All the observations in the world are insufficient for proof: we must rely on our senses and instruments, which are necessarily faulty, and repeated experiment, which relies somewhat on the assumption that after seeing a hammer fall hundreds of times, the five hundred and first time it won't turn into a bird. Proof as such only really exists in internally consistent abstractions such as maths and logic. In looking at such a world the theories we construct are based on facts, but a fact in science is only as enduring as mankind is able to determine such a thing.
Rather than seeking truth per se science builds models based on observation and experiment, and such models must by their nature be tentatively held, for there is always the potential for an observation of things acting in a manner that is unexpected. To account for this new data the model must be altered to explain this new evidence. Our theories, even those for which mountains of evidence exist, all follow this pattern, constantly.
To take the example of electromagnetism: mankind was utterly ignorant of the relationship between electricity and magnetism until a battery was accidentally turned on near a compass by a very smart chap who decided to write about it to other smart chaps, and after a lot of thinking and prodding and waving magnets with coils around we now treat the two as a single force. Hell for a long time we had no idea that lightning for instance had any relationship to electricity whatsoever, until, among others, Benjamin Franklin was able to convincingly show it. The concept of an electrical charge, its nature, that all electricity is essentially caused by the same phenomenon, was based on the gathering of decades of observations, experimental data and constantly altering models. Each scientist who contributed to this knowledge was himself in the dark about many aspects of electromagnetism.
Now after so much observation, and empiricism having spent the past few centuries demonstrating its worth, the chances of the phenomena of electricity being completely swept aside by a revolutionary discovery is roughly the same as the falling hammer turning into a pigeon (I worked out the statistics of both using a science calculator). Even if we find our models are mere shadows of a much wider phenomenon, or that we have gotten the wrong end of the stick entirely and our explanations are the ridiculous mewling of man-children, the evidence we have gathered already - the data of the thing - is not thereby suddenly wrong, rather it is incorporated into a much wider body of knowledge, and the resulting explanation may be only dimly recognisable to us as electricity as we know it.
After centuries of changing models, our understanding of what electricity actually is has become vastly different to that of its first discoverers. In later centuries the future discoveries of mankind (if we hang in there) will similarly make our understanding of it seem paltry. True revolutions in science are rare, some philosophers of science would suggest they do not happen at all, but rather there is a continual and gradual series of reformations, where even the apparently sudden disappearance for the need of the luminiferous aether was really paved by decades of discoveries which made it increasingly unnecessary and an alternative explanation for the data increasingly well-shaped. Even something as socially tumultuous as the discovery of the common descent of all life on earth was presaged by the discoveries and hypotheses of Erasmus Darwin and Jean-Baptiste Lamarck, again amongst many others, who all gradually added bits and pieces of knowledge, who added claims which could be researched and tested and found wanting or were supported.
Gravity is a theory, and in many ways not a terribly well understood one. Newton's universal law of gravity (and it is literally this) is certainly not a theory, but that is because there is no explanatory model or framework there. It tells us very little about gravity; indeed, even at the time Newton knew it was incapable of accurately describing the movements of the planets in the solar system with the required accuracy for the system to be maintained, and Newton himself, one of the cleverest, if not the cleverest bloke who's ever existed, had to fill in the gaps at the edge of his knowledge with such things as God. He was incorrect to do so, and he was more than smart enough to solve the problem himself, had he not shied away from doing so.
The Universal Law of Gravitation is a single datum, an equation which describes a single relationship. A theory of gravitation would seek to explain a collection of huge amounts of such data. I should also point out that Newton's law of gravitation is actually a bit wrong: general relativity (theory!) has already demonstrated that the 'law' is an (albeit a very excellent and eminently useful for most of our purposes in history) approximation, and becomes increasingly inaccurate as the speed of light is approached or the gravitational potential energy increases. Its inaccuracy is demonstrable by observations we can make here on earth (and indeed such observations helped develop general relativity in the first place). To cleave to a 'law' as being a fact where a theory is not, and by that to mean a 'law' is more truthful and proven than a 'theory' by mere dint of the terminology used, is to completely misunderstand what these terms are referring to.
The fact that all knowledge about the natural world is essentially tentative does not mean it is not useful or not to be relied upon. Humanity got a man on the moon relying on Newton, but without Einstein we wouldn't have GPS. There are fundamental problems with the way science is presented in the news media and to some extent I understand that, and a deep disconnect from the increasing specialisation of academia, are perhaps the predominant causative factors of the deeply ingrained wish to remain ignorant amongst many of the kind of people on the internet folk like me enjoy picking on. But to leap from that mistrust in science, which itself is constantly a self-correcting enterprise, to placing your trust in politicians or theologians or laymen pastors to interpret scientific results, to reject a consensus you do not understand the reasons for as so many do in regards to creationism, anthropogenic global warming denialism and so on, will perhaps forever remain deeply perplexing to me.
-The Rev. Schmitt.