Firstly, I see that I've been away for a month now - a variety of tumultuous things have happened in my life recently, so I don't feel too guilty about it.
So anyway, from Raph's Website I just got the link to Canabalt and it got my thinking about short games. You see, the game is very cool - it procedurally generates levels, has nice art, and is capable of producing a "that's badass" experience with simple presentation and single-button interaction. However, it is also immensely unsatisfying, because once you get over the "ah, that's cool" reaction, there's not much else to the game. It's not simply that the game is too short or needs more features - by concept, the game cannot provide anything more than it does.
The issue here is that it takes time for me to fall in love with a game, and even when I like short or experimental games, I never really fall in love with them. This troubles me because I care about art, and I care about games, and I'd like these two concerns to overlap at some point, and it seems that experimental games are an important part of getting there.
Maybe it's just that no one has hit quite the right chord with me to make me fall in love with such a short game, but I don't think that's quite it. The problem is that games, I think, need to allow for exploration, and if I feel like I get it right away, then there's nothing to explore. Exploration here doesn't simply refer to having a big map to cover, though that can work (I'm looking at you, Bethesda games!). A player can explore a game mechanic, or different strategies, or possible narrative paths, or a single narrative path, for that matter.
I'm not trying to discourage anyone from making experimental or short games, by any means, but I do worry that games that don't at least provide a couple hours worth of exploration (whether its by having a two-hour long linear path, repeatably simple gameplay with deep strategy, or just lots of stuff to discover in a virtual space) are ultimately limited in how deeply they can affect us.
That is, for games and art to converge, I think we may need to have larger-scale experiements.