What was talked about a while back (and I believe what has been done with iPod mini microdrives) is locking the drive firmware into "True-IDE" mode. I'm just speculating that this is what Hitachi has done, but it's consistant with the reported behavior and it's what I'd do if I were the Hitachi engineer tasked with stopping microdrive canibalization.
The CF standard actually defines three different types of card interface: I/O, memory, and True-IDE. The primary difference (without getting into too much detail) is how the 50 physical CF signals are utilized. True-IDE is a remnant of the early PCMCIA days and was designed in to make it easy to build a "glueless" PCMCIA to IDE adapter.
Solid-state CF memory cards and retail microdrives provide both the "memory" and "True-IDE" interfaces. Memory is the simplest so that's what the hardware in most digicams and PDAs is designed to support. True-IDE mode requires some additional hardware logic and a matching device driver. Most laptop PCMCIA controllers support True-IDE, many USB-CF adapters do, very few digicams do, and no StrongARM/Xscale PDAs that I know of do.
So by tweaking the drive's firmware to disable memory mode (but leave True-IDE mode intact), you render the card inoperable in most cameras and PDAs. It will still work fine on a host with True-IDE support - but how many people are going to rip apart a $250 MP3 player just to add a measly 4GB to their PC or laptop?
Don't expect a workaround any time soon as it really is a hardware limitation.
