How the Linux desktop shoots itself in the foot.

This article is symptomatic of people thinking about "Linux Desktop" as a concrete thing (or project, or idea), while it isn't. It's an ecosystem of vendors (targeting different markets), projects, technologies, and you can't directly compare it with standalone companies like Microsoft or Apple that (mostly) have a consistent and focused product/marketing strategies.

The five ways outlined are:

Lack of vendor support - author mentiones SLED, but conveniently skips over the fact that Canonical also provides commercial support for its "arguably the most popular Linux desktop" desktop systems.

Lack of advertising/marketing - a bunch of video commercials (RH's Truth Happens, Novell's PC/Mac/Linux spoof), Ubuntu's CD shipping program and release jams/parties many of the distroes organise & support is not insignificant; it probably is pocket change to Microsoft's marketing machines, but the entire budgets of these companies are probably a pocket change to Microsoft's marketing machines.

Too much bad techie attitude - surprise, someone's a jerk on the internet - you can find l33t people in most online communities. More than average historically in linux communities, but the author does acknowledge there are good sources of online help nowadays (e.g. Ubuntu forums). Esp. the Ubuntu's code of conduct (and similar more or less formal codes in other communities) tries (and IMHO goes a long way to) minimise this problem.

Too much infigthing - goes on to recite a recent developer-oriented flamewar and concludes "no-one outside certain developer circles cares". So, why does the author care? Open development means not only is the source code open, all the joys and flames related to it are as well.

Not enough developer co-operation - again, if you compare the cooperation between different teams in a same company, and between teams in different independent (and possibly competing) companies or projects, the latter will not look as good. I agree with the basic premise, "not enough developer cooperation", because we can never have enough of that, but the arguments are just wrong. And, the given example - that fedora and ubuntu have different package systems - has nothing whatsoever to do with developer cooperation.

Two positive examples of cooperation:

1. KDE and GNOME using (or planning to use) same underlying technology: glib, webkit, telepathy, and the rest of the freedesktop.org stack, and actually co-hosting their main conferences last year.

2. All the distros working closely with their upstream projects (e.g. many of debian developers, ubuntu MOTUs and fedora developers actually are the upstream developers too).

(Note: I'm biased, as Linux has been my primary desktop for quite some time now and I'm actively involved in GNOME and freedesktop, but I've tried to look at this as impartially as possible. My biggest pet peeve is people really thinking about a "Linux desktop" is a single project).

That article is terrible. It's subjective and in many cases downright false. Linux suffers from interoperability problems? I have never come across a single Linux program that wouldn't run on my machine. Linux even runs a lot of Windows programs through Wine. "Bad techie attitude"? A few people are jerks and they tend to be the loudest. This is true in every field, not just Linux. Lack of vender support and advertising? One of the main appeals of Linux is that it isn't (primarily) a commercial product.

Wow... the comments are even worse. We even got a Nazi reference in!

In my opinion, the first two reasons are valid. But the author forgots the main reason: There simply is not compelling reason for the majority of people to switch to a Linux desktop.

Nearly all the good Open Source desktop apps run under Windows as well. But there are quite a lot of good Windows and Mac apps that won't run under Linux. There's no good replacement and Wine is too complicated to be a feasible solution.

Commercial desktop ISVs don't support Linux for 3 simple reasons: (1) There's no standard decentral installer. This is the _most important tool_ for desktop ISVs in their sales funnel and Linux has none (or rather: too many half-finished ones). (2) It's not a single coherent platform but several different ones. And it's not going to change without a decentral installer. (3) It's easier and more profitable to develop web and mobile applications today.

There's nothing the Linux community can do about (3). But (1) could be solved and in the long run, also (2). If these two problems for commercial ISVs could be solved, the supply of desktop applications could increase.

Then, desktop distribution could also make money by reselling these applications. Then, it would also be rationale to advertise.

But that's not going to happen, for one cannot rationally discuss the need of a decentral installer with many Linux users (or Linux distributions).

This is what happened when non-engineering people become part of engineering. Everything is looks wrong.

'90: "Dear X, your program does not work in my system, here is a patch."

'00: "WTF, i cannot run this program, what should i do ?"

There is nothing wrong in Linux as OS, in fact it's become more mature and usable since i first install it. Most problems that i read in forum now is about not working hardware or from people that does not know how to read manual or to lazy.

I glanced at the article long enough to know that I didn't need to read it all ... what does he mean by "the linux desktop". He links to Corel's 1999 offering, which is a distro using KDE. Does he mean distro? I noticed he mentions KDE and Gnome disagreements but he doesn't appear to say why devs disagreeing are ruining "the linux desktop", again whatever that is supposed to be.

He could have started by defining what he meant?!

Read the manual is actually sometimes the best response. I answer questions on the Inkscape "answers" page and "how do I erase" comes up about once a week, sometimes only a few lines above the previous person asking the question:

<rant & car analogy = bonus!>If I describe a car problem and someone on teh internetz tells me "it's your head gasket, it's blown mate" I'd expect to have to look up what that means. Yet with computers if someone says "it's your menu.lst entry, you need to add ACPI=off [or whatever]" people can't be arsed to read up on it.</rant>

I can't help but wonder if it's a mistake to try and push widespread adoption of Linux distros in the desktop space. It's possible there's insufficient money to be made in that space--Apple is a hardware company; Microsoft is a de-facto monopoly but rumor has it that they're going to have to struggle to recoup costs on their desktop OS in a world that cares less and less about the details of the host system.

Maybe the right solution for desktop adoption is "If you want to run Linux on your desktop machine, that's neat. But don't think you're going to be pushing Windows off of your neighbor's desktop."

The thing that bothers me in Ubuntu is this: Sometimes I want to do something with a desktop app on a file in /var/ or /bin/ the desktop app tells me "permission denied" or what have you. How the hell do I type "SUDO" and my password? Even if there is a way (which I doubt) if it isn't obvious to me, then it certainly will not be obvious to other potential users.

My read of the article describes a Linux desktop as a system that can be maintained without opening a terminal window. That entails a lot of GUI and a distribution that works well underneath it. Why exactly is that a good thing? Are all of these GUI users going to start donating to free projects? What is the upside?

Linux is a kernel, GNU/Linux is a server operating system. It's all based on UNIX and no amount of hacky patching can make it a coherent and integrated desktop system. It'll always have some shitty spot that ruins the experience somewhere.

Things like Haiku are the future.