I'm amused that people have been using "not a game" as a means to criticise games they don't like, and now here's Jim saying they
are games ... as a means to allow him to criticise games he doesn't like.
More seriously: Where do you draw the line when the lines are so blurry?
What's the difference between a simplistic interactive fiction "game" that requires you press a button to progress, versus a TV series on DVD that requires you press a button to view the next episode? How is stop/slow/medium/fast in a train simulator different than pause/slow-mo/play/fast-forward for a movie?
And more importantly, why do we need to strictly define things at all?
Is it for buyer awareness? Well, if they want to be aware, they should be reading reviews, which will tell them exactly what the game is anyway.
Is it for criticism, i.e. the ability to judge them on game standards and compare them to other games? This seems to be one of Jim's points, but I don't think it's that simple. If you're taking one of these "non-games" ("walking simulator", "interactive experience", etc.) and trying to judge it based on standard expectations of gameplay, you're going to run into the same issue as if you try to judge (say) a MOBA based on MMORPG standards.
"WASD controls don't work; no open world hub; long queues to get into instances, only 10 players per instance, and my skills and inventory reset every time I enter one; needs a PvE mode because I'm trying to farm mobs and level up, but people keep sneaking up and killing me; 2/10, not a good MMO." At best, this sort of criticism is irrelevant; at worst, it stifles new forms of media by demanding they conform to existing ones.
As far as I can tell, the main reason for defining and labelling things is just as a sort of intellectual shortcut, as a way to say "here are the things I'm interested in and here are the ones I'm not". That's fine, that's useful. There are certainly genres I generally avoid unless a particularly stand-out title comes along.
What I wish people would stop doing is pretending these definitions are anything more than vague categories to suggest who
might be interested in them. Instead, people keep trying to come up with strict definitions -- definitions they've usually cherry-picked to exclude things they don't like.
Yes, it's true, definitions can be diluted to the point of being useless as a label. "Roguelike" is a recent example of this, where any game that involves any sort of procedural generation and/or permadeath now gets billed as being one. But what if, instead of trying to argue whether a game is or is not a "roguelike", you just treat it as a keyword to maybe pique your interest, and then read the feature list and see what parts of "roguelike" they've chosen to use? At worst, you've wasted a couple minutes looking a list of features for what might be an interesting game -- certainly far less time than you would spend trying to convince the dev they should drop the "roguelike" label.