N.B.: This review was of Minecraft in its earliest outing; see here for a more recent (and favorable) take on the game
Back in the day, I entered Habitat and played around with it. And I remember Chip Morningstar telling me that it "wasn't just a game," and thinking "Too bad! It could be a pretty good game!" That is, if you bolted on some, you know, actual gameplay. Something to do. An objective.
Some years later, as part of a project, I did a survey of just about every online visual chat environment, from things as simple as Comic Chat to actual 3D environments. And came away thinking "Why?" Because even though there were some cool things, like The Palace, in the final analysis, text is better and the other stuff just gets in the way, if what you want is, you know, chat.
Now, there are lots of definitions of "the game," and I don't intend to flog that particular horse any further today, at least; but one thing that is common to virtually all of them is goals. Not that all games need a win condition, but if the structure doesn't enable goal-directed activity on the part of the players, either explicitly or implicitly, it's not a game.
Not that being "not a game" is itself an inherently bad thing; but let's unpack this "goalness" of games a bit. Games are interactive, and game players have goals; an interactive structure without goals has no -- well, no goals. No point. No reason for further interaction.
Which is why I get very irritated with people who want to do "more" than or "go beyond" or "not be limited to" games; what they think they are saying is "I want to create art beyond the pathetic dweeby little things that you degraded gamers seem to enjoy," but what they are really saying is "I want to create pointless applications."
And it is a source of enduring wonder to me that Second Life has attracted so much venture money and press attention even though, you know, it's just Habitat with a scripting engine bolted on.
Which brings us to Minecraft, which has gotten a fair bit of attention in the indie community.
Minecraft is an impressive technical achievement; it's a 3D construction kit that allows you to create your own virtual worlds by placing and removing textured, cubical blocks. You can then move through your construction, or spawn "mobs" controlled by idiot AIs to wander semirandomly through it. You can host your "world" on your own server, and many people (a couple of dozen, anyway) can be in the world simultaneously -- and, of course, they can chat.
It's particularly impressive because it runs, at a reasonable framerate, in the browser, using Java -- and is created by a one-man team. Ten years ago, a much larger team would have been necessary, and you certainly couldn't have done it in the browser (or you could have, using VRML, but it would have been as slow as molasses in January).
The term "sandbox" tends to come up a lot here, and that's what this is -- but of course when we talk about sandbox games, we mean entities that still have metrics and allow us to establish implicit goals -- games like SimCity or Roller Coaster Tycoon played in sandbox mode. The only real possible goal here is "create something visually interesting," or possibly "destroy someone else's creation" (griefplay). Which is okay, and if you like wasting time this way, fine, but me, I stopped playing with Legos a long time ago.
So yeah, impressive application. But pointless, of course.
Now mind you, some of the directions Persson is taking the game are less pointless; e.g., he's announced a "Dungeons & Levers" expansion pack that would allow you to create a trap-driven dungeon environment, which at least holds open the possibility of player-created puzzle-based gameplay. But I feel about this much the same way I feel about games built inside Second Life: Why bother? Why not make a real game? In fact, why not make a real game with the world creation tools for people to create their own levels? Wouldn't that be less, you know, pointless?