.comment: Luddite Linux - page 4
The Role of the Modern GUI
We're all of us excited about the latest developments; DVD for Linux is sexy, indeed. I'm tremendously enthusiastic about KDE2. But let's all of us recognize that the GUI isn't the answer to all problems. And let's do something more; let's return to first principles.
There are thousands, millions, probably, of machines out there that are perfectly serviceable and that are now useless for lack of software: decent, relatively easy-to-use word processors, spreadsheets, and so on. The thing is, with a few simple applications they could do much to bring computing to places where it either doesn't exist or else is saddled with decade-old software. Imagine a small, tight Linux distribution for such machines. It would, among other things, put the lie to the notion that Linux is difficult to use.
For a few years, Linux distributors offered text-mode installation and configuration utilities. Some worked well and were easy to figure out; some didn't and weren't, but not through any problem intrinsic to the interface. The new, graphics-based configurators are no easier to use for being graphical. But they are bigger and slower.
Nowadays, it seems as if everyone who has a CD burner is a Linux distributor. It is possible to have favorites among the distributions--I certainly do, though I haven't tried them all--but there is a certain sameness to them. How lovely it would be to see someone strike out in an entirely different direction, not to support the hottest new hardware but to maximize the utility of Linux on low-end machines. I betcha it would be possible to put together a distribution that would cook along on a 386-16 with 16 megs as well as DOS ever did. A little menuing program and a group of useful applications--some are already written--would make it user-friendly. (And this Luddite Linux, as I'll call it, would simply scream on a fast machine.) Graphical programs? Can you say runtime X?
Eric S. Raymond has noted that programmers create in large measure out of a desire to scratch an itch. That itch is always going to be in hope of producing something that doesn't exist, sometimes out of need but often because it's cool. By this reasoning, there's not likely to be any stampede toward the production of text-based, easy-to-use applications.
But anyone looking to crank out a truly unique Linux distribution could do worse than produce a console-only product for low-end machines. It would be a productive desktop operating system. And it would resurrect a lot of ancient hardware while at the same time spreading the word about Linux and putting to sleep the notion that Linux is difficult and troublesome for mere users.