On 8/30/06, Steve Summit <scs(a)eskimo.com> wrote:
You forgot
// If you make a copy of $db on your laptop for testing,
// we will flay the living flesh from your bones
and more importantly
// We are holding everyone on the planet hostage to ensure
// that they never make backups of $db, but if something
// happens to us, you're on the hook for that, too.
Well, you'd have to give instructions to anyone with shell access,
too, and keep careful track of all . . . okay, but it would be roughly
possible. We already do it for checkuser IPs.
One of us is being very obtuse here, but I'm not
sure which.
Neither am I . . .
If I'm trying to count "approximate number of
people who have
viewed the page at least once", without possibly overcounting
people who have viewed the page multiple times, then depending on
the numbers it may be true that the server probably doesn't have
to, but under other circumstances it might very well have to,
and it's certainly strange to say that it "can't".
If you're only sending one view in every X to the view-counting
server, then it doesn't know how many times you've viewed the page.
It therefore can't exclude multiple consecutive views by the same IP
from its count. Yes? No?