(Excerpts from a message dated 12-01-99, Will Honea to Murray Lesser)

Hi Will-- 

ML>     Available-space discussions make sense only when one remembers
ML> that they are purely relative.  You really ought to understand that
ML> 8 MB is only 0.1% of an 8 GB drive (0.2% of a 4 GB drive, in case
ML> arithmetic is difficult for you), so is in the noise as far as
ML> usable space on such drives is concerned.  One tenth of one percent
ML> isn't worth worrying about, let alone making a serious(?) suggestion
ML> that legacy drive usage be modified to "save" any part of it.
ML> 
ML>     It was Euripides (fifth century, BCE) who wrote "The gods visit the
ML> sins of the fathers upon the children."  You will just have to
ML> learn to live, patiently, with those early architectural sins!  Of
ML> course, I suppose you could move to another platform and start over
ML> with another set of guesses about what the future may bring.  Or
ML> become a monopoly so you can ignore your legacy customers :-).

WH>I'm surprised that no one has thought to mention the IBM 'policy' of
  >truncating the disk allocation such that the last cylinder was also
  >not available until a few fixpaks back.  I don't recall the exact
  >details, but it was along the lines of forcing the last partition to
  >end no futher than the last head, last sector of the last full
  >cylinder on the disk.  If the last partition ended exactly st the end
  >of the disk, a full cylinder was reserved.  I think that there was
  >also a minimum reserved space required based on disk size.  While
  >this was fixed about the time the first >4.1 gig version of
  >IBM1S506.ADD (it applied to SCSI as well) there was considerable
  >discussion with Sam Detweiler on the news groups at the time since he
  >was writing the drivers.

   AFAIAC, the loss of that last cylinder is also "in the noise" :-(. If
I wanted to guess, I would say that the practice you describe dates from
the XT days when the last cylinder was reserved for a test track used
only by the hard-drive read/write diagnostics.  I had a small
assembly-language routine for my XT that moved the heads over to that
last cylinder before shutting down, so that if there wasn't a smooth
landing, there would be no damage to usable data.  As I understand it,
this was done automatically on the AT.  (I never owned an AT; I skipped
the 286 generation as it didn't make any sense to me.)  I would guess
that hard-drive reliability (at least for drives from reputable makers)
had improved (by the time OS/2 came along) to the point where that "test
track" was probably unnecessary, but nobody noticed it was still there
until relatively recently!

    Regards,

        --Murray

___
 * MR/2 2.25 #120 * If you are not confused, you don't understand the
situation

--- Maximus/2 2.02