> In this case, there is a tighter coupling between the mice/keyboard
> and the manager.  This can be separated out more easily than
> scrapping the whole idea...

This sounds like the right thing to do.   The ADB is just a bus, after
all.   Also, if it's possible to tell where a given piece of
information on the ADB came from, it's probably desirable to do so -
if we can tell that a given ``keypress'' actually came from the mouse,
we can treat it differently than the same keypress from the keyboard -
a big win.

The MAXine (DECstation 5000/25 and /33) has a desktop bus similar to
ADB, and I believe the drivers there are implemented as seperate
drivers that talk to the desktop bus, so there is a precedent for
this.

			       _MelloN_


------------------------------------------------------------------------------