David Hurst

It would appear that Microsoft is losing faith in Windows’ ability to cope with changes in modern computing, and are now working on a new project called “Midori”. I’m glad they realise this. Windows is archaic behemoth of an operating system, built on an unstable foundation to start with. Someone once told me that the rise of the PC and Microsoft was largely due to companies not wanting their staff to be “playing games” on their computers. At the time, the marketplace choice was plentiful: machines like the Amiga and Atari ST offered windowing, multi-tasking, operating systems with colour displays. Instead, in their ignorance, the company executives chose the IBM PC, and with it Microsoft DOS - a product Bill Gates bought from elsewhere and re-branded to fulfill the IBM contract. Whether this is the real reason the PC took off or not, the fact remains that the IBM PC and Microsoft DOS marched hand in hand to market supremacy. Many other platforms were faster, and many had better operating systems, but it was Windows that prevailed. Therefore the software companies wrote for Windows, further tieing everyone to the OS.

In today’s computing world, people are increasingly mobile, and often have more than one computer. Take me as an example: I have a desktop computer at home, a laptop that I use primarily for work, an ultra-mobile laptop I use when I don’t want to carry my big laptop around, and an iPhone, which is essentially another computer. I also work on a number of different machines at work. Suddenly, having an OS that is tied to a piece of hardware seems a bit of a misfit, and this is precisely what Microsoft are researching with Midori.

I’m not convinced that this is a new concept or amazing innovation by Microsoft, although they will no doubt claim such when/if Midori ever comes to market. There are other companies who have tried variations on this theme, but I think the solution doesn’t need inventing, a similar solution already exists: Linux. The beauty of Linux is that it separates out the components of the OS, and it doesn’t really care where those components are. In simple terms, there are three levels to the Linux OS: level one is the Kernel - the heart of the operating system itself; level two is the windows server - the code required to present high resolution images to the display, and the core windowing behaviour, and; level 3 is the window manager or GUI. In a standard Ubuntu installation you have a kernel based on the Debian Linux kernel, X11 as the window server and Gnome as the GUI. You could just as easily run KDE as your GUI. Or you could have a Red Hat kernel and run KDE or Gnome on that. It’s all interchangeable.

Here’s where it gets really interesting. The GUI does not need to be output on the same physical computer as the OS. You can use a basic terminal and network connection to work with your OS. (This is not the same as using Windows remotely via VNC or Remote Desktop, as in those cases the GUI is on the same computer and images of it are shunted through the network connection.) This opens up other possibilities, namely: supercomputers. Supercomputers that have thousands of processors and huge volumes of RAM, can be worked upon by large numbers of people at the same time. With the progression of Internet bandwidth, we could rapidly see a scenario where we rent space on a supercomputer and then use this from any terminal we like. Likewise, the software we need could be rented also. Then, your personal hardware becomes largely irrelevant.

Some would argue that this already exists with Windows using Citrix etc., and to an extent it does, but you still need a computer with an OS on it in order to use it, which sort of defeats the object.

So, I welcome the Midori project and the fact that at least one part of Microsoft is actually thinking about what people will really need from computers as we move forward. The main issue for Microsoft to overcome is that 80% of its income comes from PCs that are shipped with Windows pre-installed. Currently they have a captive audience and a steady revenue stream that requires a comparitively low advertising budget. They are consistently losing customers to Linux and Mac as it is, although despite the increase, these numbers are still a drop in the ocean. Midori and other similar concepts are likely a long way off from the mainstream yet as this form of computing requires something of a paradigm shift across the whole industry. In the meantime, Microsoft have a real task ahead of them to make their next version of Windows a success. Another (comparitive) failure like Vista will give a real boost to the OS migration. Ideally, they will build elements of Midori into the new OS, to start a phased changeover.

It’s not just Microsoft that need to think ahead like this, it’d be nice if Apple did too. I don’t want to pay Apple for their MobileMe syncing service (formerly .Mac) - I believe I should be able to sync my desktop, settings and select files between my iMac and my MacBook Pro. It’s a real annoyance when you try to keep both machines set up identically.

I guess the only downside to all of this is that your entire computing experience will hinge upon the quality of your connection. But perhaps a change to centralised computing would force the telecoms companies to up their game too, and that can only be good for computing as a whole.