OpenSourceUsability

MeatballWiki | RecentChanges | Random Page | Indices | Categories

Originally on LiveJournal:sunir/17773.html...

Reading ["Ronco Spray-On Usability"] and nodding my head, I started to think about why MacOSX? could make FreeBSD work and why Linux will continuously flounder around in user interface hell for the rest of its existence. Obviously it's the limitation of copyleft.

As John Gruber writes above, user interface design requires selling the software itself, not services or documentation, as easy-to-use software doesn't require support. Thus, only a closed source model can support it. However, BSD is a very successful open source project, and as a StableBase (to mix categories) to build upon, it allowed a commercial entity to graft on a new user interface. CopyLeft is designed specifically to prevent the investment of capital into the codebase (only into the people).

But the catch, of course, is that Apple had to be beaten over the head before releasing anything back to BSD, which although might be the economically wise thing to do in many cases, is something that requires convincing for skittish executives.

An alternative is something like the LGPL, that allows one portion to be copyleft and another portion to be closed source, but there is no philosophically clear boundary between one piece of software and another, so in practice this becomes clumsy. (Zeno's paradoxes in the digital age.)

Nonetheless, I think an ideal would be a patchwork of open (and free) code and commercial code, as good commercial code will be cloned by open source hackers quickly enough to keep pushing the frontier forward. But I digress until I think about this more.

As always, missing in these discussions is the nature of the social forces that actually build the software.

I don't think it's theoretically impossible to make beautiful interfaces in open source. I think on average many factors weigh against it so that, on the whole, it does not happen as frequently as in commercial sector.

One commonly raised problem with GUIs is that they get in the way of hacking. Hackers like the pipes and filters and files metaphor of Unix. I can see perhaps why a person who uses Linux as a hobby might value software they can hack over software they don't have to hack, but therefore can't, with a consequential reduction in potential.

But I think also that visionaries get bored working for free. Once you build a résumé with an open source project, you can get a job for money, and so your creativity goes with you. And with you goes the unifying vision. Unlike the rest of open source where lots of people can hack the living snot out of the backend without ever knowing or talking to one another--which does leave an ugly pile of mismashed code on occasion, but it compiles and runs--hacking the snot out of the GUI experience will turn it quickly into a hellish interface. I think GUIs are just not amenable to the mongrel approach to coding that is open source.

For what it's worth, and to be a weasel, I'm hardly convinced by anything I just said, but I wanted to write more here than a quiki. -- SunirShah

CategoryInterfaceDesign CategoryOpenSource?

References

Nichols, D. M. and Twidale, M. B. (2003). The usability of open source software. First Monday, 8(1). Available from http://firstmonday.org/issues/issue8_1/nichols/index.html [uncooked]

OpenSourceUsability

References

Discussion