Re: New required base packages for Amiga, Atari, ... detection

[kaih@khms.westfalen.de (Kai Henningsen)]

martin@debian.org (Martin Mitchell)  wrote on 25.12.97 in <[🔎] 87ra727ibw.fsf@mail.usyd.edu.au>:

> kaih@khms.westfalen.de (Kai Henningsen) writes:
>
> > I seem to recall that the case in question (it _was_ Atari vs. Amiga,
> > right?) still allowed you to run _the_very_same_kernel_ on both systems.
>
> This has nothing to do with the kernel, please do not confuse the issue.

But it has. If you can use the same kernel (and nearly all other  
packages), the machines aren't incompatible enough that further  
distinction makes any sense.

IMHO, of course.

> > > specific programs that only work on one kind of hardware architecture,
> > > even though the same CPU is used for more than one hardware
> > > architecture.
> >
> > Well, yes. That _is_ the same as ISA/EISA/PCI/VBL/MCA.
>
> No, there is more to a hardware architecture than just a bus standard. If
> I were to apply that principle, then you would say the only difference
> between an Amiga and a Mac is Zorro vs Nubus.

I don't know enough about 68k Macs. Furthermore, I didn't say anything  
about "the only difference"; the Bus isn't the only difference on Intel,  
either. It's just the major part of the difference.

> The reality is the architectures are quite distinct, each having their own
> different video, memory and I/O subsystems integrated on the main board.

I _really_ don't think this is seriously different from the Intel case.

> Sometimes userspace utilities are needed to manipulate some of these
> custom chips: why should they be installed for architectures for which
> they will be irrelevant? Why shoud the Mac fdisk, for example, be installed
> on an Atari by default, through the base disks?

The same problem exists on the Intel side. In fact, on the Intel side,  
there are cases where the same kernel _won't_ work for all machines; see  
the Tecra kernels for an example.

As to superfluous tools - pnp tools on machines that don't have pnp,  
pcmcia tools on machines that don't have pcmcia - I really don't see the  
difference to the Atari/Amiga case.

> > Well, I don't.
>
> Could you please elaborate on your reasons? I fail to understand why you
> think this proposal is a problem.

I fail to see why the Atari/Amiga case is any different from the Intel  
case.

> > Actually, I already knew all that. It's _why_ I think this isn't a good
> > idea.
>
> What is? Again, you haven't made your reasons clear.

The reason is simple. You haven't demonstrated why we should do it, in  
view of the fact that we don't do it on Intel. You haven't demonstrated  
that the Atari/Amiga case is really any different from the Intel case,  
either.


MfG Kai


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org . 
Trouble?  e-mail to templin@bucknell.edu .