GNOME hacker: Culture isn't holding desktop Linux back

If you're going to write for the Linux desktop, don't write for gnome use GTK or KDE. Either way it'll run on any of the major desktops.

That's what I commentarded on the last article...

"Consumers won't come to Linux until they're sure they'll have access to a wide range of applications, and developers won't write applications for Linux until they're sure there will be a large market to buy and use them."

"Open source bigwigs like Alan Cox, Miguel De Icaza, and Linus Torvalds can bicker all they want, says GNOME hacker Michael Meeks, but changing developer culture won't do a thing to attract more consumers to the Linux desktop."

Linux has failed to take a large share of home users for 4 reasons:

(1) There is no, or very little marketing. Your average user barely knows what Linux is Let alone the different brands, with the exception of maybe Ubuntu and Mint. And there is no reason for the average, non techie, home user to want to find out more. Linux may be far superior to other OSs, but who is telling people this?

(2) Being free is a double edged sword. Sure, the advantage is the user saves money. But then, buying something gives the user a sense of security, even if it is false. i.e. if they are allowed to sell it, then it must pass some quality control checks, they must have some customer support, there must be someone who takes some responsibility off my hands if something goes wrong. Yeah, that sense of security is often false, but it is still there. That's why even after average users learn about Open Office (for example), and see it, and see it has more or less the same features as MS Office, even after playing with it, liking it, when it comes down to the wire, they'll still buy Word.

(3) No advertising or marketing arms mean stupid names get released. Yes, Ubuntu means friendship in some language, or whatever. That's great. It doesn't sound very good though. By which I mean it doesn't roll off the tongue. Techies often don't understand this point, because they don't understand people. "Natty Narwhal"? Please. This naming convention is cute, and hip in a geeky kind of way. It is not marketable though. Google and Android get away with it, because they have a marketing arm, and Windows 8 aside, they have the market to themselves for non iPhone phones. Ubuntu may be a great word, with a great philosophy behind it, sharing and working together or whatever. But it is not marketable.

(4) Linux is still not as user friendly as Windows or OSX. It's getting there, but it can still be confounding for new users who ask a question about how to do something and then be told they need to run a bash script, install something with apt-get. Yeah, no big deal for techies, but for an average user, running a bash script, installing something with apt-get is about as annoying as you can get. Trying to configure PPPOE on Linux is like being shoved back into the 80s. Average users don't need this hassle. They want simple GUI solutions to their problems. Linux still relies to heavily on the command line.

Anyway that's my 2 cents.

Linux will never succeed, at least not commercially.

First, let's clear up the terrain so I'd like to be very clear on this: I'm a Linux fan, I made my teeth with Linux From Scratch (fine piece of work, by the way) and I despise Microsoft (I always spell their name properly) as much as I can without going into extremism.

Now getting to the real business, I have a bad news for anyone hoping and waiting for a Linux desktop success: it will never happen. The reason is very simple, nobody is taking Linux for what it is. The desktop computing device market right now is dominated by two powerful species. On one side we have Windows, a somewhat dull but sturdy and reliable workhorse across all industries and in all homes. On the other side we have MacOS, also powerful and reliable but stylish and elegant, destined to those who want to differentiate themselves from the Windows users mass. In order to occupy a relevant place between the two, Linux must bring something that the others don't have. It is not enough to be as good as Windows or MacOS.

Incidentally Linux had something to offer, it is something that has been built in it but there are tow major obstacles which locked it out of the enterprise and consumer market. Linux is unique, it is the only OS that gives you a PC that you can fully trust as your own and more important than that, a PC that fully trusts and obeys you as its owner and master. No other OS would even come close to that and this is something no software and/or hardware vendor will tolerate. The two obstacles I mentioned above are large masses of users who don't care about their digital rights and enterprises who freak out at the idea of consumer having a choice or any kind of rights.

The Windows (TM) of opportunity (pun intended) for Linux on the desktop is closing faster than ever. UEFI and secure booting will fire the last salvo into our beloved penguin.

By the way, Michael Meeks forgot to mention two more reasons for Linux desktop failure: Microsoft's total control of computer OEMs and retailers.

Win 98/EEEbuntu

I've still got some Win98 and Win2K machines, to which I can add and remove features and applications, they still get (old and office) automatic updates.

On the other hand, the EEEbuntu machines are now out of support, add/remove/update (apt-get) is now broken, pain in the butt reconfiguring them to point to alternate update servers etc.

It's not just the installation point that is painful. Linux is painful at end-of-life too.

Maybe in 15 years time, old Ubuntu machines will be as easy to maintain and use as old Win98 machines are now, but I haven't seen any sign of it yet. And -- for crying out loud, I'm comparing a 2 year old laptop to Win98. That is not a high hurdle.

...the forest through the trees

When I ran windblows I could always put the same program/disk back in the next machine and install the same software again even though it's a later version of windblows. This customer/home users like.

On Linux I have to find all new recompiles of every fscking program every time because the compiles for the last version of that Linux wont install and run on the next. Customer/home users hate that.

So until programs will install and run across multiple versions of a chosen flavor of Linux without a recompile for each I'm not expecting things to change.

The change won't come from Linux itself

The ONE reason I ever kept Windows in the past was games. If Valve ever makes most games available on Linux, then it becomes much more viable as a fast gaming platform.

