The DSM is not the "bible" of psychology, but psychiatry. They are two completely different areas of science, and the therapists I've talked to over the years have all, more or less, been very useful for getting a handle on my depression and understanding myself and why I operate like I do. I know a tremendous amount about this topic, having gone through several nasty bouts myself - and having a B.S. in psychology (and computer engineering, as well, which is why I'm on HN)
Psychiatry, however, IMO, is for the most part, useless. Psychiatrists should only stick to dealing with things that have a predominately biological basis, like schizophrenia or bipolar disorder. Garden variety stuff like the "common cold" form of depression should be dealt with using talk therapy. Every psychiatrist I've talked to in an attempt to get a mild antidepressant has resulted in frustration - they are all arrogant SOBs as far as I'm concerned who want to stick you onto the heavy duty addictive crap before trying the easier to deal with drugs - as an engineer, my attitude is try the simpler solution first and only if that doesn't work, rewrite bloody everything (if my brain was a software program...) So I've had to self medicate - I found research indicating the particular subtype of depression I suffer from would be best treated with bupripion, but the last doctor ignored me... so now I am going to pretend I'm a smoker and buy Zyban over the internet to get what I am sure I need. Or I could use nicotine patches as a mild antidepressant (according to the latest research). As a hacker, I see nothing wrong with hacking my body :)
But adding a drug into the system known as "me" isn't going to magically fix everything - I have to change my behavior as well, which is harder and that sort of thing require understanding psychological principles, not psychiatric blather about neurotransmitters being the root of all the problems. It's like people want to fix a software problem by replacing chips on the motherboard! And they aren't even sure exactly what chips!
Scientology is a crock of BS, but they do have a good point about psychiatry being a problematic profession. I would never discard what I've learned through psychotherapy, but drugs... treating mental illness strictly as a medical / mechanical issue is very shortsighted.
The trend over the past few decades has been an increasing emphasis on the biological aspects of 'mental illness' - when you examine the history of the two fields, you can see a swing between the two extremes - e.g. when Freud's theories started getting attention, an increased emphasis on the psychological aspects of mental issues became prominent in society and it become fashionable to be seeing a psychoanalyst.
But psychoanalysis, like that promoted by Freud, isn't that helpful for the most part in solving problems, so the trend started swinging back to biological especially after research etc started uncovering the biological aspects of mental illness (serotonin, LSD research, etc). Thus the 'magic bullets' became drugs and since they are cheaper and more profitable than talk therapy (and "easier" for the patient as well) that became the emphasis.
For some reason, viewing the whole issue as the interaction between psychological issues and biological factors isn't viewed as important - as an engineer who understand this stuff better than the usual layperson (though I'm not an expert) it is frustrating because it is so obviously just good engineering to view a person as a system - biology matters but so does environmental issues (family history, current life of the patient...) Drugs are cheaper and more profitable...
But yes, the DSM has gone quite downhill over the years in its usefulness. So if you change, in your post, "psychology" with "psychiatry" I won't down vote you.
And if one's depression is caused by environmental issues, like work related stress, or family, or "meaning of one's life" - they are not going to go away once you start taking the right drug or zapping the depression away. And thus it'll only come back eventually.
Another big problem is that "depression" is such a fuzzy label - there are many subtypes and each need to be dealt with in a specific way. I came across a book a few years ago that correlated what was known about different areas of the brain and their function and how different drugs affected each differently - i.e. something leading to theories on what type of antidepressant would be best for a patient given the symptoms they were exhibiting. Thus my theory on why bupripion would be better for me than Paxil (which was useless). But like with the low fat vs low carb controversy of dieting, medicine is afflicted with fads and trends - just like IT (are all problems solved with NoSQL? of course not, but lots of people seem to think that. same with the virus known as Agile)