"Six years ago, Matthew Thomas wrote a short piece on Why Free Software usability tends to suck. He's now expanded that less-than-700 word essay into a magnificent 2,400-word feature, Why Free Software has poor usability, and how to improve it.
It should be compulsory reading.
We've known for ages that the best way to design usable software is to do the design first, test it on real users (eg with paper prototypes or mock-ups) and then write the code. However, it's engrained in the open source process that you code first, test later and design never -- except maybe by copying whatever Microsoft or someone else did already.
We also know (from people like Jakob Nielsen) that you can learn a lot by doing usability testing with just 20 real people, which is a process within the reach of many free software projects. Maybe Matthew Thomas could add or expand this point in the next version.
When free software is a server-side and/or programming phenomenon, usability doesn't matter much. And while there is a vibrant commercial software business that employs usability experts and interface designers, and does proper usability testing, there will always be things to copy. But if free software destroys the commercial software business (which would be a reasonable if unlikely expectation, given the Free Software Foundation's agenda), then we could be in trouble.
It's certainly possible for open source projects to do very good work on design and usability: Mozilla's Firefox is living proof. However, Mozilla is absurdly rich by open source standards. I can't think of any other projects that are pulling in something like $50 million a year. However, that's also an absurdly small amount compared with Microsoft's $60 billion a year. And it's hard to see anyone getting to that level with open source code that, in many cases, you can't even give away." (Continued via guardian.co.uk, Matthew Thomas) [Usability Resources]