Too Busy For Words - the PaulWay Blog

At the risk of perpetually turning this blog into a Pascal Klein rebuttal forum, I have to disagree with you, Pascal, about your interpretation of Hugh's One Frickin' User Interface rant. This is my rant.

You yourself, Pascal, work on the Tango Project - trying to get a consistent set of icons that mean the same thing across Linux distributions and across windowing interfaces and websites. You yourself acknowledge how important this kind of consistency is: that when you click on a button that has a little starry glint in its top right corner you expect it to make a new something, for example. If this was true only part of the time, then it'd be useless. I cannot see how open standards or DRM issues can hold back widespread use of Linux one tenth as much as a lack of an interface that is both consistent between Linux distributions and (vaguely) consistent with what the rest of the world uses. No matter how secure, speccy, feature-packed, free, unrestricted, modifiable, configurable, and well-supported Linux might be, if it has an interface that people find hard to get used to quickly, then they won't use it.

The argument that having KDE and GNOME is necessary for competition is a fallacy. We already have competition: Windows and OS-X. Having more than one GUI for Linux wastes development time, causes holy wars, and makes it much harder for developers to make applications that work across the range of installed systems. That's stupid, and pretending that it's useful is just denying that we have a problem. How much time did Canonical spend getting KUbuntu working because the KDE fanatics whined? How much time is spent keeping it parallel with Ubuntu now? Is that time we can waste? No!

To argue that the commercial failure of competing user interfaces is irrelevant because "that's just companies competing" is likewise fallacious: it's a straw man argument. Linux is competing for market share just like the proprietary operating systems; the fact that its source code is open and that by and large it's under non-restrictive usage licenses is just a development methodology. There are plenty of companies that would be more than happy to have a single GUI for Linux to standardise on - Canonical, Red Hat, Sun, IBM, SGI, Oracle... But no, we have more widget libraries and APIs than we know what to do with.

I offer an explanation to your observation that Windows users don't notice the consistent user interface: because it should be completely beneath notice! It would be noticeable if it was bad! Plonk those people down on a Linux distro and what do they say? "All the icons are wrong!" "Everything's in a different place to what I was expecting." "Everything looks funny." That's attention; unwanted attention. If you could plonk a Windows (or Mac) user down on a Linux box and they didn't notice at all, then the interface would have succeeded in its goal of allowing people to use it. This doesn't have to be at the cost of having all the nifty features that we all know and love, such as workspaces or tabbed windows. But it does have to be at the expense of "difference for the sake of being different."

I think it's pointless debating which was the top feature for Apple, a rigidly enforced hardware standard or giving top priority to the user interface. They're equally true. The Macintosh has continued to be a going concern, in spite of almost irresistible strongarm tactics from Microsoft and selling consistently more expensive equipment that was often still slower than same-era PC equipment, because usability is always given priority over flexibility. "Working out of the box" and "providing a pretty login screen" is just part of that usability priority.

The point Hugh, and I, are making here is that to have a plethora of APIs for GUIs on Linux, we're dividing the developer base. Microsoft and Apple couldn't want more division in our ranks. We should have developers saying, "how can we make it so easy for everyone to write Linux GUI applications that they'll do it quicker than for Windows or Apple?" Instead, we seem to have developers sitting in their little couch forts throwing abuse at the others in their other couch forts and refusing to move because they "provide much-needed diversity". All it does is give proprietary OSes time to consolidate and marginalise us again. And that's just stupid.