The whole evolutionary outlook has multiple layers of rather surprising propositions. These are often mixed up, so an argument against one layer is taken to refute another layer. It's worth going step by step, to take care to see how the various puzzle pieces might properly fit together.
The foundation of the theory of biological evolution is paleontology, the study of fossils from various layers of rocks. Using many cross-checked methods to date the layers, it seems clear enough that our planet earth has been around for several billion years, and that the organisms living at one period in that time span are very different from the organisms at other periods separated by millions or tens of millions of years. Species appear in the fossil record at some point, and then disappear at a later point. Some species have a relatively short span, others much longer.
This was already well known in Darwin's time. The raw picture of species coming into existence at various times and then disappearing at later times - that doesn't even merit the label "evolution", nevermind Darwin. Yet this basic picture presented by paleontology already conflicts with traditional views of the origin of the world presented in various myths and religions. The science of paleontology will surely continue to correct and refine its picture of life across the history of the planet. Whatever the merit of maintaining traditional views might be, it ought to be clear enough that the ever-fresh look cultivated by science also has great value and deserves a prominent place at the table of ideas.
The next layer of scientific ideas about life on earth brings deeper structure: moving from mere change, to evolution. Looking again at the paleontological picture, one can observe that the appearance and disappearance of species is not disconnected. New species that arise at one time are almost always quite similar to some other species that already existed at some slightly prior time.
There are several ways that species might arise in such a pattern, in sequences where each new species is slightly different than the earlier species. Of course, it is logically possible for this correlation to have no cause at all - it could be some arbitrary bit of luck. This seems to be rather an abuse of logic, though - it fails to help us work effectively.
Another possibility is that similar circumstances have similar results. The paleontological picture, and even the astrophysical picure, show us a planet with a beginning. Species arose before which there were no species. If some species arose once without a prior species, surely this could happen again and again. Occam's razor would recommend the fewest causes! Perhaps each new species arises just as the first species did, through whatever obscure process that might be. This process could well produce species that are well adapted to whatever the environment happens to be at that time and place - in any case, poorly adapted species that happen to be produced will not survive, so the paleontological record will only show us the well adapted survivors. A slowly changing planetary environment would thus result in a slowly shifting collection of species.
The basic problem with this spontaneous generation theory is not paleontological but more a matter of everyday biology. We just don't observe spontaneous generation. Nature shows us a bewildering variety of ways that existing organisms can give rise to new organisms, but we never see new organisms arising without some prior organisms as their source. Invoking Occam again, we might even guess that the first organisms to appear on earth must have arived from some other planet by way of comets etc. Leaving such speculation aside, though, it seems clear enough that new species have similar prior species as their source. This is the theory of biological evolution.
The idea of a gradual shift in the characteristics of a species seems unproblematic. Clearly most species include quite a variety of characteristics across the organisms it encompasses. Clearly children can have characteristics significantly different than their parents. Does a species incorporate some kind of genetic anchor that prevents the population from drifting too far in any direction? How far must a population drift before we decide to apply a new label, before we decide the population now constitutes a different or new species? If a species is unanchored, then clearly a population that became geographically divided would be capable of bifurcating into distinct species. The theory of evolution is essentially the idea that no anchor exists that limits the drifting of the characteristics across the generations of some population.
Whether or not things have some inner essence that makes them what they are - this is a core metaphysical issue, deeper even than the conflict between science and religion. It is our rich traditional cultures that make us human, that give us contexts in which we can survive and thrive. A core component of any culture is a picture of world, a categorization of components of experience, a way to understand the world and to coordinate our work in the world. Instability in that picture of the world is a threat to our survival. But inaccuracy is also a threat. And that is the source of the conflict between science and religion. The movement from a less accurate picture to a more accurate picture is at least temporarily destabilizing.
Perhaps there is a simple resolution. Let us build a stable and accurate picture! Can't science finally work through all the various errors and reach its endpoint, the ultimate theory, the true picture of things as they are? It becomes clear at this point that the deeper debate is not between science and religion. Science and religion are both institutions incorporating wide diversity. Neither have resolved the question of whether a stable and accurate picture of the world is achievable. To what extent does faith in inner essences contribute to effective work in the world - to what extent does it blind us and make us less effective?
One curious feature of the battles over evolution is that the most crucial conflicts have nothing to do with Darwin's particular theory! The lack of an anchor for a species, i.e. the ability of the characteristics of a population to drift so far that a different species label becomes appropriate - this is a cornerstone of evolution, but not Darwin's particular proposal.
