No, but you're missing the fact that direct input bypasses the TI-OS entirely; I need to come up with a way to detect when direct input is (or is not) occurring and handle things appropriately. I guess I could do it the opposite way, by doing something to some ports (via a keyhook) when getKey/getCSC is being used, and assume that at all other times I should do direct input. The microprocessor connected to the new keypad would just have to watch for the particular signal.
Edit: On the recommendation of a lab tech at my university, experimented with fiberglass as a possible material from which to build the body of the device. Tried cutting out a key-sized hole and putting a key in it to test how easy the material is to work and if it would bind with the key. The results are below, and seem quite positive; you can see the ghostly shape of a membrane pad underneath the key.
Edit: On the recommendation of a lab tech at my university, experimented with fiberglass as a possible material from which to build the body of the device. Tried cutting out a key-sized hole and putting a key in it to test how easy the material is to work and if it would bind with the key. The results are below, and seem quite positive; you can see the ghostly shape of a membrane pad underneath the key.