Approaching Content Creation

Get as much of the content in there first. Then fine tune and polish it. This way you can see how different mechanics balance with each other. Makes for a more robust product.

Otherwise once you polish enemy A, you may not want to address issues with it later. Which tends to produce very linear predictable gameplay as each piece is added trying to build on the last...you won't get a true picture of the overall product until the last content piece is in place and by then going back to address issues with earlier polished content could break the game.

depends on the game. For an FPS where the enemies are mostly just cannon-fodder whose only real strength is in their numbers, then filler art can do the same job as polished models. On the other hand for a fighting game where each opponent needs to be fine tuned for the gameplay, then yeah you need to do that polish up front before proceeding.

I think generally speaking, the first method (yours) is the better option given the scope of creating a game. If you were making a car from scratch, would you focus a great deal of your development effort into making the interior as plush as possible? Or would you get the entire car up and running well before focusing on the nuances?

Quote:

Original post by m4rv4 My idea was to press ahead and create as much of the game's content (e.g. enemies, maps, guns) as possible and focus on polishing it afterwards. This way you would have a good overview of what the full game will look like early, and you could take all things aren't fun yet and fix them. You would complete the game by filling all the gaps of quality.

I believe that that's a smart approach. Things will be more integrated in the whole, overall quality will be more consistent and - perhaps most important - you'll be able to modify things that don't quite work early on, without wasting too much time and effort. It's not fun to scrap a few months of work because it doesn't work well with other aspects of the game - so people often work around it somehow, which leads to a lower quality game.

Quote:

On the other hand, my friend suggests an approach of taking one piece of content at a time (for instance one certain enemy) and polish it up until there is nothing more to do.

There is always something else to do. A game is never finished. At some point you'll have to decide that it's good enough.

Quote:

That idea has the advantage that the developers get an impression of the final quality of the game early. The game would be completed by adding final quality content to the game piece by piece.

Previous parts will start to look bad once newer art comes in - quality increases over time, which leads to an inconsistent whole - or a vicious circle of work. Also, when things don't work well together, resources go to waste. This is how I used to approach level-design, but it left me with good-looking, yet half-finished levels. Things didn't work together.

However, it can be a good idea to create a few top-notch assets early on, to get an idea for the final quality. I sometimes create small style test maps, where I spend a lot of time on the aesthetics rather than the big picture. They serve as examples and guidelines for the real levels I'm working on.

In many ways you can't polish something on its own; it has to be polished (and balanced) in concert with everything else in the game. It'll depend a lot on the level of interaction (eg: RTS games cannot be balanced until everything is in place due to the interactions). To a certain extent this also applies to art work, where you need everything to look consistent and achieve a cohesive graphical style; focusing on high quality art rather than cohesive art is a trap a suprisingly large number of games fall into. Since making it all work together is the most important part, doing art work one bit at a time doesn't really make much sense either.

This will, as ever, depend on what you're making. RPG quests don't need to all be made together, since they can be wildly diferent (and should be different, in my opinion). Similarly with level designs.

I go for a middle between those two solutions, but lean toward your friend's. I don't mind leaving finishing touches for later, but I want things mostly done when I do them. I want to know that I won't have to go back there and mess with that anymore.

When you're designing and testing, it's different. You need to set things up and play around to see what's what. But if you already know what you're doing and where you're headed, then leaving things half done won't cause anything positive to happen. It's more likely to make you dread finishing it, in favor of working on something new.

A lot of the inspiration for working on game content is seeing it come alive in the game. Seeing it come alive in the game when it's half done rips that inpsiration away for the second half of it.

Get something working. Use debug boxes, temp art, no textures, reuse the same effect/bullet/model/sound/etc. everywhere that you would otherwise need to make something new.
99% of your time will be spent debugging errors
.9% of your time will be spent ballancing things
.09% of your time will be spent making things look good
.009% of your time will be spent adding "i wish" items/code/art to the game
This means that you have to make the most of your time. If you predesign every enemy, every prop, every type of scripted action,
then you might have time to get 10% of that actually into a game. But the predesign, and prototyping of everything into the game before anything
is finished lets you see what was wrong with your ideas before it is too late to fix things.
Expecially if you see that some enemy added a characteristic to the game that means an overhaul of all the other enemies (imagine not designing around "and this guy HEARS threats instead of seeing them!" but then getting to that enemy last)
And besides, you ARE going to get bored of working on some trouble spot, and effects and polish to stats can be done at almost any point
with minimal setup. This means you can polish things as you go if you get hung up on thinking about another system, and thus all your time is
better used.

Depends what you mean by "polish." If you're just talking about art, like models, textures, lighting, and the rest, it's usually best to have alot working before you start doing that.

However, if you mean fine-tuning and polishing by controls and gameplay, I think it's more important to have a character (if it's a platformer, fighter, whatever) working as smoothly as possible before starting design. Imagine if halfway through the design of a platformer, say Mario 64, you wanted to add a backwards jump to your character. All the design you'd already done would be affected by a simple addition, and there'd be alot to fix.

