If their are people testing the apps in the market, the company is going to have to pay them. They have to pay people, they'll end up charging for all apps. Might as well go to a company that already does that if that's the case.
You take the flexibility of an open software, you're going to have to deal with user made apps that can bug your system. This is where most problems come from considering you are getting weekend programmers putting up half assed apps.
We like the Droid because we can get the customization out of it, so we deal with it.
There are plenty of things they can do as a middle ground. For starters they can offer a 'preferred' or 'approved' cert or something for developers that wanted to charge for their apps. They could also have a user driven/monitoring system that would track issues and make the appropriate recommendations and monitor such conflicts between programs and/or resources. You have to remember the 'average user' out there doesn't care how or why it doesn't work, they're just pissed when it doesn't.
Just because you want an "open market" doesn't mean you have to let it run a muck. Especially when it doing so will be a HUGE turn off to your 'average' user. The idea of being totally 'open' is great for something like a desktop OS where, generally speaking only knowledgeable people are going to be using the product. That not the case when you are trying to get a product to be adopted by the masses and become a common use item. In fact it's just to opposite. Look at how Apple did it (and I HATE having to use them as an example). They use Linux as a base but then make it 'theirs' and they control it (granted probably too tightly, but that what has to be done to make a product that will be embraced by the masses)... you have to remember the 'average user' out there doesn't care how or why it doesn't work, they're just pissed when it doesn't.
I agree with you that people don't care about why it doesn't work, they just want it to work.
I think this would be an awesome suggestion to the Android developers. Maybe we should start focusing on what we might be able to bring as actual suggestions for future ideas or fixes for current problems.