I disagree with this definition. We have yet to produce a perfect model of the world (aka, a theory of everything). All models produced by "science" thus far are "wrong", at least on some level (ex. Newton's model doesn't cover relativity). I think "Creating models with predictive power is also a precise definition of science." is a fair description.
I think it's fair to say that a "theory of everything" is sort of the great work of any particular field of science. In practice that means refining models, but the model-building is ancillary to the truth-finding, not the other way around. Of course, if the truth wasn't predictive we're all just screwed, but that doesn't mean that whatever is predictive is necessarily the truth. It just means we might all be screwed.
