This is interesting as this is the type of post I usually post in other forums ;-) Traditionally i'm a seat-of-your-pants kind of guy. I realize that at first your game probably isn't going to be that fun, unless you're very very experienced. So generally I start with an idea, jot a few notes, and dive into coding, making sure not to worry about graphics until I get something that works. Then I test it out if it's not fun I tweak it 'till it hurts ;-)
I used to do this alot, until just lately I started on a new game http://www.games-afoot.com/touchtiles.html
So I started by making a design outline and a mock screenshot, so far it's helped me in deciding what I want to do and how I want to do it. Although I haven't gotten to a point of actually making it a game yet. It's still in the early stages.
Once I had an idea for a robotron/berserk style clone and I tried to code the collision stuff for it, unfortunately I got nowhere, so I figured hey maybe I *should* design everything I possibly could. So I spent days writing out everything that would happen in my game, how it would look, how it would act, etc. In the end I got so tired of planning it I didn't want to do it anymore. Moral of the story is - planning is good as long as it doesn't result in burnout.
Currently with the game I mentioned above, i'm planning on an evolutionary planning process even though I did some design work early on. Basically once I get this into a working "game", i'm going to test it out, see if it would actually be fun. Then I will decide if i want to add anything, if I do, i'll write a new doc based off of the old one with my changes, then code it and test it and tweak it, and so on. It may turn out to be totally different than I started with.
Another idea i'm toying with... I'm a nut for anything to do with retrogame development, and I mean actually reading and learning how the pioneers started. Most of them being programmers and not all into design and art (especially with what they had to work with originally) would code a cool feature they liked, like one would do with a demo effect in demoscene-style graphic/music/art demos, and then they would base a game off of it. Well I know there are several different things I don't know about programming, say pathfinding for instance - I'm actually trying to learn how to implement that right now, but say when i'm done, i'll build a game around it. Or basically any concept you don't know about programming, you create it and test it then attempt to build a game around it. It might turn out decent, but it's just an idea i'm toying with at the moment.
I used to do this alot, until just lately I started on a new game http://www.games-afoot.com/touchtiles.html
So I started by making a design outline and a mock screenshot, so far it's helped me in deciding what I want to do and how I want to do it. Although I haven't gotten to a point of actually making it a game yet. It's still in the early stages.
Once I had an idea for a robotron/berserk style clone and I tried to code the collision stuff for it, unfortunately I got nowhere, so I figured hey maybe I *should* design everything I possibly could. So I spent days writing out everything that would happen in my game, how it would look, how it would act, etc. In the end I got so tired of planning it I didn't want to do it anymore. Moral of the story is - planning is good as long as it doesn't result in burnout.
Currently with the game I mentioned above, i'm planning on an evolutionary planning process even though I did some design work early on. Basically once I get this into a working "game", i'm going to test it out, see if it would actually be fun. Then I will decide if i want to add anything, if I do, i'll write a new doc based off of the old one with my changes, then code it and test it and tweak it, and so on. It may turn out to be totally different than I started with.
Another idea i'm toying with... I'm a nut for anything to do with retrogame development, and I mean actually reading and learning how the pioneers started. Most of them being programmers and not all into design and art (especially with what they had to work with originally) would code a cool feature they liked, like one would do with a demo effect in demoscene-style graphic/music/art demos, and then they would base a game off of it. Well I know there are several different things I don't know about programming, say pathfinding for instance - I'm actually trying to learn how to implement that right now, but say when i'm done, i'll build a game around it. Or basically any concept you don't know about programming, you create it and test it then attempt to build a game around it. It might turn out decent, but it's just an idea i'm toying with at the moment.