Insidiously, more and more of the routine functions of scientific experimentation have been computerized and even *automated*, and that's actually very bad for #science. The overuse of #software and computer simulation in science is driving a wedge between the scientists and the direct perception and observation of the physical world which is the bedrock of the scientific method.

If some *computer program* is doing all the actual observing and interpretation of data and so forth, then how much confidence and trust can one place in the scientist's conclusions based on that computer work? If there's some terrible error or systemic bias being imposed by the computer processing, how is anyone going to know, especially with opaque corporate software?