Job activity monitoring

[utzoo!duke!decvax!ucbvax!clyde at UTEXAS-11]

	I agree with Lauren's comment on this matter. There is one thing
that we have done that minimizes the inactivity problem somewhat and
that is to install an input wait timeout in the shell.  This is settable
(and can be disabled) by the user, but very few people change theirs.
	Our shell gives a one minute warning, then if no input has
been given it, exits.  This of course only works for people at shell
level, but most piddling seems to be at that level.

Clyde
-------