re: What is life ?
On the subject of negative entropy, I think dyson objects would fit the bill.
On the "philosopher approach" : "a machine has no goal"--I would dispute that this is always the case. We often give our "intelligent" search algorithms goals. We could easily program a machine with goals to survive and reproduce, which could be argued is equivalent to life. Many secondary goals would be established in a competitive environment to survive and reproduce (i.e. evolution would emerge with attendant positive sub strategies).
I tend to think that most all quantities that we measure exist on a continuum(down to the quanta) and that includes life and and all the qualities that we tend to associate with intelligent life.