A hundred and twenty-eight pages of reference on how something in the real world is measured and turned into a number: converter types, accuracy versus resolution, multiplexing, sampling theory, electrical measurements, temperature sensors, strain gauges, isolation, filtering and noise reduction. There is no credited author; it is vendor material, with the bias that implies.
What I took from it
The distinction between accuracy and resolution, which looks like textbook pedantry and settles real arguments. An instrument can display four decimal places and be wrong in the first. When a system receives a weight with plenty of decimals, nobody questions it; the number looks precise, and that appearance is exactly what needs checking.
The noise and isolation material gave me a rule I apply before touching code. If a peripheral returns odd readings only at certain hours, or only when another machine starts up, the problem is not in the software receiving them. Asking that first saves days spent reviewing an integration that was working fine.
And sampling theory leaves a transferable idea: measuring less often than the thing changes does not give you an incomplete number, it gives you a false one that looks true. That holds for sensors and it holds for management indicators taken once a month.
What I left out
Almost all the component and board detail, which is old and was written to sell. And it is worth being blunt: this is not a book, it is unsigned reference documentation, the kind you leaf through when there is a problem. I did not read it end to end and I do not think anyone should.