I propose that obviousness should be tested with an expert in the field. If that expert could explain how the invention worked by only looking at the result, but not at the patent itself, then the patent is considered obvious!
this is where the ideology between us differ - the classic example i use when i argue with friends on this topic is 'the wheel'.
Some of my friends argue (and i assume this is your point of view), that the wheel would've been granted a patent should it be invented today. I argue that even tho the idea is novel, there is no "secret" that you can hide with the wheel invention, and hence, there is no incentive for society to give you a monopoly, because you can't both use the invention, but hide its implementation from society. The price for said monopoly is the devulsion of the secret, and if there are none to give, you do not deserve the payment.
If the patent was for some complicated industrial process to produce better paint, an expert looking at the paint can't work out what went on in the process. But if you look at the sort of "design patents" that have come out, the amazon one-click is the classic example here, there is absolutely nothing that they can hide about the workings of the patent, and thus don't deserve a monopoly.
Patents aren't intended to protect an existence proof. They're intended to protect the underlying methods. If, after seeing a swipe-to-unlock gesture on an iPhone for instance, someone of ordinary skill can implement it without referring to the teachings of the patent, then the patent is bullshit.
(Of course, that's not the way it works, but it's the way it should work. Remember that patents exist to give inventors an alternative to keeping trade secrets. Something that would make a laughable trade secret should also constitute a frivolous patent claim.)
Along the same lines, what I think is important is to see if one of ordinary skill  in the domain of the invention could come up with the solution proposed in the invention when facing the problem at hand. Often, the problem is new, and any solution, obvious or not, gets accepted as a patent.
 An expert would be better, the current law only asks for ordinary skill
This sort of logic is supposed to already have been done prior to granting the patent so I'm assuming you mean that, for each patent granted, this is to be performed before granting the patent. Are you saying that you want this done also at the trial/hearing?
I'd further ask:
- Who are these 'experts' you mention?
- How are they identified as such
- Who pays for their time?
Let's take the case of a drug patent - walk me through how they find an expert who doesn't work for either company, how long it takes the person to do discovery, how much they charge, and who pays.