Professor Richard V. Andree - better known for his work on promoting math and computer science education after he retired - was a computer scientist, mathematician and cryptanalyst. As an undergrad, I worked with him on a number of projects in several fields, as well as taking courses from him. The courses had a variety of titles and subjects, but what he really taught - in all of them - was computer-assisted problem solving. I learned a great deal about problem solving from him, and could write about it for pages. I want to mention one lesson - summarized by the quote "It's much easier to prove something after you know it's true" - because that resonates with something much more recent in my life.
During the past five years or so, I've become enamored of a type of TV show I call the not-a-cop show. They all have the same theme: some expert in a field other than law enforcement winds up partnered with someone in law enforcement to help them solve crimes. There were probably a half-dozen or more on at any given time during the last five years.
What I recall as the first of them - and in my opinion still the best, though no longer in production - was Numb3rs. It involved an FBI agent dragging his mathematical genius brother Charlie into cases to do, well, math. One reason it struck me was that - at least during the early seasons - they explained the math in terms a layman could understand. Even better, the math was real. In some cases, it was work I was already familiar with. In others it wasn't, but if I checked it was real. I didn't check all such cases, but every one I did check was real.
Another reason it struck me was that I could see the lessons I learned from Dr. Andree being applied by the characters in the show. For instance, once Charlie knew something was right, he stuck with it. If the answers came out wrong, he didn't start over - he went looking for the mistake in the work he had done. Clearly, Charlie has clearly taken the quote from Dr. Andree above to heart.
After I graduated, I worked with Dr. Dwight Pfenning, one of the worlds foremost experts in flammable fuels forensics. The work done by this group illustrates how well the not-a-cop show Bones has captured what it's like to do forensic science. Yes, there are differences - since it involves a forensic anthropologist, the Bones group usually has a body, and their problems are figuring out how to extract all the parts of it without destroying any evidence. In dealing with flammable fuels, we mostly dealt with civil and not criminal cases, and in those rare cases where there was a victim, they were maimed instead of dead. More importantly, our material evidence had probably burned or blown up. So instead of trying to preserve it, we were trying to recreate it. Fortunately, we usually had architectural plans and engineering designs for the critical pieces. While in Bones they occasionally got to try and recreate some aspect of the event, that was our normal mode of operation. Recreate whatever had burned or blown up, instrument it, then burn or blow up the recreation to recreate the results of the incident under investigation - and then describe them in detail because we had instrumented things. Where the two overlap, the show feels right. Watching the characters on Bones try and figure out how to measure something reminds me of going through that same process with our tools.
A recent episode of the not-a-cop show The Mentalist had a "scientist" in a guest role. Unlike the previous two shows, the protagonist of The Mentalist is not a scientist. He's, well, a mentalist. He uses his keen observational skills and deep understanding of human nature to unmask the villain, or get them to unmask themselves. The "scientist" depicted in this episode is typical of media depictions of scientists from my childhood, with a "the numbers will show the truth" attitude that science as a whole dropped with the uncertainty principle. Worse yet, she doesn't "do the science." She takes evidence at face value without testing it against her hypothesis, doesn't compare values between alternatives to decide if a conclusion is correct, etc.
It's surprising how right the depiction of scientists at work in the better shows feels when compared to real scientists at work. Especially when I look at how most shows got it so wrong until relatively recently.