I'm a med student and a programmer. I think there are a lot of similarities between these two disciplines. The body is an incredibly complex system that you o

I'm a med student and a programmer. I think there are a lot of similarities between these two disciplines. The body is an incredibly complex system that you or a pathological process can perturb in a variety of ways, just like any other system. The results of those perturbations may be evident immediately or may take a long time to present themselves. The results are output in the form of physical signs and symptoms. The way this particular system works is that there are more potential pathologies than there are physical signs and symptoms, so sometimes you end up with collisions in the pathology to sign/symptom mapping. In these cases you can use instruments like labs, radiological tests, and sometimes invasive methods like surgery to further investigate. These tests are imperfect and are themselves perturbations to the system. If you can come to a final diagnosis of the underlying pathology using the tools at your disposal (history, physical exam, labs, tests, and procedures) you can then engage in a therapy. The more precise your diagnosis, the more precise your therapy can be. Again, therapies are also perturbations to the system. Ideally these perturbations move the system back into its normal functioning state though they may also cause undesired results that must then be diagnosed and treated. This entire process takes place within the context of a social interaction that can itself help or hinder. Both doctors and patients have their own social idiosyncrasies that may or may not match up well with each other. I believe that most physicians operate by generating a gigantic hash table. Essentially they know several patterns of physical signs/symptoms that act as the key. The value is whatever knowledge they have of that disease process. As medical knowledge progresses some physicians update their hash table keys/values with new information, most don't. Almost all physicians update their hash table (i.e. learn) using information gathered in the process of seeing their own patients and recognizing the difference between the predicted outcome and the actual outcome. This is often a subjective process. I think it is because of this approach that most physicians find it hard to believe that medical diagnosis and treatment can be highly systematized and is at least partially such an obvious idea even needs a study to support its use.