It is a clichÃ© worth recalling when designing software. Ninety percent of users, it is said, use ten percent of the features in a software package. For that vast majority of “average users,” the other ninety percent of the features only add needless complexity that make the key ten harder to learn. These extra features simply serve to increase the TCO of software deployments and headaches at the helpdesk, not productivity of the business.
Mozilla is a classic example of a project that has seen the light in recent times and refocused on the all-important ten percent of features. The original Mozilla suite had too many problems for me to attempt to address here, and, for that matter, many of them do not even relate to the point of this piece. Regardless of that, one of the most serious ones was the unwieldy, bloated suite it came packaged as when the user only wanted a fast web browser. Mozilla simply did more than the average user wanted — so much more that many of those users were intimidated by it. Firefox on the other hand, is much leaner and has fewer, but more desirable features. Most readers already know of Firefox's powerful extension system that allows power users to meet their needs with Firefox, a much superior way to keep the ten percent of power users satisfied than including everything in the standard distribution.
Apple is likewise renowned for its simple, easy-to-use products that get the job done, but usually don't burden the ninety percenters with unnecessary features. Apple products typically have lots of features hidden below the surface for the power user to find, but leave the user interface unmarred by frequently unused functions that serve only to insure that Tylenol will always be in demand at the helpdesk.
For example, Finder is an excellent and functional file manager. Take its search functionality, which is easy-to-use and is placed logically on the tool bar — this is the kind of feature everyone will want easy access to! Many other functions that are less likely to be of much use to the average user — and that even includes cutting and pasting — are moved off the toolbar. This leaves Finder with a mere seven toolbar items as compared to eighteen in KDE's Konqueror. Now I am sure many are wincing at the idea of moving cut and paste functionality off the toolbar, but remember a good interface should aim to make the ninety percenters' features the ones that are prominent. Average users get dragging-and-dropping a lot faster than cutting-and-pasting files — most that I know never adopt cutting-and-pasting files at all. Most power users I work with prefer cutting-and-pasting, but they almost exclusively rely on keyboard combinations to manipulate the clipboard. So we must ask: is including cut, copy and paste on the toolbar really helping anybody?
Now, not to make Konqueror an object of attacks, because it is a fine tool and handles remote file systems better than any file manager I know. But, it is also an example of a disaster for the average user. Of those eighteen items on the tool bar, we have numerous ones that have nothing to do with normal file browsing — the zoom in and zoom out functions have minimal usefulness, as does the printer button and the CVS browsing icon. Yet all of these are there, by default, to confuse the average user. If anyone actually uses these functions in relation to file browsing, it is most definitely the ten percenters who could easily handle adding those icons to the toolbar manually, accessing them in the menus or by invoking them through key combinations.
In Konqueror's case, a lot of this has to do with KDE's focus on more experienced users, and while I wonder if even many more experienced users wouldn't prefer something more like Mac OS X's Finder, I will give the KDE team credit for that point. The beauty of the Free Software desktop is that I can choose the one best GUI for my desktop, my users' desktops and so on. In the big picture, however, for the GNU/Linux desktop to be adopted by the masses, more time must be spent on the ninety percenters and less on the ten percenters. In just four years, Mac OS X has gone from a delay-laden buggy OS to the most popular and polished desktop Unix-like system available. Sure Apple pays dozens of interface gurus, but I do not believe that this is what gives Apple the edge. Instead, it is an ironclad focus on the average user. Not by limiting power, but by molding each tool's interface to how the average user will work. iMovie came out long before the average user was doing DV editing — most still do not, for that matter — but by creating a sensible, simple interface, the average user's abilities can be expanded.
This is really the UNIX concept of applications being applied to GUI tools. As console users will tell you, the power of the UNIX and Unix-like shell has to do with the ability to take highly specialized, powerful tools and combine them to do what needs to be done. Grep does not support web browsing and lynx does not manage photos. Likewise, the Apple conception of the desktop involves applications that do just one type of thing — say organize photos or create movies — and do only that thing. As with the Unix-like shell, this separation of applications does not hinder the user because the applications are designed to work together and feature predictable, standardized interfaces.
The key is not to skip powerful features, but to arrange them in a way that makes them useful to the most people possible. Firefox and Thunderbird are arguably more powerful, where it matters at least, than the Mozilla suite, but they accomplish their goals without bundling everything into an incomprehensible mess. Both power users and average users find Mozilla's new applications pleasant to work with. Recent releases of GNOME also aim at this theory of user interface development, much to the chagrin of many vocal critics in the community.
GNOME, which is following Apple's lead quite a bit, need not scare power users away. Au contraire! In the case of Mac OS X, many Free and Open Source Software luminaries have found Mac OS X's friendly exterior and powerful interior preferable to a power user-focused interface. If you can still do everything you want to, but can just act like an average user when you want to take a break, you can. This is a “feature” that should not be feared but embraced by the Free Software community, which has, in the past, stood for using the right tool for the job rather than bloated suites. Rather than ignore this heritage and attempt to follow Microsoft's bloatware lead, the community should rally around projects such as Firefox and GNOME even if they are not always power user oriented. Ultimately this will bring more users of all skills to the platform, leading to more software development — that is more power and functionality — that everyone can enjoy.