CO detectors are not, in my opinion, something that should be mandated.
I dislike combination CO/smoke detectors, as they address different problems. In my opinion, this means they should be used differently.
Smoke detectors have nearly 30 years of employment, and their use is widely understood. There is, in the case of smoke detectors, something to be said for locating them where they are likely to be near a source of fire. (We place them atop staircases because smoke will be "drawn" there.)
CO detectors have a different use all together.
First of all, CO has a cumulative effect. It also, as Ryan pointed out, is essentially the same density as normal air.
Unlike smoke, CO is normally produced, and diluted to safe levels by the room air, or vented.
A major principle in alarm installation is to prevent false alarms. Placing a CO detector in a furnace room might seem like a good idea- but is sure to false alarm every time the main burner ignites (the large flame bloom produces a momentary 'bubble of CO, thet is quickly diluted or vented).
In my home, the smoke detector is just outside the bedroom; the CO detector sits on the night table, next to the bed.
I am primarily worried about CO that I might breath, while I am asleep and unaware. A review of the literature can endorse this location; while such a location would be completely wrong for a smoke detector.
There is also the matter of different alarm life-spans.
A photoelectric smoke alarm can last forever.
An ionisation type smoke detector has lost most of its' sensitivity after ten years.
A CO detector has a definite lifespan, and will not operate after five years.
CO detectors will also alarm if:
-exposed to natural gas; or,
-exposed to freezing temps.
The NFPA has a "standard" for CO detectors, which is essentially useless. The standard says "follow manufacturers' instructions.' Keep your $25.
