Bob McGwier is an interesting fellow. I met him at Dayton many years back when TAPR was just getting into DSP devices: http://www.arrl.org/news/tom-clark-k3io-to-speak-at-amsat-ta.... I think he was doing work for a defense research outfit at the time.
Getting way off the topic, some of the events at Dayton are quite interesting. At one dinner, I got to hear Joe Taylor https://en.wikipedia.org/wiki/Joseph_Hooton_Taylor,_Jr speak about his work on pulsars. He brought some audio clips of pulsars, the first gave a "click" about once a second. The second clip was at ten per second, and the final one produced a tone of 600 clicks per second. He credits his interest in physics (along with his brother) to amateur radio as a young boy.
As a note on how computing power has improved since 1989, we can now remove impairments and decode optical nPSK at 32Gbd/s ~155k Symbols (2-4 bits per symbol in 2 polarizations) in about 8s using offline processing (data captured from a device like high sample rate sampling scope).
> A few years later when half the bad guys in the world seemed to be using AX.25 and cheap ham gear
Also, note that this awesome research was paid for by the NSA via their sponsorship of the lab and Cray machine. I assume because FSK decoding is useful for SIGINT...
I'm thinking the dates on this are off a bit, as the Pentium Pro came out in 1995: https://en.wikipedia.org/wiki/Pentium_Pro
The first PC I ever built was in 1997, and had a dual socket Pentium Pro logic board which was great for Linux, not so great for non-NT Windows.