Never Say Linux Has More Than 1.1% Market Share

Chrome OS is a Linux OS. It uses the Linux kernel and so far as I'm aware it even uses GNU components. A few simple scripts turn Chrome OS into a full on GNU/Linux desktop.

So the Linux market share then is actually 4.7%. And that's not counting Android and it's derivatives. Which are also Linux OS's. Why does the technology press constantly insist on misrepresenting Linux market share?

The above is a comment I left in the comments section of this article. I'm sick and tired of being told Linux only ever has 1.1% market share. It's just not true. But it seems to be an unwritten rule of the technology press.

The Linux kernel is used as a core component in many alternative operating systems. Some more popular and better known than others. But the fact remains. If Chrome OS has 3.6% market share and all other desktop Linux OS's have 1.1% market share. Then "Linux" has 4.7% market share.

But 4.7% sounds far too credible. Too popular. Especially as that doesn't seem to take account of the mobile market where Andoid/Linux dominates. It's far more comfortable for the Microsoft and Apple centric technology press to pretend Linux is still in the doldrums. Still only developed by geeks in their bedrooms.

It's just not the case. Chrome OS is Linux. Android is Linux. And GNU/Linux is of course far more popular than official figures show.

Rant over.

Users Don't Know What They Want

I was reading this article which quoted a comment by Richard Hughes. "User don't know what they want". Well excuse me for being to retarded to write an advanced GUI desktop on my own. But I can choose what to eat for breakfast in the morning. I manage to dress myself. Go out to work. Make it through the day and get home safely. All on my own.

When I bought my current PC. A Dell Dimension XPS 700. I did that on my own as well. I even paid for it. With a credit card. When I decided to make the move to GNU/Linux full time. I decided on my own to do it in stages. The first thing was selecting a GNU/Linux distribution. I experimented with Fedora and openSuSE. Then discovered Ubuntu. Now since I was using the hybrid hardware/software RAID array built into the 700 series system board the installation of a Linux based OS back then wasn't straightforward. Dmraid wasn't installed or configured by default on any of the distros I tried. And that's actually part of what drove my decision to go with Ubuntu.


Being stupid, I couldn't get dmraid working in either Fedora or openSuSE. Both distros use RPMs. Back then RPM hell was a term many people came to know. Basically there were dependency issues. Ubuntu with it's debs was better organised. No dependency issues. Although configuration was still a problem. But with a little research into the Ubuntu documentation I found what I needed to get it all up and running. Which meant I could now dual boot. Until that point I had been boot Linux from a USB drive.

So I'm not smart enough to write my own version of Gnome. I am smart enough though to know how to problem solve. How to make decisions. I know what works for me and what doesn't. Windows XP with all it's issues and problems wasn't working for me any more. So I made the sensible decision to find an alternative. What do you think I'll do now that Gnome doesn't do the things I want it to do? Maybe I'll find an alternative. It would seem to be the sensible thing to do.

When ever a software developer, a programmer makes a comment like "users don't know what they want". It's a clear sign something is going very wrong in that software project. Which reminds me of another comment I read once in a article. Linus Torvalds once said something like "he who writes the code gets to decide". Meaning ultimately programmers participating in free open source software projects are the ones who decide which features to include and which features to drop. Which is fine. Until that is the software you are writing targets a user group beyond other programmers. And that's what desktop environments do.

Kernel developers like Linus Torvalds have a certain luxury of rarely having to interact with desktop users. The concerns of the desktop user are rarely the concerns of the kernel developers. In a sense what kernel developers do is invisible to desktop users. So we desktop users don't complain very often when they drop or replace features. Desktop developers however don't have that luxury. As Canonical/Ubuntu and KDE found out. Gnome should be learning lessons from these two groups.

When KDE 4.x series was released there was uproar. The majority of users hated it. Some of there anger was squarely targeted at being hit with the unfamiliar. The very same issue we have to overcome to get people to use a GNU/Linux based OS in the first place. Anybody would think we'd know better. Right? A lot of the anger however was squarely down to the fact KDE had changed things too much. Now so far as I know, KDE developers aren't known for being polite about people who criticise their work. But after all the bitching was done they knuckled down and started fixing the things people were complaining about. As a result KDE is now a more pleasant desktop environment to use with some pretty cool features. Everybody's a winner!

Canonical has experienced similar anger spat in their direction for daring to force Unity on it's user base. Unity started life as the interface to Ubuntu Netbook Remix. And it's not hard to see why Canonical would think Unity would work well on a netbook. Small displays means you have to be economical with the display. Low powered CPUs meant not much was being done by way of multitasking. But on a desktop? These just aren't considerations that are any where near the top of the list of all the things to be considered. However Canonical would not be deterred. It rolled out Unity.

