[...] broadcasting at frequencies above 100 MHz will traverse the ionosphere and continue into space. The earliest television signals have already reached several thousand star systems. However, the strength of TV signals at light-years' distance will be low, given the small gain of the transmitting antennas. For VHF broadcasts, the maximum effective radiated power is between 100 and 300 kilowatts, and for UHF is 5 megawatts. At 100 light-years, these will produce signals of flux density no more than 10' - 10-31 watts/m'-Hz, even in the very narrow parts of the band where the carriers are located. The best SETI experiments today are seven orders of magnitude too poor to be able to detect a comparable signal. And note that retrieving the video components of a TV broadcast would require 10* greater antenna collecting area than required to find the carriers.
[...]
In other words, to assume that leakage automatically generates a "reply from Earth" to any SETI signal we might receive is unrealistic. Consequently, it's useful to seriously consider the general nature of signals intended for deliberate communication between star systems, as these might (1) elucidate the construction of any future replies to extraterrestrial transmissions, and (2) help to gauge what sort of signals our SETI experiments might discover. In this paper, we consider some realistic limits on information content that can be easily sent across interstellar distances via light or radio, and suggest what might be reasonable signaling strategies.