"phil hunt" wrote in message
. ..
So the frequency changes are pre-determined on a time basis?

Oh yea.

If there is a radio receiver, is it better able to detect/deceive a
signal whgen it knows the frequency in advance? Or can it "sniff"
for lots of frequencies at a time and pick out what looks
interesting?

Both.
To "sniff" takes the time needed for a bunch of transmitted
cycles to come through so the receiver can determine it's
not random noise.
The receiver for a hopper *assumes* there is a signal in the
expected slot and integrates it into a bigger signal for the
rest of the system.

If two receivers, placed say 10 m aparet, both pick up a signal, how
accurately can the time difference between the repetion of both
signals be calculated? Light moves 30 cm in 1 ns, so if time
differences can be calculated to an accuracy of 0.1 ns, then
direction could be resolved to an accuracy of 3 cm/10 m ~= 3 mrad.

Measuring the time difference between reception by two antennas
yields a curve (all points such that the distance between the two
antennas and the point is a constant). Add another antenna and the
possible transmitter locations are the points where the curves cross;
by doing the A-B, B-C and A-C comparison you should have a single
point left.
If your antennas are too close together the curves stay in proximity
to each other so long you don't have the angular resolution to get a
good fix.

You can measure the angles quite accurately by using multiple
directional antennas and measuring the phase & amplitude differences.

Alternately, would something like a pinhole camera work? What I mean
here is: imagine a cubic metal box, 1 m on its side, with a vertical
slit, about 1 cm wide down one of its vertical faces. On the
opposite face, there are detectors for detecting radio waves. If the
elevctromatnetic ratiation coming into the box can only go in
through the slit, and goes in a straight line, then knowing which
detectors are lit up would allow someone to tell where the
radiation was coming from. It may be that, depending on the
wavelength, the incoming radiation would be diffracted by the slit
and would get spread all over the detectors. If this is the case,

It all most certainly will defract.

perehaps multiple slits could be used, and the diffraction pattern
would differ dependent on the angle with which the radiation strikes
the slitted face? (because the radation at each slit would be
out-of-phase with the radiation at other slits). Has anything like
this been tried?

Sounds good, unfortunately your detectors hung on the wall still have
that problem with determining a short burst signal is really there and
not random noise. There's also the problem with which slice of the
spectrum they are listing too at any one time.
There are ways to do the listing with a really wide band but they require
boat loads of processing that's not done real time and none I'm aware
of preserve the phase information for DFing.