No it is not because a new neutral gene has only a 50% chance of survival at each generation. See post#749. It will gradually dissappear from the population. It has little more chance of surviving and taking over a population than having a dozen true coin flips coming up heads.
I'm sure it has been observed but I don't have any examples right at hand now but I think a websearch should turn up some examples.
Asking opponents to do your work for you!
Why not? You have to keep in mind that this is a parallel and not a serial process.
If your above two statements were true, and evolution had indeed proceeded that way, the examples would be readily available to everyone and you should know them. Clearly the examples are non-existent. It is not a serial process however. You cannot 'add' mutations from one gene to another as would be necessary for it to be a serial process. Each mutated gene has to be re-mutated individually till the right 'combination' is found. Of course since most of the 'combinations' in fact the large majority of them are detrimental, it would result in massive deaths in the species (including the deaths of those carrying the gene). Also to add to the problem you need to realize that while a single mutation may kill an individual, a second one together with the first one might be beneficial but you will never get there because the first one killed off everyone possessing the 'first step'.