Most people say Unity was rolled out too soon. It wasn't finished. And indeed they say the same of Gnome Shell. Rather than bitching or insulting it's user base though, Canonical it seems would rather just make Unity better. 11.04 delivered a stale turd of a GUI. With 11.10 Unity was now running atop the new and improved Gnome 3.x. They fixed some of the annoyances. Made the dash useful. 12.04 will focus more on stability and polish. Basically KDE and Canonical listened to their user base. And because they listened they could fix the problems that were pissing people off.

Listening and understanding users is the most important thing a developer working in the user space can do. If your not writing software people want to use then your playing to an empty house. I'd hate to see the Gnome Foundation playing to an empty house. Gnome has been good to me over the years. It's been relatively hassle free. Simple and easy to use and configure. It's developers need to respect the users and listen to what they are telling them.

It would be a shame to see Gnome implode and be crushed by the weight of it's own foot print. There are plenty of alternatives. LXDE, XFCE, KDE and Unity to name but a few.

deviantArt Munro

There are people in the world who doubt the usefulness of the Internet and the World Wide Web. These same people are the people who are currently talking down the transition from the desktop computing model to the "cloud" model. They are also the same people who stand to lose the most from this transition.

I've read articles in the past that talk about the security concerns related to "cloud computing" or "fog computing". There's a lot of scaremongering going on in the industry and in popular media. And disasters like Microsoft's Sidekick disaster are routinely trotted out to add weight to the claims being made. All of those articles however ignore one simple fact. We already live in the "cloud". As I've said in various forums in the past, cloud computing is simply a return to the client server model where applications are hosted by a server and interacted with via a client. The server does the heavy lifting. The client provides the interface. This is how the web works. It's how it's always worked.

So why all the fuss? Well traditionally web applications have always been limited to the capabilities of the web browser. Primarily Microsoft's Internet Explorer which has been a ball and chain shackled to the ankles of the Internet since Microsoft forced it upon it's customers in Windows 95. But that has changed. Microsoft no longer dominates the web in a meaningful way. It's market share is falling continuously. New versions of IE cannibalise market share from older versions of IE while the over all market share for IE falls. Microsoft's competitors are stealing the march on the web. Apple, Google, Opera and Mozilla have all released HTML5 aware web browsers. And it's HTML5 that will set the web free.

HTML5 is what powers Munro from deviantArt. Munro it's self is nothing special. It's a painting application that allows you to publish directly to your deviantArt account. What is important is how Munro is delivered. It's free to use and is delivered to the user via the deviantArt web site. There's no plug-ins required. No installation required. No configuration required. No lengthy serial number to input for activation. It is essentially the ultimate plug 'n' play application. The only thing the user is required to do is figure out how to use it. Which isn't hard.

Munro isn't the first HTML5 application to pop up. There are quite a few out there now. Discreetly integrated into web sites. Users simply take them for granted without even noticing how painless the whole experience was. Because it was painless. This is how the web was meant to be. Free and easy to use. Simple and transparent. Accessible to all.

Ditching Ubuntu 64-bit

Since upgrading to Ubuntu 10.04 64-bit my PC became unstable. Most of the time it was just innconveinent minor things. The clock on the top panel wouldn't display properly. However occasionally the whole PC would freeze. Especially when under a heavy work load. Couple this with the fact that Adobe Flash and Adobe Air have never worked to satisfactory degree. I decided it was time to go back to a 32-bit world.

What I had to consider was what I was losing and what I was gaining. My PC has 8GB of RAM. Normally on a 32-bit system 3.75GB is as much as you get. So I'd potentially be losing half of my RAM. But that was it. None of the software I use is exclusively 64-bit. Some of it is however exclusively 32-bit and there are ways around the 4GB memory cap in 32-bit operating systems.

What I stood to gain was a simpler installation procedure for my 32-bit only software and a more stable operating system. I use the 32-bit version of Ubuntu 10.04 on my laptop with no problems. So I already knew it was more stable.

To get past the 4GB memory cap, I intended to install the server kernel which has PAE technology enabled by default. However I found this was totally unnecessary. The desktop kernel in Ubuntu 10.04 comes with PAE support.

In the end up I now have a more stable PC. A simpler installation procedure for my 32-bit software. I still get to use 7.9GB of the total 8GB of RAM in my PC. Which is actually an improvement over the 64-bit version of Ubuntu. The PC does however feel a little sluggish at times. Particularly when launching applications. But that's a small price to pay for simplicity and stability.