As Dr. Newmark points out in his post, psychiatry remains profoundly different from the rest of medicine, where diagnosis is everything, in this respect. If a patient presents with chest pain, it is supremely important to know whether it is due to a heart attack, aortic dissection, bronchitis, pulmonary embolus, gastroesophageal reflux, or costochondritis, because each of these calls for clearly distinct treatments. Psychiatry is not like that. Deciding whether a person's diagnosis is depression, bipolar disorder, or schizophrenia will suggest moderate differences in treatment, but the latter will derive more from specific symptoms and personal background. This goes to show that psychiatry remains far more art (or "art") than science. The research-powers that be have yet to persuade the practitioner otherwise.
I agree with everything in his post, and my colleagues agree that the DSM is pseudo-scientific and pseudo-medical, designed for insurance forms and research purposes.