The first thing you have to do is make sure you have an actual problem. Consider the following question:
I have a Beagle with [z and x] skin allergies, as well as open sores due to him biting the spots that itch. I need to bathe him, what is the best shampoo for his condition?
Is this a shopping question? Yes, technically, but the scope is rooted enough that we really shouldn't expect too many answers to this, and that's the key. When you look at what has not worked well on other sites, you've got to look at why it didn't work well, or what sorts of undesirable things happen when open-ended questions are asked.
What the above question really asks is:
Is there a shampoo I can use for a dog in this condition? If so, what is it?
That's not open-ended, in fact there are probably less than half a dozen or so answers to that question, all of them likely to help someone in the same predicament.
I really suggest waiting for these examples to manifest organically and then bring up a discussion about them, it's too easy to box and compartmentalize these questions hypothetically based on what they might contain, or what you see as the perceived scope of a bad question.
Remember that Pets Stack Exchange has quite a bit of competition, the two things that are going to help us gain evangelists are the quality of our content and the friendliness of our community. We need to keep the noise down, but I don't think we're at the point where we should be making concrete plans on what to exclude.
Why not let a few questions come in, watch what happens, and then raise the discussion again? Funniest thing your dog ever did is probably not going to fly, but What's the best way to help my previously abused dog take better to a leash? is something you might grow to be proud of.
Right now, I think it's just too hypothetical. The site will be going public in the very near future, let's give it a few weeks and then circle back?