What do we want a definition of planet to be, or more importantly not to be? Knowing what we want is the first step towards a new definition... The problem is that, even though our notion of planet comes from a historical, humanocentric background, we are trying, in the prospect of our quest of absoluteness, to create a definition that is compatible with both the historical and the scientific views. But no "simple" definition will manage to do that and satisfy a majority.
If we want to do both, the historical part forces us to create a definition that categorizes the 9 objects we call planets nowadays as planets, but labels no, or only a very small number of objects, as planets. As of wether Pluto should be considered a planet or not, rewriting books and declassifying Pluto as a planet may, arguably, be a more or less reasonable idea in theory, but is definitly not in practice, for there is an incompressible and large number of people that are attached to the current classification.
Usually, we associate the notion of planet with an object that orbits around the sun and is "rather large". But in fact, "rather large" depends on the distance of the object to the Sun. For example, if Mercury was 2 times more distant from the Sun than Pluto, we would probably not call it a planet. But if Jupiter was 3 times more distant from the Sun than Pluto, its sole size would assure it the designation of planet. Besides, we could base our classification only on the size of objects, but it would allow any object the size of Pluto at the very improbable distance from the Sun of 10 times that of Pluto to be called a planet. And we do not want that, because we could one day discover a whole bunch of those and we want to limit the number of new planets to only a very small amount.
All this indicates that we have to create a classification that takes distance from the Sun in account, because it is the only means of limiting efficiently the number of eventual new planets: we can be reasonably sure that we will not discover new objects bigger than Pluto that is less distant from the Sun.
So a new definition would specify that a planet's orbit's center is the Sun, that it is "large enough" and "close enough" from the Sun.
We could formalize the "large enough" and "close enough" by an empirical coefficient, a "planetness" coefficient that would disqualify an object as a planet under an arbitrary threshold.

A purely illustrative and quick example would be something like:
Equatorial diameter: ED(planet); Semi-major axis: SMA(planet); Size factor: SF; Distance factor: DF;
Size coefficient: SC(planet) = arctan(ED(planet)/SF)*2/Pi;
Distance Coefficient: DC(planet) = arctan(DF*1/SMA(planet))*2/Pi;
Planetness coefficient: PNC(planet) = (SC(planet)+DC(planet))/2. With ED and SMA in km, SF=2 350, DF=6 000 000 000 and no care for consistent units, we could define a planet as:
A planet is an object which orbit's center is a Star, and which Planetness coefficient is greater than 50%.
Pluto would be in and 2002 UB313 would be out in this example. Mercury:85%, Venus:93%, Earth:93%, Mars:88%, Jupiter:95%, Saturn:92%, Uranus:84%, Neptune:78%, Pluto:50,5%, UB313:47%.
An object the size of Jupiter would virtually be guaranteed the name of planet, for its size coefficient is 99%, and it would have to be more than 2440UA from the sun to be disqualified as a planet. Tweaking relations would allow to have the desired behavior, like increasing the planetness gap between Pluto and UB313, or including UB313 in, or rejecting an object the size of Jupiter from a smaller distance. Finally, to take in account the notion of size of the system, we could have the size factor and distance factor depend on the system's star's equatorial diameter.

Wouldn't it be acceptable?