Linux kernel 2.4.32 is a network operating system. But you have to use repositories in the grid to run data. Linux 2.6.x is a local preferenced distributed data operating system. Its been three years since the birth of Linux kernel 2.6.x and no finish line in sight. The reason is confusion.

Linux 2.4.x is a hdd virtual memory operating system, which is not usable for local preference database, Kernel 2.6.x has to use large ramdisk for data for speed.

Linux kernel 2.4.x has all the generic drivers, but they can no longer be used. Kernel 2.6.x uses udev and /usr drivers beyonf drivers used in bios. Mepis took the first step to make 2.6.x an event to provide correct drivers.

Kernel 2.4.x has Dhcp and dhcpcd for network operating system, but kernel 2.6.x has to have unionfs to synchronize mysql data in the ramdisk to the microsecond for data accuracy. 2.4.x still uses synaptic for software update.

All the major linux distros have to use official kernels released by Linus to conform to partition, directory and file name so languages, scripts and design tool kits and emulations can not be broken in bindings.

With tiny Linux distros, this is not difficult; but for major distros who changes kernels, the number of bugs to be fixed is not worth the troubles. As it is Linus kernels are already moving targets, difficult to comply to keep dependency compatible.

We can see Novell still has work to do on script binding maintainence. Fedora is just blowing the wind. Mandiva is lost in the forest.

Hopefully, all three major look at Slackware codes and see how to integrate 2.6.x into 2.4.x kernels. Then work out the minimum requirement of partitions, directories and file names to keep all the bindings along with all the hardware ports. Scritps amy have to be updated with Linus kernels. Terminals may have to include complete linux commands as each new kernels appear.

This is just a start(first step in the 10,000 mile journey) in the right direction to keep sanity in the Linux world.