Wayne,

I wouldn't say that unloaddb/loaddb was that brilliant at reducing
space in DB.  Initially yes, we saw about 40% saving on our 4 year
old DB.  But I think there is something not very clever in the way
space is then allocated and what we found over the next few weeks
was that DB space was going up quite a lot quicker than expected.
I'm sure we saw gains in efficiency : because redunant entries had
been removed.  Also, the downtime (days) to do it is unacceptable.

What is needed is on-the-fly tuning - but whether that ever comes,
I wouldn't know.

Regards, Sheelagh