According to Apache, they stopped tuning for performance when a low-end machine could saturate a T1. The Linux box in the test peaked at 27Mbit/second. That is, the Linux box could saturate 10base-T, let alone anything reasonably between a web server and the clients. If you've got multiple T1s coming in, you may as well have multiple machines at the end.
For file service, the Linux box put out 114Mbit/second. That's more than 100base-T, and a significant chunk of a gigabit. If you've got this sort of thing on your network, file server performance is the least of your worries. Unless you intend to have a separate network for file service, like in their tests, this isn't going to be a big deal.
In any case, haven't we just seen that NT beats Linux on Intel-based high-end servers, but sucks compared to equivalent UNIX servers? Really, Linux isn't ready yet for the high-end, and this hasn't yet become all that big a deal.
Clueless users complaining are, indeed, terrible. In order to get rid of them, however, the proper thing to do is to get them to read good documentation before they talk to anyone. However, it is far better to get users to find the answers to their questions on their own than to give up on the topic.
Nobody who's failed to document something they let someone else try to use, proofread that documentation, and get someone who didn't already understand the topic to understand it by way of that documentation can really say RTFM without being somewhat hypocritical. Ideally, the answer to any question would be RTFM, with a reference to TFM, which would actually answer the question, as well as an explanation of how the user should have found the answer without asking the question. Any question that can't be answered this way is a sufficiently good question to deserve help.
On the subject of computers-as-toasters, why not? I'm perfectly happy with Apache, which I have done essentially nothing to configure. It would be great if anyone could set up a new Linux box with an afternoon of waiting for downloads. For complicated or esoteric stuff, it is fine for the process to be more involved. I'm rebuilding my whole system, and it's made my learn a lot of stuff. I've configured an MTA recently, and I had to read a bunch of documentation to do anything interesting. This is fine, and probably good. Users ought to understand anything they want to mess with. On the other hand, if what you want to do with your computer is as simple as making toast, it should be as easy as using a toaster.
Primarily, ease-of-use in the presence of customizability is a matter of good defaults. I've spent a long time by now customizing fvwm. I have strange shortcuts that are efficient but arcane (Shift-shift, Alt-alt, &c). On the other hand, when I was starting out, the default.fvwmrc was perfectly fine, and I used it until I decided I cared.
Ideally, new users should be presented with an interface that isn't particularly quirky and behaves the way they expect. They should then be encouraged to try to make things work the way they would prefer. Then they should have to slog through complicated config files, and figure out strange features. Before that, they should be able to just have things work, so they can configure other things.
Imagine: you're sitting down to configure your kernel. You start up a text editor to edit it, but none of the keys do what you expect. You try to read the man page, but it scrolls right by, and your xterm won't scroll back. You try to pipe it through less, but the TERM environment variable doesn't seem to be set. The shell you're running is clearly neither csh or bash, and neither of those seem to be in your path. Oh, and you can't get the window manager to shift the focus back to the text editor.
Some things have to work before you can do anything. I've spent enough time running off a minimal (or sub-minimal) boot floppy to know that I expect computers to be as simple as toasters for some things. It really sucks when 'ls' doesn't just work. I don't really care if 'ls' is pretty without any configuration, but I do need it to list files without any configuration.
AOL appearently doesn't think they're different; the case is against AT&T for using "You have mail." when AOL uses "You've got mail." (at least, according to the article). Since trademarks are on exact wording, and single sentences can't be copyrighted, AOL doesn't even have any reason to complain if they were first, which they were not. This is why I think that countersuits should not need to be valid-- AT&T ought to be able to countersue despite not having a trademark or copyright, just based on the fact that AOL thinks there's something wrong.
Slashdot Mirror
According to Apache, they stopped tuning for performance when a low-end machine could saturate a T1. The Linux box in the test peaked at 27Mbit/second. That is, the Linux box could saturate 10base-T, let alone anything reasonably between a web server and the clients. If you've got multiple T1s coming in, you may as well have multiple machines at the end.
For file service, the Linux box put out 114Mbit/second. That's more than 100base-T, and a significant chunk of a gigabit. If you've got this sort of thing on your network, file server performance is the least of your worries. Unless you intend to have a separate network for file service, like in their tests, this isn't going to be a big deal.
In any case, haven't we just seen that NT beats Linux on Intel-based high-end servers, but sucks compared to equivalent UNIX servers? Really, Linux isn't ready yet for the high-end, and this hasn't yet become all that big a deal.
Clueless users complaining are, indeed, terrible. In order to get rid of them, however, the proper thing to do is to get them to read good documentation before they talk to anyone. However, it is far better to get users to find the answers to their questions on their own than to give up on the topic.
.fvwmrc was perfectly fine, and I used it until I decided I cared.
Nobody who's failed to document something they let someone else try to use, proofread that documentation, and get someone who didn't already understand the topic to understand it by way of that documentation can really say RTFM without being somewhat hypocritical. Ideally, the answer to any question would be RTFM, with a reference to TFM, which would actually answer the question, as well as an explanation of how the user should have found the answer without asking the question. Any question that can't be answered this way is a sufficiently good question to deserve help.
On the subject of computers-as-toasters, why not? I'm perfectly happy with Apache, which I have done essentially nothing to configure. It would be great if anyone could set up a new Linux box with an afternoon of waiting for downloads. For complicated or esoteric stuff, it is fine for the process to be more involved. I'm rebuilding my whole system, and it's made my learn a lot of stuff. I've configured an MTA recently, and I had to read a bunch of documentation to do anything interesting. This is fine, and probably good. Users ought to understand anything they want to mess with. On the other hand, if what you want to do with your computer is as simple as making toast, it should be as easy as using a toaster.
Primarily, ease-of-use in the presence of customizability is a matter of good defaults. I've spent a long time by now customizing fvwm. I have strange shortcuts that are efficient but arcane (Shift-shift, Alt-alt, &c). On the other hand, when I was starting out, the default
Ideally, new users should be presented with an interface that isn't particularly quirky and behaves the way they expect. They should then be encouraged to try to make things work the way they would prefer. Then they should have to slog through complicated config files, and figure out strange features. Before that, they should be able to just have things work, so they can configure other things.
Imagine: you're sitting down to configure your kernel. You start up a text editor to edit it, but none of the keys do what you expect. You try to read the man page, but it scrolls right by, and your xterm won't scroll back. You try to pipe it through less, but the TERM environment variable doesn't seem to be set. The shell you're running is clearly neither csh or bash, and neither of those seem to be in your path. Oh, and you can't get the window manager to shift the focus back to the text editor.
Some things have to work before you can do anything. I've spent enough time running off a minimal (or sub-minimal) boot floppy to know that I expect computers to be as simple as toasters for some things. It really sucks when 'ls' doesn't just work. I don't really care if 'ls' is pretty without any configuration, but I do need it to list files without any configuration.
AOL appearently doesn't think they're different; the case is against AT&T for using "You have mail." when AOL uses "You've got mail." (at least, according to the article). Since trademarks are on exact wording, and single sentences can't be copyrighted, AOL doesn't even have any reason to complain if they were first, which they were not. This is why I think that countersuits should not need to be valid-- AT&T ought to be able to countersue despite not having a trademark or copyright, just based on the fact that AOL thinks there's something wrong.