Molleindustria is best known for games like the McDonald's Game and Oiligarchy, games that combine a strong political stance with actual gameplay, something that the games for change movement could certainly learn from. In Kosmosis, Pedercini, Molleindustria's auteur, has done something quite different: it's a five-day game created as part of an Experimental Gameplay challenge (the "unexperimental shooter" theme).
The underlying idea behind Kosmosis is that the modern shmup genre is an epiphenomenon of the capitalist war machine; in the designer's notes, he notes that the ur-game of the genre, Steve Russell's Space War, was created at MIT, which is heavily funded by the Department of Defense, and was based on a space-race inspired military fantasy. (In this, he is incorrect; Russell was inspired by E.E. "Doc" Smith's Lensmen series, which is indeed military fantasy, but was written in the 30s and 40s, long before the US/Soviet space race.) Kosmosis is, on one level, an attempt to imagine what shooters might look like if created in a universe where, as the creator puts it, "non-degenerated socialist values are hegemonic."
As you page through the game's short initial instructions, he makes the values he is trying to impart through gameplay explicit. I'll go through them, and relate them to actual gameplay.
"Thesis I: The task of the vanguard is to instill revolutionary class consciousness in the intergalatic proletariat." During play, you are a small red-tinged dot maneuvering across the playfield. Across the playfield are "prolets" -- small white dots -- and "reactionary war machines" -- large glowing green spheres. When you get close enough to prolets, they join you, and follow you about. You are the revolutionary vanguard, as it were.
"Thesis II: The people united will never be defeated." When you have enough prolets, you press space-bar and turn into a rotating sphere; in this form, when you intersect the nucleus of one of the "war machines," you can damage or destroy it. The people apparently remain united only for a time, however; eventually, you turn back into a cluster of revolutionary prolets, some of which break off instead of continuing to follow you. Note that intersecting war machines when not in "united" form, some prolets are extinguished.
"Thesis III: The role of the vanguard will diminish as the educated masses gain autonomy." As the game progresses, it becomes harder to maneuver your 'vanguard prolet,' and masses of revolutionary prolets will occasionally transform themselves into attack form to take out war machines. This makes the game harder to play over time, however. As the instructions say, "Don't try to dominate the swarm, join the swarm."
Swarm is the right word here, because the prolets are designed to follow flocking behavior. Indeed, the notion that this game is a "shooter" in any sense, except in the sense of an ideological reaction to the shooter as a genre, is risible; combat, when it occurs, is essentially a form of melee, the dominant aspect of gameplay is maneuvering your vanguard to convert prolets, and nothing actually shoots at anything else (not even the war machines at you).
What I find utterly fascinating about this game is that Pedercini clearly did begin with a set of precepts that he wanted somehow to encapsulate in gameplay, and has succeeded in doing so. Part of "success" is, of course, simple labelling; calling the white dots prolets, and your ability to cause them to join your swarm "instilling revolutionary consciousness" sustains the aesthetic the designer is attempting to create. But part of it is also inherent in the gameplay, that is, in the action of the player and the construct of the application: diminishing control over time, the conversion mechanism, autonomous action by revolutionary prolets in the late game.
Molleindustria, in other words, understands to the core what most developers of serious games have never understood: the mechanic is the message. You don't get across ideas in a game by embedding bits of text in an otherwise conventional game; you build your game around them.
As a game qua game, Kosmosis is interesting, but not amazing; it is, after all, a five-day project, and "thesis III" (diminishing control) makes it increasingly frustrating to play. Still, it is worth experiencing -- and notable also for its use of semiautonomous flocking behavior, a idea that a number of other indie creators have also explored (e.g., Dyson and Genocide Automation).