A noteworthy feature of life that we can observe both in the present and across the paleontological record is ecological coherence. The drift of charactistics of a population is not directionless at all, but adaptive. Species are generally well tuned to some ecological niche. How does this happen?
There are two fundamental processes that shape biological characteristics: parents give rise to children, and children give rise to parents. When e.g. two sheep mate and give rise to lambs, the characteristics of these offspring are loosely constrained by those of their parents. The children won't generally have characteristics identical to their parents, but they won't be wildly different either. Then, in turn some of these children will mature and become parents themselves. Some children will never become parents at all, while some may become quite prolific parents.
Is adaptation a result of parents giving rise to better adapted children, of children giving rise to better adapted parents, or of some combination of these processes? Darwin's brilliant hypothesis is that adaptation can be completely explained as the result of children giving rise to better adapted parents. The variation of characteristics of children can be completely haphazard. The environment is a very powerful filter, though, making some children successful in becoming parents - those with better adapted characteristics - while children with less adapted characteristics fail to become parents. Of course there is a lot of noise in these filtering process. But over the course of hundreds and thousands of generation, the noise averages away and the environment signal becomes a clear force consistently channelling the drift of characteristics of a population in the direction of better adaptation to some ecological niche.
Of course, these ecological niches are themselves little more than opportunities formed by the characteristics of other species inhabiting the same landscape. All these species have characteristics that are constantly drifting and responding to the changes in each other's characteristics. But the notion remains, that the adaptation or responsiveness of the characteristics of a population can be explained entirely as the result of the environment of the population selecting which characteristics allow children to become successful parents and which prevent that success. The variation in characteristics of the children of successful parents can be entirely haphazard, and yet the whole drift from generation to generation can still be adaptive.
The keystone of Darwin's theory is the haphazardness of the change in characteristics between parent and child. One can imagine variants of this hypothesis. Perhaps it is an iron-clad law of biology that such changes are always haphazard. Or, alternatively, it could be that change has almost always been haphazard in the past and will almost certainly continue to be haphazard in the future. Or again, perhaps the pattern adapting and responding of the characteristics of populations is indistinguishable from what it would be if the parent-child difference was indeed haphazard - i.e. any non-haphazardness has had no effect on the pattern, and perhaps is unlikely to in the future. In any case, adaption and response are due to the way the environment selects parents from children, rather than the way parents generate children.
Events on the ground have now constrained the strength of the haphazardness hypothesis that can be supported. With genetic engineering, the process by which parents generate children has become subject to non-haphazard environmental influence. It is clearly not an iron-clad law of biology that the change of characteristics from parent to child is always haphazard. The strongest version of Darwin's theory simply does not correspond to reality.
Another style of variation of Darwin's hypothesis brings in a distinction between natural evolution and artificial evolution. Perhaps the hypothesis of haphazardness is to be applied only to natural evolution. But this distinction causes more problems than it solves. How are artificial evolution and natural evolution to be distinguished? We might introduce human intent as the distinguishing factor. Is human intent then somehow outside of biology, outside nature? Probably the most shocking hypothesis of Darwin is that humans are actually natural, full and normal participants in all the processes of biology, organisms alongside all the others.
Logically there is a choice: either human beings are natural organisms whose behavior is subject to all the laws of biology like any other organism - in which case genetic engineering is a natural process, so haphazard variation of children is not universal; or, human intent is somehow not natural, human behavior is not subject to the same laws of biology that govern other organisms, so genetic engineering can be ruled out of the bounds of normal natural evolution to which the laws of biology apply.
One has to wonder, too, given the marvelous creativity of evolution, whether humans are the only species that has worked methods of genetic engineering. I am not proposing that any other species in the past have developed X-ray crystallography, chromatography, and all the other techniques devised by humans for genetic engineering. Nor even am I proposing that other species have developed the ideas and intent that would parallel our own approach to genetic engineering. But it is at least conceivable that some symbiotic system could exist, involving organisms at several scales down to the viral level, where some sort of chemical signaling from one species could steer the genetic processes involved in reproduction. So often in science some special circumstance opens one's eyes to some extraordinary possibility, and then one starts to find that possibility actualized again and again in other less special circumstances, so one is again reminded how extraordinary the ordinary really is.