In reviewing Should we have a policy about "too much downvoting"?, I noticed that it was of type discussion but also had status declined. The documentation on How does Meta Stack Exchange work? says parenthetically that this status is usually used for feature-requests. I can understand how this status might also be used for bug or even support, but I'm not sure what to make of it for discussion. Can you help me understand the conditions under which moderators would "decline" a discussion question?
The discussion tag itself is often used as a "disguise" of sorts for a feature-request when the author isn't quite sure if the feature should be implemented and wants discussion from the community on whether it is a good idea or nailing down the specifics of the feature.
Sometimes the tag is omitted simply on the basis that it takes up another tag, and there are other, more prevalent tags that the author would rather use. Other times it may just be that the author didn't want to attach two required tags to their question. Generally speaking, the tag refers to the body of the question. If the question itself is requesting a feature in some way (even implicitly), then the question is subject to a status-declined tag like any other feature request.
The mis-tagging here is that the question is marked as "discussion" despite the fact that it's a "feature-request".
Having a "status-*" tag is appropriate, because it is indeed proposing a new feature.
As far as I'm concerned, the feature-request tag should be reserved for actual feature requests. You know, ones that you've put a lot of thought into, have a proposed implementation, and really think are in a finalized ready-to-be-implemented state. You're not really angling for community consensus, you're essentially demanding that the feature be implemented. (Although "demanding" sounds like a bad word, I don't mean it that way or to imply anything bad.)
On the other hand, a discussion question is one where you are looking for community input and trying to solicit a discussion. That discussion can be on anything related to the Stack Exchange network, including ideas for possible new features that have not yet been fully fleshed out. These are basically feature requests that are still in the incubator—you have a glimmer of an idea, something that you think should be done, but you're not entirely sure if you're right, and you have no ideas on how it should be done, but you want to open up a discussion about it anyway. Answers to this type of question would disagree that anything needs to be done, disagree that anything needs to be done now, propose a solution to the problem, provide additional data in support of the problem, etc. etc.
For example (and I pick on my own questions because they're handy and I know what I was thinking when I wrote them), this question is a lot like a feature request—I want something on the site to change. But I'm not proposing a concrete implementation, I'm just throwing an idea out there for discussion about something I'd like to see changed. I'm open for input from the community on my ideas, and hoping to get some good ideas on how to implement it. As such, I've tagged it simply discussion.
This one, I was pretty certain that I was right. I had a concrete feature I wanted to see implemented, and I knew exactly how I wanted it to work. It was simply a matter of writing the code and unit tests. In my mind, there wasn't any need for discussion about how it should work, just whether it will happen. Consequently, I tagged it feature-request. (And then that Shog9 guy came along and shattered my dreams.)
Of course, even questions tagged feature-request still often elicit discussion. It is perfectly reasonable for people to disagree with the inherent validity or proposed implementation of your feature request. They can post answers that explain the rationale behind their disagreement or propose alternative implementations, and they can downvote the question to signify their disagreement. This is a common and (in my mind) perfectly acceptable use of downvotes.
Both of these are perfectly valid uses of the tags. You could tag a question with both tags if you like, but you are not required to do so.
As far as the particular question you linked, it appears to fall into the second category: when originally posted, it was the beginning of a feature request, but not yet a full-fledged request. After the discussion took place, an actual feature proposal crystallized, which was subsequently declined. It would have probably been appropriate to substitute the discussion tag with a feature-request tag when it was marked status-declined, but that didn't happen—not a big deal in my mind.
The tagging system is not absolute. Just about every question here could be tagged discussion because this is a discussion site. Lots of discussion questions contain feature-request-like elements. Others are actually support requests in disguise (because someone doesn't know or assumes that a feature does not exist). Some even turn out to be bug reports. Categories are fluid, and that's why we allow questions to be retagged. Don't read too much into them. The content of the question itself—and the answers—is where the real magic happens.
Summary: If this discussion item really morphed into a feature request at some point, then it should have been retagged as such and the status declined tag was appropriate. If the intent of seeking feedback did not change, then no matter what Jeff's role or "leaning" was, it should not have been given the status-declined tag.
Background: I think a lot of this and related consternation is traceable to different views of the discussion tag as it relates to questions about feature changes and, to help with the illustration, bug fixes.
Consider the following matrix with "topics" as rows and "intent" as columns:
The question is, what tags do you associate with questions that fall into each of the cells above? I contend there is no controversy regarding the "Requesting implementation" column. Everyone seems to agree that the bug fix cell should get the bug tag and the feature change cell should get the feature-request tag. I think also that there is no controversy around the fact that although users should be prepared for some "discussion" when submitting either of those, these should not also be tagged with discussion.
The disagreement seems to rest with the "Just want dialog" column. I'm aware of two different positions on this:
There's a third possibility, namely that both the topic and intent tags should be included, but I haven't found that view promoted yet.
Perhaps I'm an outlier here, but until this is clarified, I think it's going to be difficult to have meaningful discussion about policies that are specific to these tag types.
As a final aside, I'm not aware of any controversy regarding the tagging of support questions or the tagging of discussion questions not related to bug fixes or feature changes.
Update: Just had a long chat with another user who is squarely on the "Topic takes precedence" side. For the record, I'm on the "Intent takes precedence" side. Their primary argument was ease of finding questions related to features and was based on the assumption that people would agree/disagree vote feature-related discussion items anyway. My primary argument concerned promoting free exchange of ideas and was based on the assumption that we'd figure out a way to greatly reduce agree/disagree voting on discussion questions.