<Non electronic digital _computers_ are a lot rarer. If we decide that a <computer has to have a sequence of operations and some kind of <conditional branching, I can't think of any at the moment. Mechanical <sequencers, based normally on camshafts are common, of course. You kidding, air logic has been around for years and frequently used in atmospheres or other considerations that would prohibit using electronic or electically power controls. The basic elements such as switchs valves and oneshots all exist and can be combined into digital functions. I know I demonstrated a system where several inputs (switches) had to be true before a sequenced action using two bistable elements and a oneshot. It's not fast but fun to watch! <Are you thinking of 'Digital circuits are built from analogue parts' ? Not a valid concept. both OR and AND gates can be done using totally non amplifying devices (diodes). It was Vonda that postulated that digital was analogue with a precision of two states, true and false. The realm of analogue is one of infinite precision but possibly of limited accuracy. The digital realm is one of limited precision and absolute accuracy. <As a practical point for this list, if just about all (analogue) circuits <are analogue computers, then they are on-topic here :-). No I don't As a final thought, analogue circuits often perform complex functions. This very aspect and the realization that precision and accuracy are very differnt commodities are why some fuction are easy to do using an analogue approach and can be very difficult to do digitally. Allison