Yep, thats great advice, I always carry some paper or a notebook and a pen with me everywhere, even when I think i wont be able to write for some reason (darkness, etc).
I suppose what should be considered is that games can be made in many different ways and methods. The only rule is that almost always (even this has some flexibility) you need to design the setting and character first. In games with more than one setting, that is they move forward in location, i think it's enough to first design just the first setting and then move forward. Contrasts between location moods or backgrounds are nice to have, especcially to give a message about the character or the game/story.
Anyway, after this, going straight from the end, something we know needs to be the conclusion, could work aswell, while thinking what caused something and why.
Another way is to first have an undetailed story flow without any detail on the character or setting, and fill those in afterwards, once the story/plot is finished.
Yet another way would be to first design interesting locations, characters, etc, and then link them with a story.
I am personally eager to try out a method which would in theory allow a lot of designer sympathizing towards the player. You could build the story bit by bit from the begining, statring with the begining status quo and moving on to a certain conflisct or problem that occurs/is caused by an NPC or the protagonist, and then put yourself "into the players shoes", so to speak, and imagine an area populated with certain logical items, people, concepts, etc, and how one could use it in imaginitive ways to solve a problem.
Now that I think about it, is a game possible without a conflict, or something going wrong/being wrong? Will a game about a perfectly happy person, who decides to walk 100 km to the next town, for no particular reason, be interesting? Of course for it to be not just an interactive story in which you need to click to progress, there need to be puzzles and obstacles, or meaningfull choices along the way which present conflict. But does the actual story require something being wrong or messed up? Hmm, I think it's just the wording that bugs me, of course the story needs a challenge, a goal, a problem, except it doesn't have to be something going wrong.
Anyway, thats about all my thoughts for now. Discussions like these really give me a lot to think about. Btw, I find a lot of my ideas come when I'm sleep deprived too, in the middle of the night or just the next day after very little sleep. Sortof like insomniactic ideas, heh.

There was one night on which I was sitting with lots of artciles on game design, and writing everything that came to mind (and there was a whole lot of it). "I'll just read a few more lines" became "I'll just finish these four articles..." I went to sleep at 5 in the morning.

P.s. Anyone know where people study game design, anywhere? I know there are people studying it, but dunno where.