The ultimate solution is an adaptive system with btu monitoring.
It works like this... you have a computer that actually manages your heating system. Each zone has a btu monitor, which consists of a flow meter (just like a water meter for those with town water, but designed to withstand high temps) and a temp sensor on the supply and return for that zone. Outdoor reset is simple and inexpensive, but what we actually care about is the number of btus being used... the outdoor reset curves are just approximations.
Now, we know how many btus are being used by each zone, and we have a room temperature sensor, and an input device that lets you tell the system what temp you want that room to be. So, the computer sitting in the basement says, "hmm... the temp in the room is falling by 2 degrees per hour, and will drop below the desired temp in 15 minutes," and starts circulating water using a variable-speed circulator. It knows from the last time that it heated this zone, that the zone absorbed heat at a rate of 15kbuth, maintaining the desired temp. So, it looks at the supply water sensor, and adjusts the circulator flow rate to deliver 15kbtuh.
This time, however, at it watches the room sensor, it notices that the room is not heating as fast as it did last time. So it anticipates that the remaining heat in the boiler will not be enough, and fires up the burner, while increasing the flow rate to use the current water temp more effectively, until the burner has time to heat the boiler. Now it is modulating the flow rate to the supply water temperature such that it is dumping 20kbuth into the zone, and the temperature of the room is behaving as-desired in order to keep it within 1 degree of the desired temp. It decides that the room is warm enough and shuts down that zone, storing the 20kbtuh figure for the next time.
With a system like that, it doesn't matter if the heat loss is due to cold outdoor temperatures, or high winds, or because someone is painting the room and left the window open without turning the "thermostat" down. The system will adapt to the conditions, as defined by the behavior of the heating zone, itself.
To go one step further, we add that outdoor weather station, giving the computer weather data (temperature, wind speed and direction, precipitation, insolation) and now it can compare the monitored behavior of the zones to the outdoor conditions. After a few seasons, it will have a very good idea of what effect the weather has on each heating zone, and will be able to anticipate behavior based on the weather, not just the last cycle. But it will still monitor and adapt during each cycle, in case someone does leave a window open, or someone is sick and turns the room temperature up, or someone cuts down a tree that had been shading a portion of the house, or there is a snow drift against one wall. Or whatever.
The same could be applied to cooling, as well, particularly with advanced chilled-water systems and variable-speed equipment.
Why don't we do that? The answer is, of course, money. That's a very expensive system, just to get a handful of percent more efficiency. You get a lot more "bang for your buck" with other upgrades (like better windows and insulation, and wood-burning equipment, and radiant heating, and such).
For larger applications (eg, schools, office buildings, warehouses, etc.) those few percent might be enough to justify the cost of a system like that, but the development cost would be high, and it would take many years for a company developing that technology (to the point that it is reliable, user-friendly, and profitable) to ever see a return on the investment, due to the low level of demand.
Joe
