Isn't complexity really subjective?
What we habitually consider complex are those things that are arranged to our values.
In an objective sense, it may be that a desktop PC is no more ordered, no more complex, than a cloud of gas.
Thanks for the ping.
See if you can get a cloud of gas to download your e-mail ;o)
Complexity does have strict measures, measures that are what is actually being used in thermodynamics, mathematics, and elsewhere, but very few people who are not mathematicians have any idea what the measure looks like and intuition about such things is almost always badly mistaken.
The most useful general measure of complexity for a finite observer -- a general subjective metric, in other words -- is something called predictive error complexity. Given some pattern, the PEC describes the mathematical limits of knowledge a particular algorithmically finite observer can have about the nature of the system that generated the pattern. This is an obscure concept that is only about ten years old, but it has quickly proven to be an extremely important unifying concept with broad and deep ramifications. Formal definitions of intelligence and measures of the same are based on it.
Kolmogorov complexity is a better known measure, but has little meaning in practice outside of machines that are infinite in space and rate (time). Theoretically very important, but you cannot engineer anything with it. PEC is the purely finite case equivalent.
Well, yeah. It is a concept. As such, the word used could be "fluffiness". The fact is, something discernable exists between a cloud of gas and a PC. And that something is more(different) than "color", "temperature", "mass", or "position".