I saw a documentary where they inserted the gene of a mouse that basically is the starting "build an eye" command into a fruit fly, and a fruit fly eye grew. My question is, if eyes of different types of animal had different evolutionary beginnings (I don't know much about arthropod eye evolution, but for a different example, as I understand it, cephalopod eyes evolved from skin cells while vertebrate eyes evolved from brain cells), then how is it that the same gene controls the growth of two such evolutionary different eyes?
Both fruit flies (Drosophila) and mice (Mus) are classified under Bilateria.
The presence of such a highly conserved sequence in both species suggests that they share a urbilaterian ancestor who also used a similar gene to turn on eye formation (not create the whole eye, just start the process).
Even though the structure of insect and mammalian eyes diverged evolutionarily after the common ancestor, the switch that starts eye development worked so well there was little need to change it. I suppose one could say that it remains backwards compatible.
The documentary was probably talking about Halder's paper in Science (Science, Halder et al., v.267). If you'll read the abstract (especially the bottom part) You'll notice that the gene is a control gene that is responsible for starting a chain of interactions at the end of which the real fly-specific 'eye-building' process begins. Basically this is like using the same kind of electrical switch to turn on the light in your room (mouse) or an air conditioner (fly).
So the gene that was implanted from the mouse to the fly is called a homologue/homeotic/hox gene which doesn't actually build the eye it just switches on the genes that build the eye. The reason that this worked is because hox genes are found in most or all multicellular organisms. In terms of evolution this is most likely evidence for a common ancestor.