In message: <Pine.NEB.4.33.0212221521100.15890-100000@vespasia.home-net.internetconnect.net>
            Bill Studenmund <wrstuden@netbsd.org> writes:
: > Any reason not to do a GEOM-like method of letting the disk tell you
: > what slices are on it?
: 
: I'm not familiar with what GEOM is and isn't. Could you point me to info
: that explains it?

phk's paper on GEOM describes it fairly well.  I don't have a pointer
to it (I'll post one later).  Some have said it is too complex, but it
is infinintely expandible.  I'm now able to read/write my pc98
(japanese PC) hard disks on an i386 box, which helps a lot when
testing new software.

: My main thought is that we want something that works in a wide-variety of
: partitioning schemes, so we don't want to go with a one-size-fits-all
: idea. Though as one of the many labeling schemes we know, it'd be fine.

True.  GEOM isn't a labeling scheme, but rather a framework for
dealing with the multiplicity of labeling schemes that are extant
today.

Warner