>The problem is that, while ISO 9660 CD's will mount fine (due to
>architecture independence), a real FFS-type CD-ROM for the NeXT (on
>black hardware) has big-endian byte order, whilst you're trying to
>read it on the 386 platform, which has little-endian ordering.  This
>means that all the data structures in the disk label are all in the
>wrong byte order.

>Alistair G. Crooks (agc@uts.amdahl.com)

I began thinking about the discussion on using disks from one NetBSD with 
other NetBSDs, and I suspected this.

The byte order issue was the whole reason behind NeXT's support only for 
486s and higher, Intel has some sort of byte order swap mode built in to 486 
and higher, thus making an Intel NeXT port, from the 68k black box, possible 
or at least more reasonable.

I wont even go into my thoughts on writing machine dependant code, at the 
least UNIX (I know, device driver are an exception), but then it could be a 
Mach-ism, with which I'm not familiar.

thanks
jam