PCOS is often looked at and talked about as if it were a disease, but IMO classifying PCOS as a disease is not accurate. This is because PCOS is actually a syndrome. A syndrome is basically a bunch of simultaneously occuring related manifesting symptoms, not a disease. A syndrome is something that results from disease-causing agents, it isn't a disease but it is caused by a disease (sort of how to have HIV is to be diseased, but AIDS is the syndrome that results).

A woman is said to have PCOS when she exhibits symptoms of androgenization, in absence of any obvious cause (say endocrine disease like cushing's, or steroid abuse, something like that). The androgenization has to impair fertility, which is why the emphasis is mainly on ovulation (the cysts on the ovaries are caused by failed/incomplete ovulation), however other symptoms of androgenization are usually required present before diagnosis is made, such as excess hair growth, acne, hair loss on the head, etc.

I harp on the disease/syndrome distinction for a reason. I fully believe that the label PCOS is entirely inadequate. The diagnosis of PCOS is made exclusively on a bunch of symptoms, but not on any pathological disease. Naturally this can cause lots of confusion for patients and researchers.
I think PCOS is much like diabetes. Up until suprisingly recently doctors didn't know the difference between type 2 diabetes and type 1 diabetes... all diabetics were diagnosed soley by their symptoms of having high blood sugar and general inability to use sugar. Doctors soon discovered that there were actually two types of diabetes, and they were completely unrelated to each other in pathology. Type 2 is caused by IRS. Type 1 is much less common and caused most commonly by auto-immune disorder/disease which causes the body to destroy its insulin producing beta cells. Type 1 usually occurs in early childhood (but not always) whereas type 2 takes longer to set in and occurs in adulthood (again, but not always especially now that carb consumption is way up). Type 1 diabetics have no IRS, they merely lack insulin because they can't make it. Type 2 diabetics on the other hand were born with relatively "normal" capacity to make insulin, but because of a biological susceptability or poor/uninformed dietary choices this results in their body to be unable to respond to insulin due to IRS, thus causing type 2 diabetes. Sometimes type 2s can overwork their insulin producing capabilities and become insulin dependent like type 1's, but it's not by or for the same cause.

PCOS, like diabetes, is a disease defined by symptoms but not any disease causing agent. Just as in the example of diabetics above, classifying a syndromes as diseases is going to have problems. I don't think all PCOS women have it for the same reason, and I believe in the future there will be different "types" of PCOS as there are different types of diabetes. The most common type of PCOS is probably caused by IRS due to the number of women who improve on LC diets. Though most women who have PCOS symptoms are insulin reistant, curiously not all women who have PCOS symptoms are. It also seems that not all PCOS women respond to any particular type of diet or insulin-sensitizing medication better than the other. This tells me there is strong evidence supporting the potential for as yet unidentified diseases that can result in PCOS, which may be totally unrelated to IRS.

However, because of the nature of our diets (they are VERY poor in composition and replete with disease-causing agents), it is unsurprising that most PCOS women are actually suffering from IRS... the PCOS is merely a symptom of uncontrolled insulin resistance. There seems to be an especially strong link between PCOS and tendancy towards hypoglycemia. It seems that whatever results in the high insulin levels causes both the hypoglycemia and the PCOS. The disease causing agent is actually high glycemic load carbohydrate food. To resolve IRS-related PCOS, total carbohydrate intake should be minimized, especially higher glycemic index carbohydrates.