a long rebuttal to the Linux-is-the-engine fallacy

[Gordon Messmer]

Marko Vojinovic wrote:
> On Sunday 27 July 2008 03:40, Alexandre Oliva wrote:
>> On Jul 26, 2008, Marko Vojinovic <vvmarko at panet.co.yu> wrote:
>>> But the system without a kernel has *precisely zero* usability.
>> Yet you provided and cited the counter-example yourself: the boot
>> loader required to load the kernel, and that actually provides some
>> usability, including the ability to load a fully-functional program
>> such as memtest86+, that runs without a kernel and is also included in
>> the distribution. You could get many other useful programs that run 
>> in "real mode" started from grub.
> 
> I don't understand what you mean by "real mode". Memtest runs under the bios 
> operating system. The fact that it is a part of a distro and being invoked by 
> grub is just a matter of convenience. No program runs without some sort of 
> kernel, except the kernel itself.
...
> Also for grub itself. It is an application that runs under bios.

There is no magic in Linux that distinguishes it from GRUB.  Both are 
"applications" that run in real mode.

Real mode is an operating mode for general purpose processors wherein an 
application has full access to interrupts, hardware IO, memory, etc. 
Most kernels run applications in protected mode, where they do not have 
direct access to hardware.

That's the only thing that really separates the kernel from other 
processes.  It runs in real mode.  So does Memtest86+.  And GRUB.  And 
Xen.  They're all just processes.  Your belief that the kernel is a 
magical element of an operating system stems from the fact that you 
don't understand what real mode is, or how processors work.  That's OK, 
most people never will.  They don't have any reason to.  But since you 
don't understand that, please believe us when we tell you that GRUB, 
Memtest86+, Xen, and Linux are all applications that run in real mode. 
A kernel is not required in order to run a real-mode application, which 
makes Alexandre's comparison fair.  The system has *some* functionality 
without or before Linux runs, since users can interact with real-mode 
applications.  (They can actually do more with those real mode 
applications than they could with Linux and no GNU)

> It's main 
> purpose is to load some other, more sophisticated kernel, that, once running, 
> eliminates the need for both bios and grub. So you cannot exactely consider 
> grub to be a part that defines an os.

Well, the kernel's purpose is to load other, more sophisticated 
applications.  Again, there's no magic in the Linux kernel.  It's just a 
program, like all of the others.

> I agree that grub is as essential as the Linux kernel according to my 
> measuring stick (that is why I mentioned it in the first place), but it is 
> eliminated from the contest on other grounds. And I hope we all agree that 
> giving the name to a distro based on the bootloader is plain silly.

No more than naming an operating system after its kernel.  I'm not aware 
of any other operating system which is referred to by a name which is 
clearly the name of its kernel and nothing else.