I just want to add a couple of comments to this after seeing the appends on
this thread.  The FORMAT=DRIVE means, as said, pick the best format that the
drive can use.  This means "compressed" for those drives that support it.
However, if you have one library/autochanger with two drives, and one supports
compression and the other doesn't, DONT USE FORMAT=DRIVE.  Same goes for
mixed-capability drives (eg DDS and DDS2 4mm drives).  The cartridges written
with the higher capacity won't be readable by the lower capacity (and you
can't control which drive gets used).  Finally, DRIVE is the default format,
so if you didn't specify it when you define the devclass, that's what you've
got.  Again, if you used DRIVE and you have mixed drives in the autochanger,
then those volumes that are loaded in the lower-cap drive get formatted to
the lower capacity, even when they are later used in the higher capacity
drive.

Along these lines, once you start writing to a cartridge in one format, you're
stuck with that format for the rest of the cartridge.  So any vols in the
filling state with, say, format=8500 are stuck that way.

The Estimated capacity doesn't really do anything except guide migration from
the storage pool to the next pool.  I'm not totally sure it even does that.
If you get more data on a volume than the estimated capacity, adsm just
updates the estimate as it fills.

Craig Bell