You mention that OS X beta should be pretty stable because its built on BSD. This doesnt mean that they cant break it. I had problems DP3, specifically with 'su' and other pretty important system binaries. I hope that whatever the hell they were doing breaking the unix base has been fixed.
It is possible to ruin a stable unix base.
Also, you comment about redhat as a beta release. I really think that we should sit down and really talk about what the word 'beta' means when we use it to talk about linux. From what I understand, RedHat doesnt claim that something is relatively stable untill the x.2 release. For Debian, they do things differently, more like *BSD but not quite. Of course, i dont know if i would call anything from RedHat stable.
Keep in mind that the BSD's are getting serious help from BSDi with SMP code. I'd expect to see something like this from NetBSD pretty soon. Lets not forget, if it ends up in a BSD distro, that means that anyone can use it. -> good for linux.
Lets not forget that bootsrapping isnt just as simple as loading the kernel.
Under FreeBSD, booting is usually a three step process. The second stage loader must switch repeatedly between real and protected mode on the i386 so the size of the bootloader is limited to 64K. Sounds to me like it would be a real pain to implement an XML parser that could be included in the second stage bootloader to read config files written in XML. They even implement a very stripped down version of printf() just to make everything fit.
We are sacraficing something useful (a configurable bootloader) for something that is easy to use (a bootloader without a config file because the XML implementation wont fit into ram durring boot).
And just to be safe I'll site my sources. Check out the/sys/i386/boot/biosboot directory on any FreeBSD machine.
You mention that OS X beta should be pretty stable because its built on BSD. This doesnt mean that they cant break it. I had problems DP3, specifically with 'su' and other pretty important system binaries. I hope that whatever the hell they were doing breaking the unix base has been fixed.
It is possible to ruin a stable unix base.
Also, you comment about redhat as a beta release. I really think that we should sit down and really talk about what the word 'beta' means when we use it to talk about linux. From what I understand, RedHat doesnt claim that something is relatively stable untill the x.2 release. For Debian, they do things differently, more like *BSD but not quite. Of course, i dont know if i would call anything from RedHat stable.
Keep in mind that the BSD's are getting serious help from BSDi with SMP code. I'd expect to see something like this from NetBSD pretty soon. Lets not forget, if it ends up in a BSD distro, that means that anyone can use it. -> good for linux.
Lets not forget that bootsrapping isnt just as simple as loading the kernel.
/sys/i386/boot/biosboot directory on any FreeBSD machine.
Under FreeBSD, booting is usually a three step process. The second stage loader must switch repeatedly between real and protected mode on the i386 so the size of the bootloader is limited to 64K. Sounds to me like it would be a real pain to implement an XML parser that could be included in the second stage bootloader to read config files written in XML. They even implement a very stripped down version of printf() just to make everything fit.
We are sacraficing something useful (a configurable bootloader) for something that is easy to use (a bootloader without a config file because the XML implementation wont fit into ram durring boot).
And just to be safe I'll site my sources. Check out the