It seems to me that the issue is not literally, "Is DF a game?", but more one of "What would DF's status as a game/simulation/something else mean for development direction and priorities?" Tarn of course has his own conception of this, but I know if I were in his shoes, these types of discussions would remain fascinating because our conceptions of our own projects is always changing.
So given that, here is an attempt at summarizing the patterns I think I'm noticing in this thread. It seems to mostly be a matter of where our priorities are between a high quality interactive experience, and the creation of a rich and fascinating environment. Great games typically excel first and foremost in the category of interactive experience. The game could be simple or complex, but what it primarily does correctly is it gets the interactive portion tuned just right. In contrast, great simulations (at least the type that are relevant when considering DF) typically excel first and foremost in the category of creating a rich environment, full of interesting details and elements that interact with each other in fascinating ways; it may include some form of interaction with the user, the user interface may be complex, elegant, painful, whatever, but the rich environment is what makes the simulation shine.
I personally find myself attracted to both of these categories. I am nearly obsessive about the quality of user interaction, but I love a rich environment as well. What I really want from games like DF is to have an interesting world with which I can meaningfully interact. Both elements must be high quality for me rate the overall experience highly. A high quality game in a simplistic world might be great for some people, but I get bored easily. Similarly, a rich environment with limited user interaction tends to be a novelty that loses that novelty quickly, and a rich environment with difficult user interaction becomes frustrating. This last is the most annoying to me because I'm so tantalizingly close to a high quality experience, but I have this roadblock in the way. At least with simple games, they were designed to be simple, and at least with simulations with limited interaction, they were designed to have limited interaction. But when it is designed to have a lot of interaction, but that interaction is low quality, ugh. I can put up with it for a while, but eventually it just becomes too much. I spend more time thinking about how it could be fixed than I do actually using and enjoying it.
My opinion about DF's current state is that it is doing pretty well on the simulation side, but is far behind with the interactive experience. Thus, it's no surprise that I prioritize the interactive parts of the game. I have no problem with the addition of more simulation components, other than the fact that doing so takes time away from the interactive components. Other people might share my opinion about what they want out of a game like DF in general, but disagree about the current balance between interaction and simulation for DF specifically. Others of course will straight-up disagree about the importance of one or the other component in general. I'd like to imagine that Tarn shares my roughly equal concern for both, and that it will effect future development in a way that I appreciate, but that's obviously up to him. But for now, I definitely spend more time thinking about how to fix DF than I do playing it. (I haven't actually played DF at all for nearly two years, and I don't see that changing any time soon.)