On Thu, June 24, 2010 11:08, Valdis.Kletnieks () vt edu wrote:
The answer to that kind of question is quite often related to the
industry average. For example no more failures than one standard
deviation below the industry average.
Ahh.. but that doesn't really help either. Consider that not all
are created equal. Should a failure to detect some unknown basically
strain that's only been seen on 4 machines in Zimbabwe count the same
failing to notice that a machine is still infected with Code Red or
that's virulent and malicious and on a very large current burn? Do
care it didn't detect the Zimbabwe strain your machine has never been
Of course any way of measuring it will be fundamentally flawed in
certain ways. There is always that pesky 80/20 or 90/10 rule. And you
can of course figure out a way of correcting for corner cases, but
that will only create additional corner cases. That's what makes
lawyering on product liability a craft at best and usually some form
of black magic.
For that matter, do you really want to create a situation where the
A/V companies now have an *incentive* to make sure their competitors
detect something (either by failing to share data, or resorting to
malware custom-crafted)? The only reason the whole A/V industry
And yes, there may very well be unintended consequences. Nonetheless,
I feel the era of complete exoneration from product liability is
coming to an end for packaged software. Especially in the security
industry. It is just a matter of an 'unsafe at any speed' moment
occurring and there will be legislation, however braindead such
legislation may be from an engineering viewpoint.
Call me a pessimist, but we've been putting way too much critical
stuff on internet connected systems while at the same neglecting basic
hygiene at every level not to have some disaster to happen. It isn't
so much a question of if but when that will happen.
Full-Disclosure - We believe in it.
Hosted and sponsored by Secunia - http://secunia.com/