Bart Schaefer wrote:
These two foci are in conflict. For Rich, it's obvious that end users
have limited expertise in distinguishing undesirable traffic; and to
allow the users to express their opinion, he must first allow that
traffic to pass, which is unacceptably dangerous, possibly disruptive,
and violates the "as early as possible" constraint. Only an absolute
classification helps to solve Rich's problem.
Your description actually means that nothing can be blocked without a
user first saying it should be blocked. IOW, the foci you've described
can't block, ever, because the user hasn't told you yet it's bad, and by
the act of seeing it, the choice to block or not is moot.
That's obviously silly.
Remember: TiS buttons are intended to catch those things that your
filters didn't _already_ capture. They're to refine your filters for
future repeats of the same thing (however "same" is defined). _Not_
_be_ the filters.
If the filters can tell something is bad, you block it, no matter how
you've determined it to be bad. No TiS hits necessary.
TiS is an _adjunct_ to your filtering strategy. Not the whole thing.
Rich may be overly dismissive of Mike's problem, but to declare concerns
about maintenance and security to be anachronisms not part of the "actual
job" is to repeat the mistakes of the past.
Mike is not dismissing it. I certainly ain't, because I have to
maintain and secure the darn thing. TiS buttons certainly don't
compromise our maintenance and security.
Asrg mailing list