Exactly. A big part of medical/pharmaceutical software/system development is (formal) "Validation": is this device/process GUARANTEED to perform as specified? This is especially true as software/systems become more complex -- can you ever know, for sure, that you've verified every aspect of the system/software?
By way of comparison, the Therac-25 was trivial piece of software -- yet inspired most of this paranoia!
https://en.wikipedia.org/wiki/Therac-25
BT isn't the issue -- though any non-wired interface presents additional challenges (
e.g., what if there is "invisible" interference with the radio transmission and your "STOP!!!!" command gets lost?). OTOH, you could probably come up with a scheme for tying into the RNet control on the chair (
e.g., by emulating the seat controls in the PC).
I'd be more concerned about the voice recognition software misrecognizing an utterance. Or, failing to recognize it at all! I'm sure Dragon (or any other company) would disclaim any liability for screwups in their software in such an application (many expressly prohibit such use in their EULAs).
You also have to consider the reliability of the operating system (which is typically not certified for such types of applications):
http://www.coneslayer.org/wopr/mph/h.../msg00026.html
If the DoD doesn't have enough clout to get those sorts of "guarantees", what makes you think a chair manufacturer would?
