The intuitive human interface
Usability is too often defined by familiarity. Critics talk of intuitive behaviour, but what we mean by intuitive behaviour is coloured by our familiarity with the recognised way of doing things.
What we call intuitive behaviour when talking about computer desktops is neither intuitive nor natural, but depends almost entirely on learned behaviour. 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.
Predictable behaviour and a minimum of choice is desirable for infrequent users who don't want or need to learn new behaviours. Familiarity dictates that once we learn a way of doing things it becomes intuitive. It follows that if a certain set of behaviours become second nature, we don't like change, because change necessitates a different behaviour.
So changes to the behaviour of interfaces don't always go down well, especially where they're done for the sake of change, or to "add value", as in the case of Microsoft Office 2007.
Microsoft's problem has been to find ways to persuade users to upgrade the Office suite for each new release although the new features, however attractive to the end user, seldom justify the upgrade. A new interface may or may not justify the upgrade although incompatibilities between data formats usually force the issue.
In search of the zeitgeist
Converts to the GNU/Linux desktop are probably more likely to welcome change. The varieties of choice on GNU/Linux, the release cycles of Ubuntu, and the excitement of distro-hopping, offer a new sense of adventure to computer users who have been trapped inside the Vista bloat.
Linux users are much more likely to download the latest beta release of any particular application just for the thrill of it. The upside is that the users are part of the review process and mistakes and errors are found and fixed more quickly. The downside is that many users "expect things to just work".
Change for the sake of change isn't desirable. If Linux is to break into commercial environments on a large scale, there has to be continuity and consistency. At the same time the desktop cannot afford to ossify, and the developers cannot be expected to ignore the potential that the technology offers. There are different users to please.
Compiz has been massively popular with certain classes of user. Its popularity has little to do with utility or functionality. Nobody needs wobbly windows or spinning cubes to get their work done, but aesthetic pleasure, which Compiz provides for some, is as much a part of the usability equation to some users as the transparency of the filesystem is to others.
The computer users of today are very different to the users of ten or twenty years ago, and are familiar with desktop metaphors, icons, buttons and ancilliary gadgets that were entirely alien not so long ago. Familiarity breeds contentment. Assessments of what is usable and what is not have to change accordingly.
I can't program
KDE 4.0 introduced Plasma and widgets, and determined to makes extensive use of the 'sweet spots' of the screen, the "edges and corners, which are easier to aim at". The GNOME developers are planning to base the GNOME 3.0 release around Zeitgeist and GNOME Shell. Both have encountered mixed reactions.
