Preventing date rollback

[Marcus J. Ranum]

 Chuck.Phillips at FtCollins.NCR.COM (Chuck.Phillips) writes:

>[...] if the computer manufacturer were to add a non-resettable piece
>of mechanical hardware measuring the elapsed uptime of the computer and
>provide a system call for accessing the elapsed time, then at least your
>software could check for consistancy.

	Of course, if someone were to try to sell a computer with built-in
support for copy protection, nobody in their right mind would buy it because
a) the cost of the copy protect hardware would get passed on the the buyer
b) the pain in the a** factor of the copy protect would get passed on to
	the buyer. the cost of hardware failure in the copy protect hardware
	would be considerably higher.

	I know a large hospital that had a total computer lapse on several
critical systems because some vendor's dbms protection scheme didn't handle
leap years right and "expired" at the wrong time. Unfortunately, they didn't
lose the entire account as a result, but they bloody well deserved to.

	Copy protection's plain silly for anything other than PCs, since
any organization that has enough money to buy a, say, $50,000 machine
isn't going to be stupid (I hope) enough to risk the lawsuit that a
software vendor can bring against them. Deep pockets have a lot more to
worry about that the average PC user - if they're being dishonest. If
they're being honest (which a gentleman must assume until convinced
otherwise) they shouldn't have to suffer the insult of distrust - especially
from a service provider - what ever became of "the customer is always right?"

	On the wacky side, I'd love to see a hyper-dongle for a CRAY. ;)

mjr.
-- 
	How can something that is almost 2 Megabytes in size be called
a "kernel"? The truth in advertising laws should apply to software. Let's
call it a "gourd" or perhaps a "watermelon".
		[From the programming notebooks of a heretic, 1990]