Is your software usable? I don't mean in the sense that it’s not unusable – most software can be made to function if you are willing to take enough time and use enough swear words.

A "usable" program is one which does what you, the user, expects. As the hardware engineers are still working on a mind-reading module to be fitted as standard to all PCs, this is a difficult challenge because each user is different. The science of usability and the art of user-interface design are attempts to make programs which are more usable by more people.

Historically, good user-interface design has not been a hallmark of free software, and usability issues have caused particular controversy in the community. There are a number of reasons for this.

First, designing graphical user interfaces (GUIs) for ordinary people is a relatively new problem in free software, and in cases of conflict we are still working out effective arbitration mechanisms. The common way to solve an irreconcilable difference in software development is to add an option – make the code work one way or another, depending on some flag. This works fine for server or command-line software, but with GUIs all the developers try to get their 15 pixels of fame and add a button to turn on their features. This is not good for the overall design of the product – you end up with a bloated, inconsistent mess.

Secondly, user interface design has a strong aesthetic component. Free software hackers are adept at assessing the technical merits of a piece of code, and can usually come to an agreement using objective criteria hammered out over many years of software creation. However, even leaving aside the possibility that the aesthetic sense of a software engineer may not necessarily be representative of the population as a whole, judging the relative merits of two interface designs can be a frustratingly subjective process.

Thirdly, free software is a meritocracy – and the merit that counts has historically been that of implementing your own ideas. "Show me the code" is a common way to end a discussion where someone has proposed a new feature. Because user interfaces are designed on paper but implemented in code, this means that if plain usability input is brought to the table, the person bringing it can end up being labelled as being all mouth and no trousers – or ignored.

Fortunately, over the past couple of years things have been changing. The "15 pixels of fame" problem has been addressed in Firefox, for example, by making it relatively easy to write, package and distribute extensions. There is consequently less pressure to include niche features in the core product.

The development of Human Interface Guidelines and the rise of academic software usability research has fostered a growing consensus on the right ways and wrong ways to do interfaces. Large desktop projects now have their own HIGs and are working on making their software more consistent with itself and with what people have learned to expect. While some positions are still subjective, others can be supported with evidence.

Finally, the culture is evolving. The community has started to see non-code input as valuable and something to be appreciated. It's more common for developers to have the humility to accept that others may know more about the design of interfaces than they do, even if those interfaces are implemented in code. (This change has also benefited other side projects, such as documentation, which has historically been another weak area.)

So all is not doom and gloom in the world of free software usability. Project participants are waking up to the idea that it's not enough for software to have capabilities if users can't figure out how to activate them. Contributors of usability advice are given more respect than previously. And the interfaces are improving. One less barrier between us and world domination.

Gervase Markham works for the Mozilla Foundation, a non-profit organisation dedicated to promoting choice and innovation on the internet. His blog is Hacking For Christ

Explore Personal Tech

• InGear