The other major way to allow Linux is to have MS make an official port of Office to it. Then suddenly all the pseudo-tech business people start talking about Linux "coming of age" and the hipper ones give it a try. No-one (including Mozilla, apparently, with Thunderbird) cares about office tools on Linux that will interoperate easily in a corporate environment, so it's up to MS :)

My opinion

I am a Windows Sys Admin but I "try" to run some Linux machines at home to get used to it incase one day my job requires it.

My experience:

To much choice, what flavour do I pick, I had to install 4 or 5 distros before I settled on Mint, the average user wont do this when they can just install Windows. The enterprise most likely doesn't have the resource to go through all the testing and user acceptance to do this (I know we wouldn't) especially when we have a platform that already works.

Applications, Windows I download, double click and follow the wizard, easy. Linux I need to use apt-get (Or the software repository in Mint which goes some way to solving this), hope that the dependencies are correct, edit some "unknown" file to set the application up how I want it to work, if I want it to start with the system I have to jump through hoops (still cant get deluged to start at startup) rather than set a service to automatic or put a shortcut in startup. Then what if the app I want isnt in the repos for my distribution, I now have to run some commands to add them, I cant just go and download the app from a website and click to install.

Applications part 2, I have several times at work tried to get to grips with nagios, this generally entails installing nagios then finding it wont run, or the install completely buggering up the system, start again and the same happens. I guess this is version compatibility, fine but if its in the repo and available then it should "just work"

Mapping/mounting a drive, horrendous time to get this working, once you know the commands its fine but the first time I think it took me over an hour and I had to install packages to get smb working properly, alternatively I can click Map a Drive and fill in the GUI. To make it persistent in Linux was easy but no where near as easy as checking a box on the GUI in Windows.

I dread to think what setting up a printer would be like, thankfully I don't have one.

Enterprise level control, no i don't mean "just run a script on your clients", I want centralised control, organisation of my assets, I want constraints set set if I make a mistake it wont propagate to my client (With a script, you don't get this), delegation of control for certain task (I am sure this exists at the client level but what about the enterprise) basically I want a Linux version of AD with all the tools that comes with it.

Thats some of my experience

Linux isn't User friendly for the consumer or the enterprise, running command line to perform simple tasks may be quicker for the elite but it is a dark art to the end user. The Linux community seems to be incapable of seeing this point and believes because their way is faster or considered "correct" then anyone who wants a simple tool/GUI needs to read the manual and gtfo of their forum (I have experienced this nearly every time I have asked for help in a Linux forum, Windows I get a polite thread with helpful responses).

Microsoft provide both command line and GUI for all but the most high level of actions, the GUI is just a wrapper for the command line but it covers all the bases, people are scared of the command line because it allows for mistakes and is harder to remember than looking through control panel for an icon. In the Windows world everyone can do what they want to do with little effort i.e install a printer, new hardware, software, change settings etc in Linux they can't and saying search google/read the manual is not acceptable to perform a simple task.

No matter how good Linux gets unless it becomes more user friendly and the community start to accept that not everyone understands the OS and most just wants things to work, it will never take any significant portion of Microsofts empire.

That way way longer than i meant it to be.....

Linux on the desktop isn't going to happen?

The funny thing is, Linux has been a great desktop OS for over a decade now...

As I commented elsewhere on this article, Meeks is pretty much spot on; the thing I miss most when deploying Linux desktops in business is software that makes it easy to configure large numbers of user's desktops and applications at once, and to lock them that way.

Yes, it's partly possible already with the major desktop environments but it's also far from quick or easy to do and that's a shame.

De Icaza blames Torvalds over too much change?

Well that's the pot calling the kettle black!

Bah!

People calling for a corporate switch from MS Office to OO or one of it's forks need to understand that they are *not* compatible office suites, and switching may not be an option for offices that have heavilt developed MSOffice infrastructure.

You can do many of the easier jobs with the free alternatives to MSOffice, but formatting in Write is nowhere near the clever stuff you can do in Word 2003. Calc runs slower than Excel when heavily loaded with functions, uses a cursor-movement shorthand that is not derived from the Lotus 123 standard that everyone knows and loves, has a different function language punctuation standard and "imported" Excel workbooks lose macro functionality. Not disabled. Lost. I'm impressed with the database, but it ain't close to being as clever as MSAccess - a useful tool for small systems and for prototyping/proof-of-concept work of larger scope. Slide shows are of no interest to me other than in their simplest form so I cannot comment knowledgeably on that.

Fair Disclosure: I use OpenOffice for all my personal projects and MSOffice 2007 for all work-related stuff.

You cannot discuss Office Suite compatibility if you think people use MSOffice just for memos and balance sheets any more than you can discuss Windows Vs Linux is you don't understand the semantics of the Kernel Model and the difference between root and user accounts.

Clerk

Kill "real value"!

Michael Meeks and several other commenters on the Linux desktop adoption challenge "supposedly" due in part to problems with Gnome never addresse two critical aspects of this vital discussion.

One is that there is "another" Linux desktop choice in KDE, which is omitted as if this alternative does not exist. What role that plays in the "adoption" moniker is yet to be considered or known.

The second point concerns talk of the apparent difficult situation of Linux adoption here in USA specifically, and does not necessarily address or mention larger Linux deskstop adoptions outside the USA in places like Germany, several countries in Central and South America and Russia, for examples.

The conversations on this topic are too vague and chaotic.