i think gameplay is everything.. so gameplay should always come first..

awesome graphics and no gameplay is not a game its a tech demo.. and lately on the AAA games department i have found many games which has breathtaking graphics and yawn giving gameplay ..

i still like alot old games which you still play because they are fun not because they impress visually.. examples are  mario 64 ,mario kart 64 , star fox 64 , zelda ocarina of time , mario world, final fight , f-zero , star fox and some others

some games were legendary just because of its hooky game play .. graphics were good at that time but right now they are just dull or even very bad and only its gameplay was left..

we will see if all current multi million AAA outstanding graphics games remain that amazing in 5-7 years..

I wanted to add my two cents that I feel that XNA is such a perfect framework for people to explore new or improve upon older and existing gameplay methods.  I think that's one of it's major advantages.

Sure, you want it to look nice, but since I don't have an art person yet, my simple 2D game is all about creating interesting gameplay and I'll worry about the rest later.

While I believe that "gameplay" (I really don't like that word) is more important than graphics, I actually like to focus on the graphics programming first.   I am horrible at 3d modelors, and have given up on being an artist, mainly because I want to focus my programming skills.  So I've resolved to just have other people do my art.

It might be just because I really want to be a good graphics programmer, but I also think having a good graphics foundation from the beginning helps you get a good feeling of the end product of the game should be.  I still end up spending way more time on the other parts of the game.  I just like starting with something looking decent right away that way I know my game sucks because of the gameplay...not because of crappy graphics, and then work from there =)

On the 2D game I have been working on, I do both at the same time...For example when we were testing death with“lava“ or what ever other dmg tile we had, I first just used a red tile, then upgraded it to loook more like fire.

 

At first when I made my plyaer it was just a box, but as I started adding player elements and a player class, I wanted to plish it a bit. To me it is more rewarding and keeps me motivated to see the game slowly become“polished“ (polished to my standards anyway)

 

I've taken a different approach. What I'm doing is building a graphically sophisticated engine now, then when I get ideas for games, I can get immediately to the gameplay part. I have tried to make the engine a flexible as possible, in order to accomodate the crazy ideas I get for games.

I guess I have to fall into the graphics first category. I get an idea, and ill start some basic coding on it.. But I'm never satisfied with a blah background, or a red square.. so i go and get my graphics that I need right now (get or make em.. heh).. then code more.. then when I need another, ill jump out to the graphics again.. I typically try to do the graphics once for each thing, and get them to look as good as I can. It keeps me motivated.. particle effects? its not polish, its a step closer to being done and has to be added when Im at that point.. No placeholders, the final effect is done then.

If i have been working on a project for a week or two, and then come back and run it and all I see is boxes.. and red squares, well Im not going to have much enthusiasm for that project.. It needs to grow graphically as it progresses in the code..

Gameplay is sure important but the graphical output of a game is vital for me too.
The reason to that is because I'm not only a programmer, but also an artist and I think it's awesome to see the creations I make "live" in the game.
Games should produce emotions to the player and to make the player feel afraid or excited is incredibly hard to do without good graphics.
Mario 64, zelda or other old school games is fun to play but do they bring those emotions?
I can't say I'm excited when playing mario 64, but I think it's great just to play and enjoy it.
So the combination of both really make the good game, not just one of them.
And you're right about newly created games. Crysis really has extreme graphics but there's much complaining about the gameplay. (And I agree)
The AAA game departments need to find a balance between these two parts when developing a game. Otherwise it will never be the "WOW" feeling.

I like to focus on the graphics first, because well it doesn't matter how good your gameplay is, not many people will play a game that doesn't at least look good (2D or 3D). It's like sitting a guitar that looks old and beat up next to a new one and asking someone which one they want to play, unless they know that the wood sounds better the more it's played, they are going to go for the new looking one.

However I'm not saying game play isn't important, you should have a firm idea on what your game play is going to be like even before you start working on graphics, and vice versa. They have to fit together for the end product after all. Having really pretty graphics and good game play but feeling like they kinda don't belong together isn't to common, but I've encountered it before and it really does slowly get to ya as your playing a game like that.

you should start with gameplay and then graphics

Although everyone will say gameplay, its not exactly true.

Great gameplay generally is an evolution of the assets and code. You may think of ways to make a game fun, but it really takes constant revisions to get a game really addictive.

I go with gameplay first everytime i start a new project.  In my opinion a game with great gameplay and bad graphics is far more fun to the user than a game with ultra-realistic graphics and bad gameplay.

I always work with asynchronous steps. But gameplay is most of the time the 2nd of the 5 things I do..

1) Develop a scalable "engine"... whatever I'm doing I need to create a scalable engine. This contains game states (playing, pause, etc). Render states (backbuffers). Input states. Reusable shaders, camera logic, animation logic etc.

2) Start with gameplay (game logic). If your first step is right this is a lot easier

3) Almost simultaneously you need to start with basic graphic and sound assets like menu's, dialogs, effects and dummy objects.

4) When you have finished the gameplay logic, you can start by replacing the dummy content with final content. And start filling it with more graphic/sound content

5) After touches. Create shaders for cinematic effects, color corrections, bloom, etc etc.

Or simplified: start with a solid base, then you can create all other stuff independend of eachother.