I do not think that a GDD is possibly the most important thing in game design.. It IS the most important thing.
A Game Design Document is not only the basis for your game and help you improve your own developmental direction, it is also a living document.
Not only will it help you realize aspects you may not have thought of before, it also helps you realize the opposite; what you cannot currently do. This might be because of limitations of time, tech or just cost (not just monitory, but also time, effort, resources for team and or client, etc.) Either way, this is actually a good thing since it will help bring balance to your development and help create a stable development plan from simple tasks to milestones.
Remember, a GDD is meant to be adjusted and worked on through out. And you can even plan for further development goals and designs for later on.
A rule of thumb that I was taught years and years ago.. and still applies today...
A wise old man (I Hope he is not lurking around on these forums to hear me call him old) once told me years ago, "if it is not in the GDD it wont be in the game."
Though this sounds harsh, it really is true. Parts of your idea might make it, but the full concept of what you have in mind and what it could be will not, else you sacrifice something else as seen below.
I have seen development teams say, "Eh, We are going to do this, i made a few notes, but we will get this working". Granted one team did in fact get that one aspect into the game, but it was incomplete and buggy as heck. Since it was not well planned out, with the team leads working together to make sure this one game play aspect was not noted just once, but in all categories of the GDD it should have, it was later removed from the game once it went live.
This same development team also suffered, sadly, because some were so focused on this one new aspect (while others were not even clued into this idea because it was not documented) that other parts of the development were ignored or pushed back too many times. This means some very important aspects of game play (and rather simple in the end) were not completed and were not optimized (art, scripts, terrains). This causes some major latency issues both server side and client side in one particular area of the game.
Now that might be a drastic example. However it is really true in the long run.
If you spend that time working on the GDD and documenting everything you can- and keeping the rest of the team updated, then you can only have good results in the end. Sometimes, taking it slow at first will really help you win the race later on- yes, i know another bad analogy, I seem to be full of them tonight.
I know a post was given about a few styles of GDD documentation, but do try to remember a few things:
Document everything in order of categories and try to stay organized in your layout.
It is okay to repeat some information if it is relevant- or reference back to that section or additional documents.
Keep a running list of all attached documents that are part of, but not in the GDD.
Keep sections updated for your development/design team.
Keep record of what has changed in each version- be it spelling or new additions or changes.
Remember a GDD is not your Game Design Proposal, this is meant to be a living document and perfected to help you and your team reach your goals.
A few often overlooked things in addition to what the others have said, but I hope that helps some. I will not even state how large our GDD is, or what additional documents we have attached to it, but I do know that despite everything, and working on this for well over three years (with two full revisions) I am still not done.
Just my 2¢ worth,
Peace,
Sarrene'