Here’s a great article on game writing from Susan O’Connor via Gamasutra. She makes a great point about the focus of games (and television) on the middles of stories. While I don’t agree with her definitions of when the story begins and ends (I don’t think that the primary conflict is the entire story), there are a lot of good gems in there.

For instance, there is this on multiple endings:

“If you have multiple [endings], I think two is the maximum. It either works out or it doesn’t. When it comes to multiple endings — well, you might as well just break your writers knees with a crowbar, as trying to make just the one amazing ending is hard enough.”

I’ve always had problems with multiple endings for a variety of reasons. The primary reason is that you are creating a lot of unresolved detail in most cases. Let’s say a player spends the first 50% working towards the “good” ending and the last 50% working towards the “bad” ending, to simply. He can only get one of those endings, so all that writing work making the character out to be a saint in the first half will never be resolved. Additionally, writers have to make multiple endings that logically follow from the events. Usually when you watch a movie, you have in your mind (at least I do) a number of paths that the story can go down: the hero can win, the hero can beat the bad guys but lose something of himself, the hero could be wiped out (almost never happens), et cetera. As details come to light, some of these endings disappear as incongruent or non-viable. But if your game supports all these endings, you can never include details that shy away from a particular ending (or more importantly towards a particular ending) without creating an incongruence.

Writing, like audio seems to be vastly under appreciated in the industry, which is why I suppose they have a separate GDC to segregate them.