Using a DDS would require devising some sort of computer control over its output frequency. The same control computer could also be used for T/R, keying, and any sort of bandswitching that needed to be done, as well as provide frequency control options such as memories, T/R offset, and RIT. Computer control could be through the parallel port of just about any old computer, or through a PIC chip.
The more we talked about this project, the more we began to realize that we were talking more about a philosophy of interconnecting radio building blocks than just a project about a VFO. In fact, our philosophy could be employed using a crystal oscillator, a toggle switch to handle T/R, a straight key for keying, a direct conversion receiver, and a simple RF output amplifier. Although limited to a single frequency, the resulting system would be a reasonably well functioning transceiver. Throw in the DDS VFO and its computer control requirement, including the ability to provide features such as T/R offset and bandswitching control, and the transceiver could now function more adequately, and (depending on the RF output amplifier's frequency range) do so on a variety of HF bands. The various modules could be built and upgraded one by one, and upon completion, each module could be put to use with the rest of the station, right away. Designers could design individual modules using a known interface, and everyone's designs could be used together. We've decided that we'll focus on defining this interface philosophy and designing some of the basic building blocks, then see if others are interested in helping the idea grow. We believe it will be a satisfying, efficient, and effective approach to homebrewing.