Despite all of the vast gains seen in technology over the last few decades, user interfaces on desktop computers have remained pretty much the same. We have insanely fast CPUs and graphics cards to make our windows wobbly, and we have truckloads of RAM to run countless apps simultaneously. But we still tend to have the same kind of setup on our desktops: a program launcher, a taskbar or list of running programs, a "system tray" area, various windows scattered around, and so forth. Even the desktop environments that have bucked the trend in recent years, like Gnome 3 and Ubuntu's Unity, still keep most of the basics that we've been using since the Amiga Workbench days.

Now, you could argue that this is a good thing. User interface design has been refined over the decades, and we're all used to how WIMP (windows, icons, menus, and pointing device) setups work, so why change it? Why throw all of this away for something novel and "experimental" that requires learning a whole new workflow?

Well, sometimes that workflow can make you faster. A lot faster. Look at the command line, for instance: We Linux and Unix geeks know that it's not as pretty or welcoming as a flashy GUI, but it gets many things done much more quickly. Prodding the mouse, going through tabs in a dialog box, fiddling with sliders and radio buttons – who wants that when you can just type in a command and get your flippin' work done?

[...]