An excellent point about whether composition should be an important part of determining planet status.

The definition of planet, in my mind, should be geared toward classifying things we haven't found yet - when we come to another object in another star system, how should we designate it?

If it's as big as Jupiter, but is essentially a giant ball of methane (forgive my ignorance - Pluto is mostly methane, right?), will we deny it planetary status?

Similarly, are the giants they are finding now, which have highly elliptical orbits, not planets because their orbits are so eccentric?

If it's as big as Jupiter, but is in the middle of that system's Kuiper belt (an impossible situation, granted, but for the sake of argument...) would it not be a planet?

If it's as big as Jupiter, but was captured by the system, rather than created by the leftover rubble when the star was born...?

etc.

For me, the answer to all the questions I've posted is 'still counts as a planet.' I agree with the criteria above - not a star nor brown dwarf, orbits a star or brown dwarf, bigger than anything that's not a planet.

So the debate is really the cutoff mass. Asimov's suggestion that it has to be big enough so that its gravity pulls it into something pretty close to a sphere, works for me - for solid planets, anyhow.