I'm not a fan of shopping questions, but so far it's seemed that the consensus is to leave them open (usually nobody votes to close them, nobody flags them, nobody edits them into more general questions).
A big part of the problem with shopping questions isn't that they're inherently bad, it's simply that the Stack Exchange Platform isn't designed to handle them. The voting system doesn't work right. It would be too easy to get a shopping question with page of answers with the highest voted answers being the popular choices (and earlier written choices) instead of the best solutions or best-written answers. Nobody's going to read a long list of undifferentiated answers; they don't really help anybody.
I would encourage people to ignore the phrasing of the question and pretend instead that the person asked the question we wish they'd asked. In other words, when you see "Which jacket should I buy?", pretend they're asking how to pick a jacket for their particular needs and provide that type of information. Maybe include a few specific product recommendations with the reasons for them related to the general reasons. Even bad questions are likely to be left open if they have an awesome answer.
Even better, edit the question into the better question that we wish they had asked. Make sure it will still lead to answers that will help the original questioner out, and try not to invalidate any existing answers. Edits that improve questions or answers are always welcome!
Notice on that shopping question that the upvoted answers are the ones that tell you how to find a jacket or are specific about why a given choice is good, and that the actual straight product recommendations never got any votes... But that's really because we have a good community of good answerers with a natural inclination towards awesomeness, not because the original question was very good (or good for this platform).
Some of the comments on Q&A is Hard, Let’s Go Shopping! are pretty interesting:
I agree, let's leave shopping questions open but try to explain why we're recommending solution [foo] for a given problem.
These kinds of questions is that they tend to get out of date very quickly. I'll use a fictitious company and product:
Perhaps I'd recommend the BlahBikes Greatbike! as a great tourer/commuter, but in a few months, BlahBikes has decided to no longer make the Greatbike! and the answer now recommends a hard-to-find product. If I had said in my answer why I'm making the recommendation I had and simply used the Greatbike! as an example of a bike that has all these features I suggest looking for, then the answer is still useful to the site.
When I asked because I wanted to shop, people told me what to look for using common nouns and adjectives (not using proper nouns / brand names).
The LBS sales person then suggested specific items / brands; people's answers had helped me learn what components+characteristics to look for: which is just what Jeff's blog post recommended.
I'm not saying that you mustn't recommend a specific brand; maybe sometimes it's more useful if you do (perhaps, I don't know, especially, if it's a hard-to-find or one-of-a-kind or pretty-unique...); but it's, sometimes, more useful if you don't.