Technology that requires a manual is technology that doesn't get used. At least, where mainstream users are concerned in the consumer and enterprise software markets. One of the lessons of the last 30 years of computing, and particularly in the rise of the consumer Web, is that ease-of-use trumps deep functionality most of the time.
That's what made Microsoft the billion-dollar behemoth that it is. It's what is driving Apple's iPhone into millions of consumers' hands. And it's what makes Facebook, Google, and other Web companies so successful.
They're easy. They're intuitive. They solve real problems for real people. And all without a single manual.
Can open source also deliver this kind of mindless (and productive) ease of use?
It's not a question of documentation. Documentation for many open-source projects is chronically weak, but then, most documentation for most software is pretty weak. It's not just an open-source problem.
It also doesn't matter that you and I find a given open-source project supereasy to use. The only thing that matters is what mainstream end-users think, because they're the ones who create meaningful markets.
The kind that adds up to billions in sales.
Good technology spreads virally, as Matrix Partners David Skok points out. It depends on users being able to adopt technology without really having to think about it, and then tell their friends.
It doesn't spread because of all the great things that technology could do...if only the user could figure it out. Gartner's Jeffrey Mann rails against "vendors [who] confus[e] 'You can use it to do that' with 'We designed it to do that'," and open-source advocates are guiltier of this than most.
It's not what the software can do. It's what it does. For normal people. Without training or user manuals.
The open-source world, but particularly the Linux 'desktop' crowd, attempts an end-run around this argument by, somewhere, who runs (or should run) Linux (or some other open-source project).
But this is the wrong sort of argument, and precisely the wrong way to attract mainstream users. Facebook is its own argument: it's immediately clear what it's for and how to use it. Google is the same. TiVo? Ditto. And so on.
Most people don't have any agenda for choosing technology other than a) it's easy and b) it lets me communicate/work with all their friends and family.
Such software just needs to work. With minimal or little training and no ideological motivation. At all.
Can open source deliver this sort of experience? Yes, but it doesn't happen nearly enough. Commercial open-source projects like MindTouch go a long way toward making enterprise technology easy to use, while community-led open-source projects like Handbrake and Adium make open source supereasy for consumers.
These are the exceptions, unfortunately, not the rule.
We need more efforts like Canonical's Ubuntu, which is. Every open-source project should have a UI designer involved. The focus of every project should be to be able to run with minimal or no prior training.
Even the ubertechnical projects focused on developers. If Microsoft teaches us anything, it's that there are far more "average" developers/IT administrators/architects/etc. than rock stars, and such people benefit from a nice UI rather than a nice command line.
It's time to get out of the weeds of open-source development to equally focus on the end-user experience. Open-source can do this, but it must become a priority.