In article <ndtcj09i8agkbhi8mt83mbrmo844l7aatv@4ax.com>,
10JQKA@nospam.com (possibly bogus) wrote:

If the SETI equipment was pointed at the earth from out in space
would it detect signs of intelligence or would it be looking on all

It would probably pick up signs of intelligence, but on the planet
doing the observations.  Note, though, that S@H uses a very restricted
frequency range compared with even Berkeley's own full SERENDIP coverage.
The SETI Institute's Phoenix covers yet more, but a lot of frequency ranges
have to be excluded because there is just too much local noise.  (S@H has
to consider the risk of "users" getting overenthusiastic and trying to
go to the press over what is really a local detection.)

The definition of "clear" in PR material is also confusing.  Below about
1GHz, natural noise sources get strong.  Below about 30MHz, the ionosphere
becomes opaque.  Above about 20GHz, the atmosphere becomes opaque and
atmospheric noises increase.  This is the wide definition of the
waterhole.  S@H searches in a range reserved for radio astronomy, around
the neutral, atomic, hydrogen hyperfine line, and therefore clear of man
made sources (except for spurious signals).  PR material often doesn't 
distinguish between the low noise window and the narrow range around 
1.42GHz.

the wrong wavelengths?   If it would detect a signal, from
approximately how many light-years could it detect it?  

This is extensively covered in the FAQ.  Planetary radar and deliberate
beacons are detectable to the order of 1000 LY, using current earth
based system transmit powers.  S@H will not detect TV carriers at the
limit of the solar system.

Also as our airwaves become more crowded we are using better and
better compression to maximize signal density.  If we take that to
extreme, how good of a compression/encryption would need to be on a
signal before it would just look like white noise to the SETI
detectors?  Is there a way to test for this? 

Current levels would probably frustrate leakage detection, which is
why you will find that most discussion here is about the detection of
signals intended to be detected, and special cases, like CW planetary
radar.