Given that the ordinary credit card processing of the time relied upon the embossed numbers on the card and something like carbon paper to transfer them to a paper receipt, this may well have been an improvement in resistance to forgery. For that matter, reproducing house or car keys from a photograph is not at all unheard of.

It’s just like a regular key. You can also read those out by taking a photograph.

Compared to biometrics this simply is gold as the key can be changed easily. Of course once integrated circuits became cheap you could simply build a card with a micro controller and sophisticated cryptography. Cryptography also didn’t quite exist back then.

Cryptography was alive and well since the late 30ies, especially in Briatin. I’m sure it was not forgotten by the 70ies. The idea behind that system is a different one: Lost or stolen keys could quite easily be declared invalid without changing the actual hardware. That’s not really viable with standard Yale lock cylinders…

If I remember correctly these key cards were two sided. One side would have the code for the device, so that side would be the same for all users of a device. The other side would have the code for the individual user, therefore unique for that device. Each code was dependent on making 1 to 4 contacts on up to 8 sets of contacts for millions of combinations. The problems was these things were actually kind of difficult and messy to make and it could be expensive replacing a single key if lost, stolen or just worn out.

JMyint, I’ve probably mixed up my combinations and permutations, but… 1 of 8=8, 2 of 8=56, 3 of 8=56, and 4 of 8=70 for a grand total of 190 key codes. It must have been more than that.

George – that’s eight 4-bit numbers, or 32 bits in all (4 billion plus.) But they probably didn’t allow all-zero combos, so 2.5 billion or so. Add some check-bit/crc and it’d down into the millions but still a lot.