As a paramedic, the "DNR" thing is part "folklore", part reality, but not for the reasons you cite.

If CPR is ineffective, the result is death. Dangerous? Not particularly.

Most people wearing DNR tags (aside: even a DNR tattoo on the chest is not considered legally binding against "heroic measures", unless signed by a physician) do so to provide some measure of what the patient might want (due to life expectancy, illness, etc).

Survival rate to discharge is HIGHLY dependant on many, many variables - I mention elsewhere that in certain parts of the country, due to concerted education and well managed/funded EMS systems, survival rates can approach 50% to discharge. You are right, though, successful CPR isn't going to negate the underlying cardiac condition.

If CPR buys a patient a year to five years, the pain of broken ribs is probably fairly minor in the grand scheme of things.

There are, as mentioned, a whole host of advances. My home EMS system does some fairly deep analysis of all CPR attempts and works to revise protocols, including what drugs to administer, through to 'continuing compression through defib shock' (studies done on medical students showing it to be a viable possibility).

Ultimately, education is the key - the ability to get skilled compressions more quickly on the chest, continuing to perfuse the brain and other organs has a huge impact in survivability, as well as damage done.

It's not the CPR that's particularly damaging. Yes, ribs are dislocated (though rarely broken), and cartilage is torn. Those are fairly minor complications though. The 'damage' is done by the medications we push. Drugs like Epinephrine are very good at getting your heart to beat again. A recent study in Japan suggests that that may be at the expense of your brain, and other vital organs. Other drugs commonly used in ACLS resuscitation are equally nasty (Amiodarone, for instance).

You're absolutely right that we've gotten very good at getting a heart to beat for a few more days until it succumbs to the damage done to the rest of the body (that 40% figure you mention, though out-of-hospital rates are still much lower than that). The exciting research being done is to improve long term outcomes for cardiac arrest patients.

More and more research is showing that the key to a good long term outcome is early compressions, effective compressions (push hard and fast), uninterrupted compressions, and early defibrillation (with an AED or manual defibrillator). Over the past 10 years, resuscitation algorithms have been streamlined to minimize 'off chest' time. Compression to breath ratios have changed (from 5:1, to 15:2, to 30:2). A great deal of emphasis has been placed on keeping the rhythm checks that happen every two minutes as short as possible. We're now teaching people to continue compressions while the defibrillator is charging (and under ideal circumstances, you could easily continue compressions throughout the shock). Intubation is performed without stopping compressions.

The focus of all of these changes is to decrease the number of time compressions are stopped. Every time compressions are stopped, it takes some number of compressions to get the pressure built back up in your circulatory system. Fewer interruptions means fewer 'wasted' compressions, and less time that the brain isn't getting perfused. This is obviously key to a good neurological outcome.

We like to think all our fancy drugs and advanced interventions are doing something effective, but in reality, 'basic' treatments like CPR and defibrillation are what really save lives, and more importantly, save brain.