Well, let me say this.

The technology marches on, and not all of the advances come from corporate R&D departments.

Consider th G4GUO protocol. Charles Brain (G4GUO) developed, along with a digital voice mode for use on HF. It uses a set of 36 carriers spaced (IIRC) 50 Hz apart, using QPSK to modulate each at 50 baud, 2 bits per symbol, giving an aggregate throughput of 3600 b/s. 1200b/s of this is used for FEC, the other 2400 for actual data, which is encoded using the AMBE codec. The result is slightly mechanical, but otherwise rich, clean-sounding digital audio on HF. As with PSK, a transceiver set to USB is used, basically causing the transceiver to function as a very narrow-banded transverter. Google for his callsign to find it.

PSK itself is also a good example. There is no secret about how it works, and there are basically four levels of users therein: (1) those who buy an interface to connect their radio to their computer, load the software and go (these are your appliance operators), (2) those who build an interface to connect their radio to their computer, load the software and go, (3) those who build the interface to connect the computer to the radio and then implement the PSK protocol themselves in their own code and (4) those who use the computer as a terminal to access PSK through some sort of homebrewed modem. (There may be some variations on each of these levels, but you get the idea).

Incidentally, an idea was expressed to me by one of my professors in college. He said, "In a perfect world, hardware and software are logically equivalent." Of course, this falls down at a certain point (can't run software without hardware or vice-versa) but has a certain validity, also (most tasks can be implemented either way). As such, there is much to be said, especially where digital modes are involved, for software development leading to innovations. As computers and programming languages become increasingly powerful with time, I think you will find them leading to new modes at an alarming rate.

An interesting thing here is the GNU radio project. These folks are using DSP boards with 20MHz wide D/A and A/D convers on them, and implementing AM, FM, SSB, etc. in software. With an appropriate transverter, they were able to create an FM tuner program that would (as a proof of concept) tune two separate radio stations from the FM dial simultaneously on one set of hardware. How many channels would *you* like to watch, without the lage that occurrs from conventional scanning? How about *instant* changeover to your priority frequency?

Let me close wth an analogy.

My father in law is very fond of pointing out businesses that have forgotten what their business is. He points to the railroads and says that they think they are in the train business and have forgotten that they are in the transportation business (that goes double for Amtrak).

I think that we, as hams, need to remember that we are in the communications hobby, not the radio hobby. That we develop an affinity for radio in particular is fine, but I think (and this is likely to be controversial, but so be it) that we need to include the many computer enthusiasts who have revolutionized communications amongst our own. They're among us, and we are among them; they are us, and we are them. Want proof? Just look at the fun that computer enthusiasts are having with directional antennas to extend their 802.11 range, and the fun that computer/ham hobbyists are having with new digital protocols. We are all one.

I think some of the differences you're trying to point out here come from a shift from being able to manipulate technologies.

Let's look at one of the most popular HF rigs of the past, oh, 40 years--the Heathkit HW-101. They sold thousands upon thousands. All tube (er, excuse me, thermally generated field effect transistors in a sealed glass envelope), point to point wiring, two, maybe three weekends and you were on the air.

Let's look at the latest generation, exemplified by, say, an IC-7800, which has DSP up and out the wazoo and outperforms anything I've ever had my hands on. Not even close.

Even your IC736 probably makes extensive use of surface mount technologies, and exotic (compared to back then) microprocessors and dedicated semiconductors designed expressly for these uses. We've come a long way from the days of 807's, my friends?

So is homebrewing dead? Not completely (witness the K2's), but home rolling something from scratch that will compete with what you can find at your local 1-800 emporium.

The innovation among us for ham radio will be not in designing new hardware technology, but in manipulating it; tools like PSK, etc, which allow us to use these technologies in new and exciting ways.

In the end, though, ham radio boils down to this: You have someone with a box full of various components and a piece of wire who is able to communicate with someone else who has another box full of parts and another piece of wire. Do we focus on the boxes, the wires, or the communication?