There are Linux users that have little use for productivity suites, and depend instead on a simple text editor, which has all the functions any one man or woman could need, and makes files that everyone can read. This method of communicating with the outside world is all too obvious for those of us who have learned their craft with a keyboard and a command line, and have never depended on the click of a mouse for access to a computer. Once upon a time Linux users prided themselves on lack of bloat and simplicity of use. But times have changed.

Since increasing numbers of Windows users have deserted their roots and infiltrated the citadel, Linux users have changed their expectations. We need an intuitive interface and a multitude of choices - Gnome or KDE, which do you prefer? The lighter, more innovative, options, Blackbox, Windowmaker or Xfce, have been relegated to relative obscurity in the race for the desktop, because the challenge has been to find something familiar to the Windows user, a desktop interface that the Windows user can recognise and use without a hitch.

Intuition is a word with many meanings, and what we mean by intuitive behaviour is coloured by our familiarity with the recognised way of doing things. It isn't intuitive to sit in front of computer and click a mouse. It isn't intuitive to send an email to the other side of the planet, and it isn't intuitive to tap out words on a keyboard, but all these behaviours become second nature because we know that this is the way it's done. As the anonymous coward once said, "the only intuitive interface known to man is the tit", and what we call intuitive behaviour is neither intuitive nor natural to most of us.

Thumb tacks
It is entirely learned behaviour to use a mouse or to type a text message with your thumbs. You only have to see your proverbial granny confront a computer for the first time to know that a button that says "Start" is not the obvious button to click when you want to exit - yet this is learned behaviour for those who learned their trade on Windows 9x. And the trouble with learned behaviour is that once you know how it's done, that is the way it is always done.

So, the massive effort that has gone into KDE and Gnome is predicated on the knowledge that to be accepted the Linux desktop first has to look and feel like Windows and/or the Mac, and that new users have to experience an "intuitive" response. Photoshop users who come to Linux and try the Gimp complain that its interface is clumsy when what they really mean is that it doesn't look and feel like Photoshop, so the Gimp developers have been working on creating a more usable interface that is closer to the style of Photoshop. It is arguable that neither solution is correct and that both depend on the evolution of learned behaviour. What we are really after is to give the greatest number of options with the least number of clicks and keystrokes - to get from A to B without a detour round the keyboard.

When we talk about usability, all too often we are talking about satisfying received opinion, and that our ambition has to be to create a desktop that is so close as to be imperceptibly different to the Windows or Mac desktop familiar to most users. This is a problem, because it inhibits innovation, and prevents the kind of lateral thinking that may come up with an entirely different approach. We shouldn't forget that the desktop as we know it, with its icons and mouse clicks, is almost entirely an invention of the Xerox Parc laboratories in the late seventies. Lateral thinking, daring to think differently, brought us a new way of doing things that is now the accepted way.

Caterpillar tracks
The judgements that are made against the Linux desktop are seldom based on a genuine technical assessment of the capabilities of the package, or the merits of performance, reliability or cost, but on the current way of doing things and the habitual reluctance of consumers to change - "We know a certain way of doing things and expect things to stay the way they are".

This attitude has its virtues. To be a trained user of a word processor on the employment market is taken to mean that you can use Word or Excel, and employees demand consistency - but there is also a flaw in this reasoning, because the assumption is that what is, is always right, and that that there is no room for change, and also assumes a learning curve that isn't there. OpenOffice satisfies the paradigms of Word and Excel - and requires very little adjustment to use in the real world.

This kind of conservatism is the biggest obstacle to Linux take up in the office, Reviews of office suites written for Linux are often predicated on their ability to translate Microsoft document formats - "Can I read my Word documents?", or their replication of existing features in Word or PowerPoint, rather than the different features and possibilities that they commonly provide.

Behind the facts
Conversely, the downturn in world markets and changes in licensing by the dominant vendors have encouraged some users to look beyond their conventional suppliers. The support of open source solutions by trusted hardware vendors such as IBM and Sun Microsystems has also encouraged change, and here and there, in Munich or at Peugeot, there have been large deployments.

Sooner or later, the benefits of cost, flexibility and quality will demand a more objective assessment, and faster uptake in the enterprise. The initial key to the future adoption of Linux on the desktop may, though, lie outside the traditional PC networks, to niche functions such as public access kiosks and other thin client solutions, where total cost of ownership (TCO), reliability and ease of management are at a premium. Linux desktops are flexible, configurable and manageable and, coupled with a thin client solution, are able to integrate with other desktop and server systems and make use of legacy equipment. The Linux desktop solution may often be deployed not at the expense of existing systems but, like many Linux solutions, emerge through the IT needs of start-ups and other organisations with limited resources.

A different universe
Critics like to pretend that free software lacks innovation. This assumption is given credibility on the desktop because we can't imagine a different universe until we have stopped playing catch up with Windows, and have caught a significant share of the Windows market on the desktop.

On the server side and on the developer platform, free and open source software is overflowing with imagination. On the desktop we are stumped by the expectations of users, who are still expecting their Linux experience to be a reflection of their experience on Windows (without the virus problems). Into the future we can look and hope at the likes of Looking Glass and Dasher, Compiz and Beryl, and think that maybe, just maybe, out of such projects will spring the ideas of those who dare to think differently, in an age when free software rules the desktop and the world is more open to change, and Linux gives an experience that is totally unlike Windows.

Richard Hillesley