Understanding open source Part 2

So for a given OSS product, end users have choices. Some will take the time to test several distributions. Some will write reviews or participate in online discussions to disseminate their opinions. Ultimately the word will get out that some distributions are better than others, and those will make their way to more desktops. Managers of less popular distributions, faced with dwindling user bases, will respond by joining forces with more popular distributions, changing the distribution in response to user criticism or working with developers to bring about improvements. The net result: More popular distributions will succeed, less popular distributions will disappear or adapt.

That doesn't mean better distributions will always succeed, but on average I believe it does: Most distributions will be interoperable (effectively the same product), so a monopoly situation where users don't switch because of incompatibility or cost is unlikely. Popularity sometimes derives from image rather than substance, but I believe practical considerations that substance will win out over time in most cases.

A more popular distribution is not necessarily more usable. Certain aspects of usability are undeniably misinterpreted or mischaracterized by users: They either don't know anything is wrong or can't identify the problem correctly. But poor usability tends to cause frustration, even when a user can't put his or her finger on why. Faced with more and less frustrating options, users will probably pick the less frustrating one, opting for a more usable product even when they don't know the specific reason. That is especially true in the case of OSS, since distributions are often similar in other respects (such as price) that traditionally trump usability.

So more usable distributions will tend to succeed given enough end users and choices. And distributions differ because of the innovations they incorporate. So more usable innovations will succeed, while others will adapt or fail.

Usability evolution is more than just an analogy to a biological process. It is the same principle, applied to software not because it's a good idea but because it applies itself. As long as people modify open source software, and as long as users choose among those modifications, usability innovation through evolution seems inevitable.

Usability evolution has probably been around for some time, but its effects have been minimal:

• With commercial software, other factors overpowered evolution, including cost, compatibility and investment in existing technology and training. But for concurrent distributions of the same OSS software, the user experience is similar so training costs are negligible; the financial cost (particularly that of switching to a competitor) is low or nonexistent; and all distributions are functionally the same product, so compatibility isn't an issue. Several versions of the same product are competing instead of several different products.
• OSS has been largely ignored by end users, so any usability evolution has been on a small scale and not particularly noticeable.
• In the past, there were fewer users and fewer developers. More developers are getting involved, more are interested in wooing end users, and more end users are starting to catch on. As OSS becomes more popular, those numbers should continue to increase as barriers to entry drop for both users and developers. And the more people there are developing and using software, the more powerful the evolutionary effect is likely to be.

A place for usability engineering

Traditional usability engineering still has a place in OSS, both within the evolutionary process and outside it. It's likely that some projects will have the means to do full-fledged user interface design and usability testing. That should increase an innovation's fitness and thus give it an evolutionary advantage, allowing usability engineering to continue paying off.

As discussed earlier, some companies have a commercial stake in OSS projects. Since product usability is in their interests, they are likely to conduct usability tests and do UI design on the core platform, specific distributions or individual innovations. If the results are applied to the core platform, their effect may be outside the evolutionary process. If they are applied to a particular modification or innovation, they'll simply be part of the usability evolution.

Open source software is not new. The notion of free or inexpensive software, whose development is structured so that anyone can improve on it, is exciting, particularly when commercial software secrets are too closely guarded to allow cooperation and collaboration on a large scale. What is new is open source software's serious entry into the end user space, providing consumers with equally usable alternatives to their current software, options that provide the advantages of OSS without the stereotypical drawbacks. That shift calls for a reexamination of the models of how user experiences improve. Usability evolution may be a largely untapped power inherent in the open source model that can complement our current usability methods and allow non-usability professionals to contribute to the evolution of the user experience.