.comment: Be Careful What You Wish For - page 2
Two Weeks Do Not a Linux Tryout Make
Mr. Kellner was manifestly foolish (presumably without knowing he was) in giving Linux a two-week trial. Someone who had used only Linux and who then popped over to Windows for a two-week spin would come up with equally useless results. ("Where's the compiler?" "How do I set permissions?" "What? You mean anyone can delete any file?" "They call this a file structure?")
But that's a little bit beside the point. Linux is not Windows. It is not just like Windows. If it were, what would be the point in choosing one over the other? Indeed, Linux isn't even just like Linux: the variation from system to system is enormous. I daresay that there is no other Linux machine in the world that is set up as mine is or, probably, as yours is. I'm using one of the 2.4.0-test kernels, and XFree86-4.02 with some but not all of its features enabled, and QT-2.2.3 with some but not all of its features enabled, and KDE built from source a few hours old that would itself be different if I'd downloaded it a few minutes earlier or later. I'm using glibc-2.2, meaning chances are that my binaries might not work on your glibc-2.1 machine, and somebody else's binaries built with gcc-2.96 or gcc-2.97 might not work on either of our machines. And that's fine. We accept it. We embrace it. Where necessary, we work through it.
We don't have to make things this complicated, of course. We could, and no doubt many of us do, get and install a distribution and keep that distribution on our machines until the next version comes out, and to the extent that we upgrade anything in the interim grab the RPMs or some other packaged binary. That's fine, too. But even that is more complicated than life in the Windows world. (Life could be made easier for anyone who used distributions if those distributions by default put /home on a separate partition and at upgrade time didn't touch /home at all. Frankly, better still would be distributions that didn't format partitions slapdash and instead intelligently looked at what was on the drive, upgraded as needed, and left the rest intact.)
What do we want, then? Well, we want all the source to compile perfectly everytime we try to build something; we want RPMs never to choke on a failed dependency; we want total control, but we want to do what we want to do without a lot of stuff getting in our way. That stuff can be incomprehensible screwiness in typeface handling, or it can be nonsense like bubble help that we can't turn off. There's a happy groove--not too tough, but not clogged with pablum for the clueless--that we seek. The problem is, that groove is located in a different place for just about every user--and it moves for each user as his or her skill increases.
Linux doesn't have the luxury of deciding the one size that will fit all users and then forcing all users to live with it. That is its great strength--what brought us to it--and its great weakness, because it means that it will never be the operating system of choice for idiots or for those who have only two weeks to spare, to the extent that the groups do not overlap.
Do we want everyone to switch to Linux? By this standard, we do only in that we wish our fellow humans were as bright as we are, and as interested in computing. Yet we keep holding the notion that Linux should and perhaps will win the desktop, that everyone use it, that everyone somehow see how much better it is.
Solid state disks (SSDs) made a splash in consumer technology, and now the technology has its eyes on the enterprise storage market. Download this eBook to see what SSDs can do for your infrastructure and review the pros and cons of this potentially game-changing storage technology.