> But, doesn't it seem that 1983 is a little out-dated in terms of
> technology?

Yes.  But that doesn't mean it's not in use.  The Sun-3 I use as an X
terminal at home is a -3/260 that was in regular service in 1983.  And
it's still got its original disks, AFAIK.

And yes, the kernel's concept of disk geometry was useful in its day,
but it is seriously outdated now.  I think - especially for SCSI and
similar interfaces that present the disk as a big vector of fixed-size
sectors - we should be treating the disk as just that: a big array of
sectors, with performance penalties for non-locality-of-reference.  And
instead of trying to do rotational layout and keeping things on the
same track, we should give up - disks with megs of cacheing on the
drive (both write-through, possibly reordering, and read-ahead), with
bits coming off the platter faster than they're getting shipped up to
the cpu, with variable geometry, etc, etc, it's not doing us much good
any longer.  And legacy hardware, while it's true it's still in use,
won't take all _that_ much of a performance hit from it, certainly no
worse than using a ccd to concatenate two disks of different geometries
imposes.

					der Mouse

			    mouse@collatz.mcrcim.mcgill.edu