1. Major problems scanning blue sky from color slide (8 bit scans).
No problems for example with black, which was my major concern.
Haven't really done anything else than slide scans so far...

Hi, I've scanned loads of different blue skies no problem .... what do you
mean by major problem?

2. Speed, which does not exist. Because of speed related problems
I have so far limited my experiments to 2700 dpi/8 bit. But later when
I start "serious" work with the device, thinks need to change!
I ONCE tried 5400dpi/16 bit with ICE with multisampling; I left the
device scanning over night, and it was not ready in the morning!
Had to cancel it, which was not that easy...
OK, I admit that I use ICE often...have to check what Minolta really
claims the scan times to be with ICE...

Sounds not right .... multisample scans with ICE do take at least an hour,
but why bother, I and several in this group report there is no discernible
benefit; if anyone can show a good benefit of spending that time, then let's
hear about
Typical normal exposed slide on ICE at 5400 +16 bit depth should not take
more than about ten minutes ... but scan time seems to vary depending on
slide characteristics

your machine spec. similar to mine though I use XP ... I wonder whether 2000
gets the best from USB?