http://www.8052.com/forum/read/178653

Trying to analyze any random chip isn't easy - or actually totally impossible. A LED driver or analog amplifier chip for example can on one hand consume huge currents and have the power pins at just about any place. But they would also need advanced programmable load circuits to behave as expected.

The scope I was discussing was autodetection of most logic chips in the 74xx and 4xxx families, where the user may have to assume that the chips are working with 5V supply voltage and that the user may either have to guess at the supply pins or hope that it's ok to use processor pins to drive the chip. And what I was saying is that the testing logic don't need to run 200 full sets of rules just because the rule database may support 200 different logic chips. Many tests may quickly reduce the number of potential chips. If toggling a single pin will change a different pin between high-impedance and activelly driven, you can suddenly ignore a huge number of chips.

Not too much difference from the hw setup of the originally lined device, but with the difference that it identifies the chip instead of just testing the chip based on an entered chip ID. That linked unit showed such huge lack of creativity.