Marco Tha Don
05-03-2006, 11:17 AM
There are quantifiable areas where Linux is superior:
1. Actual operating system performance shreds OS X. The hardly-concurrent XNU microkernel, according to a Tiger-vs-Linux, PPC-vs-Xeon test Anandtech did a while back (keep in mind, they also ran Linux on PPC, just to drive the point home), Apache could serve 10 times as many pages under Linux.
Now, granted, OS X is not designed for server usage (although I know people who think it is). But if you break down the reasons why this disparity in performance exists, the problems XNU has (such as terrible system call performance and locking design barely better than the "good 'ole" days of BKLs) can kick you in the balls in a lot of different ways.
Mach, on which XNU is based, is probably one of the worst operating system designs ever. I like Linus's take: "[Mach] contains all the design mistakes you can make, and manages to even make up a few of its own."
I'm just talking out of my ass on this final jab here, but I'm idly suspicious that Windows could probably give OS X a real run for its money in the performance department too.
2. The stack in Linux is still very much free (speech and beer). The free speech aspect is important; short of going into Stallman mode, Eric Raymond's open source principals come heavily into play when comparing software quality. But what really got me when I used OS X was that I had just gotten free from Windows and the world of ****ty piracy and shareware, and to do anything useful in OS X it seemed like I'd have to go right back.
--
Now, there are some areas where personal opinion comes heavily into play for me. Frankly, I think the OS X user interface, as played up as it is by its fans, is utter crap. I still can't get over lacking the second mouse button. Additionally, I'm a person who leans heavily on the "multiple desktops" concept of organization - I have my Gentoo desktop here configured for 6. The desktops stay in use, my windows stay organized, and I can leave a zillion applications open without even the slightest noticeable hit to the performance of whatever I happen to be currently doing at the time.
Every time I've tried to use OS X to do, well, anything, it feels like the GUI is getting in the way of my productivity. (I suppose that is related to my preference for KDE over GNOME - I like having a million clever switches, tucked away just right)
And the brushed metal of OS X looks tired and dull to me. I don't know why people love it so much; one of my primary frustrations with KDE is that when I go to kde-look to look for new themes, the most downloaded themes put that "skin" onto KDE.
More than anything, though, I'm a strong believer in free software. If I can't tear it apart, read it, understand and modify it, I'll pick something I can, unless reason forces me not to. (Example - I'm unfortunately dependant on nvidia.ko to make my 7800GT scream).
Apparently Jobs approached Torvalds in 1997, asking him to help make UNIX mainstream again. Linus didn't accept. It doesn't sound like a thing I'd figure Torvalds would accept; still, I'm infinitely pleased he stayed home and continued to work on the fastest, most powerful and most portable kernel there ever was.
1. Actual operating system performance shreds OS X. The hardly-concurrent XNU microkernel, according to a Tiger-vs-Linux, PPC-vs-Xeon test Anandtech did a while back (keep in mind, they also ran Linux on PPC, just to drive the point home), Apache could serve 10 times as many pages under Linux.
Now, granted, OS X is not designed for server usage (although I know people who think it is). But if you break down the reasons why this disparity in performance exists, the problems XNU has (such as terrible system call performance and locking design barely better than the "good 'ole" days of BKLs) can kick you in the balls in a lot of different ways.
Mach, on which XNU is based, is probably one of the worst operating system designs ever. I like Linus's take: "[Mach] contains all the design mistakes you can make, and manages to even make up a few of its own."
I'm just talking out of my ass on this final jab here, but I'm idly suspicious that Windows could probably give OS X a real run for its money in the performance department too.
2. The stack in Linux is still very much free (speech and beer). The free speech aspect is important; short of going into Stallman mode, Eric Raymond's open source principals come heavily into play when comparing software quality. But what really got me when I used OS X was that I had just gotten free from Windows and the world of ****ty piracy and shareware, and to do anything useful in OS X it seemed like I'd have to go right back.
--
Now, there are some areas where personal opinion comes heavily into play for me. Frankly, I think the OS X user interface, as played up as it is by its fans, is utter crap. I still can't get over lacking the second mouse button. Additionally, I'm a person who leans heavily on the "multiple desktops" concept of organization - I have my Gentoo desktop here configured for 6. The desktops stay in use, my windows stay organized, and I can leave a zillion applications open without even the slightest noticeable hit to the performance of whatever I happen to be currently doing at the time.
Every time I've tried to use OS X to do, well, anything, it feels like the GUI is getting in the way of my productivity. (I suppose that is related to my preference for KDE over GNOME - I like having a million clever switches, tucked away just right)
And the brushed metal of OS X looks tired and dull to me. I don't know why people love it so much; one of my primary frustrations with KDE is that when I go to kde-look to look for new themes, the most downloaded themes put that "skin" onto KDE.
More than anything, though, I'm a strong believer in free software. If I can't tear it apart, read it, understand and modify it, I'll pick something I can, unless reason forces me not to. (Example - I'm unfortunately dependant on nvidia.ko to make my 7800GT scream).
Apparently Jobs approached Torvalds in 1997, asking him to help make UNIX mainstream again. Linus didn't accept. It doesn't sound like a thing I'd figure Torvalds would accept; still, I'm infinitely pleased he stayed home and continued to work on the fastest, most powerful and most portable kernel there ever was.