Try not to forget that many people only have one machine, and even if they have more than one, an NFS server might be a little hard to find, in a Windows shop. An FTP install will probably be easier to set up.
Also, naive users may not appreciate having to configure their network connection before installing. Granted, the network install is a great feature - I love it dearly, and use it whenever I can, but it's not for everyone.
Google tries to determine how important a web page is by counting the how many other pages link to it. This works most of the time, because, for example, if you look for the word "Linux", Google will determine that www.linux.org must be a good site, because so many other sites link to it. Yahoo doesn't do that, and brings up www.linux-net.ch, with www.linux.org a close second.
Comparing Google with other search engines, using more obscure search terms, usually results in Google getting you where you want faster. Usually, but not always. Coming up with www.microsoft.com is a good example. According to Google's rating alrogirthm, this is an "important" site, because so many other pages link to it, and it might very well be, but it doesn't necessarily contain the information you need.
I wonder what it would take to 'emulate' a Pentium on a Pentium, and forge the ID?
Not much, probably. Ultimately, it's the communications software that's trusted, not the hardware. If a web site wants to know what your CPU ID is, it can either: 1) Ask the browser, or 2) have the client download a piece of trusted code (a signed ActiveX, perhaps) which queries the CPU ID and sends it back, possibly encrypted.
Either case is easy to spoof. In the first case, you just patch the browser, and have it send a spoofed ID. In the second case, you modify the browser to trap the ActiveX download, and then have it patch the ActiveX in memory to spoof the ID. The patched ActiveX then happily encrypts your spoofed ID, and sends it back. There's no way the web site can know what happened.
Granted, the second exploit is harder to pull off, but no harder than taking advantage of a buffer overrun, or disabling software copy protection, and both are provably doable.
I believe you're missing the point. Non Microsoft proprietary operating systems will be able to support proprietary hardware. Free operating systems can't, because you're not allowed to publish hardware specs, or driver source, for proprietary hardware.
That's not entirely true. Linux needs a sufficiently large installed base to get the hardware verdors' attention. The code may very well be free, but what good is it if all the hardware you can buy is proprietary, and requires you to sign an NDA to develop drivers for? Free software needs the market share, because market share is what hardware vendors care about. If we can convince them that by supporting Linux they'll sell more hardware, Linux will continue to be a viable option.
Slashdot Mirror
Try not to forget that many people only have one machine, and even if they have more than one, an NFS server might be a little hard to find, in a Windows shop. An FTP install will probably be easier to set up.
Also, naive users may not appreciate having to configure their network connection before installing. Granted, the network install is a great feature - I love it dearly, and use it whenever I can, but it's not for everyone.
Google tries to determine how important a web page is by counting the how many other pages link to it. This works most of the time, because, for example, if you look for the word "Linux", Google will determine that www.linux.org must be a good site, because so many other sites link to it. Yahoo doesn't do that, and brings up www.linux-net.ch, with www.linux.org a close second.
Comparing Google with other search engines, using more obscure search terms, usually results in Google getting you where you want faster. Usually, but not always. Coming up with www.microsoft.com is a good example. According to Google's rating alrogirthm, this is an "important" site, because so many other pages link to it, and it might very well be, but it doesn't necessarily contain the information you need.
I wonder what it would take to 'emulate' a Pentium on a Pentium, and forge the ID?
Not much, probably. Ultimately, it's the communications software that's trusted, not the hardware. If a web site wants to know what your CPU ID is, it can either: 1) Ask the browser, or 2) have the client download a piece of trusted code (a signed ActiveX, perhaps) which queries the CPU ID and sends it back, possibly encrypted.
Either case is easy to spoof. In the first case, you just patch the browser, and have it send a spoofed ID. In the second case, you modify the browser to trap the ActiveX download, and then have it patch the ActiveX in memory to spoof the ID. The patched ActiveX then happily encrypts your spoofed ID, and sends it back. There's no way the web site can know what happened.
Granted, the second exploit is harder to pull off, but no harder than taking advantage of a buffer overrun, or disabling software copy protection, and both are provably doable.
I believe you're missing the point. Non Microsoft proprietary operating systems will be able to support proprietary hardware. Free operating systems can't, because you're not allowed to publish hardware specs, or driver source, for proprietary hardware.
That's not entirely true. Linux needs a sufficiently large installed base to get the hardware verdors' attention. The code may very well be free, but what good is it if all the hardware you can buy is proprietary, and requires you to sign an NDA to develop drivers for? Free software needs the market share, because market share is what hardware vendors care about. If we can convince them that by supporting Linux they'll sell more hardware, Linux will continue to be a viable option.