Telling scientists things is hard. They love to be experts, and can be resistant.

Even computer scientists. Had one come to me, looking for an API for fast data copy in RAM. I asked what he was achieving, he told me. I asked the bus width on the CPU and the clock speed. Told him, that's what it can do. He went away, said he remembered there was a library routine that could do better and he would keep looking. I just shook my head.