> On Mon, 10 Apr 1995 16:19:36 -0700 
>  Ted Lemon <mellon@vix.com> wrote:
> 
>  > Oh, but wouldn't *that* be a win!   Probably not brutally difficult
>  > either, although it'd be a lot of work.
> 
> Yes, it would (win and work...).  I've been thinking of the semantics of 
> this a bit...perhaps a `swap-byte' flag that's handled on a per-mount 
> basis...this flag could be set by reading the superblock, if the magic 
> number is wrong, flip it around, check again, if it's fine now, set the 
> flag...you get the idea...

some would say that you shouldn't slow down the 'default' FFS at all
to do this.  i.e. clone the current ffs into a different mount-type,
e.g. "bufs" (for backwards ufs...  8-), and do it there.

While compatibility is nice, it's not clear that you want to slow down
your 'default case' significantly to get it.