handle them is painful and difficult work. In some cases the only practical way to figure out what actually works is to reverse-engineer what Windows does (the hardware manufacturers generally make sure that Windows works with their hardware, but rarely make such information public).

Apple's main goal is their own hardware, over which they obviously have a lot of control and information, so they really don't need to worry so much about this.

A device can follow the USB standards (and be labeled as such) while still requiring a special driver. The two are not mutually exclusive.