Jonathan Stone writes:
> >As a last resort I trashed the mbr & partition table with
> >`dd if=/dev/zero of=/dev/rsd1d bs=32k count=100', booted a 1.4 floppy,
> >ran sysinst to setup the mbr and dos partitions, and quit sysinst
> >after it setup the disks but before it started extracting sets.
> >It worked. I was impressed that sysinst got that right.
> 
> Er... Luke, that supports what *I said, not what Chris said.
> 
> Read it again if you dont beleive me.

Maybe I wasn't clear enough.

My first attempts to put an mbr and dos partition on my new disk were
by hand (using fdisk, disklabel, etc). I had problems getting that
working.

My final attempt was to
	a) trash any remnants of my prior attempts at setting up
	   the mbr (using dd)
	b) boot a 1.4 floppy and try sysinst.
sysinst managed to put a working mbr/dos partition table on my disk,
and setup the netbsd partitions. I was impressed by this.

Read what I replied to (cgd's message <87btdvh73z.fsf@redmail.redback.com>)
and what I wrote again if you don't believe me. In that message Chris
quotes you as saying:

``
	> Jonathan Stone <jonathan@DSG.Stanford.EDU> writes:
	> > Last: my impression is that most of the actual problems with sysinst
	> > are due to the x86 MBR handling.  Lest some people forget,
	> > most of the feedback on 1.3/pmax was along the lines of:
''

And the first two lines of that quote are what I was disagreeing with
to some extent (since sysinst actually did the right thing for me).