Oh god, IronWarrior hit a nail on the head there. If at all posible, add EVERY feature you NEED, but not every feature you WANT. If you NEEDed a feature, adding it late in the dev cycle is going to be a pain, and it IS going to break things. Lots of things. And where it might work fine on one level, there is always going to be some interaction on another level, where you system change now breaks that level. If you WANTed the feature, but it wasn't NEEDed, then add it when you have time, but note that it will break things to add it.

One thing you might want to consider, is that if you intend to approach a publisher with this game - your time is probably best spent making a very well polished prototype for "selling" the game.

Other than that, I would say that getting the bulk of the content in there and saving polishing and tuning would be my method of choice.

Remember that a successful game design normally requires several prototypes. You shouldn't be finishing anything off until you've prototyped it such that the core gameplay is working well.

Both paths are viable if the bulk of the programming job is done.

I also lean more toward your friend's method, because it is more rewarding and inspiring to see the result of your (presumably) long and hard work in the shape of something that looks close to a finished product. Like starting with a small area/map in a RPG or FPS game, and then slowly expanding.

In theory though, your method might be a better idea, as you have a wide overview of the content and it can help you avoid some of the pitfalls others already mentioned. But it requries working discipline, and well organized and detailed design.

I've known people like your friend. They take pride in their work and produce some great output. But they never finish any projects. Therefore the notion of the "final quality of the game" becomes a total illusion anyway. More often people get bogged down with one piece that they need to spend more time on, or change their minds about something half-way through and start again, or just lose interest and give up, all before you have a truly usable set of assets.

If it's your own personal piece of art and you want to enjoy the process of working with it, you can spend as much time on it as you like, but when you're working with others, you are just holding them up as a result of your perfectionism. If I'm a programmer waiting on a sprite or a 3D model in order to test my code, I don't want to wait for the final version before I can do that. I need a usable version as early as possible. That is far more important to me than seeing the finished product.

If it's a team project, be a team player - generate useful deliverables first, and polish them later.

Quote:

Original post by Talin But it requries working discipline, and well organized and detailed design.

I would argue that this applies more to his friend's method than his

I think most of you are forgetting that the engine and game design are supposed to be finished. That means a plan has been layed out, and you simply need to follow it.

You're working on game content alone. IE, modeling, animating, and voicing an orc warrior, or using a map editor to contruct a city. We already have an idea of what the orc and city are supposed to be like from the design. We just need to construct them.

If you ask me, there's no reason to leave things unfinished. That doesn't mean you need to cross every t and dot every i, but you should at least finish the majority of each task before moving on to the next.

Quote:

Original post by Kest I think most of you are forgetting that the engine and game design are supposed to be finished. [...] You're working on game content alone.

The original post said 'the engine is up and running'. Obviously that's a vague statement but I don't take it to mean the game is finished and you're just waiting for content. In fact, I'd say it's pretty much impossible to have finished the engine without having enough content to test it with, and that the term 'engine' implies that there is gameplay content (eg. scripts, game-specific code) yet to be implemented. All this requires assets for testing and iteration.

Similarly, even just content creation itself should be done on a parallel basis rather than a serial basis. Testing that different pieces of artwork actually work together, or that the level designs are actually playable, or ensuring that you have the right level of detail on your models, should be done before you have spent a load of time polishing details that could end up thrown away or not even noticed by a player.

Quote:

Original post by Kylotan Quote: Original post by Kest I think most of you are forgetting that the engine and game design are supposed to be finished. [...] You're working on game content alone. The original post said 'the engine is up and running'. Obviously that's a vague statement but I don't take it to mean the game is finished and you're just waiting for content.

I simply took it that the engine is operational, and you are building content for it. It shouldn't be so bad off that you're waiting on resources to program code. If you were, the engine wouldn't be up and running in respect to the content that you're building.

I would agree with your advice under normal circumstances. Having a stand-in character model to finalize the combat engine is of great benefit. But when you're on the third or fourth character, it's usually no longer an issue. Do you leave each character's assets partially finished as you move on to the next? I see little reason to.

Quote:

In fact, I'd say it's pretty much impossible to have finished the engine without having enough content to test it with, and that the term 'engine' implies that there is gameplay content (eg. scripts, game-specific code) yet to be implemented. All this requires assets for testing and iteration.

But are we talking about having enough content to test the engine, or about having all of the content for the game?

I'm half guessing, but that's how I perceived the question. He would build an entire village of buildings and people that are partially done, employ scripting and behavior for them, then finish them up simultaneously near the end. His friend would have the village blacksmith fully constructed before he even starts on the village healer, with no partially constructed buildings in sight, and no clear image of what everything will look like when the village is finished.

Personally, I could see some benefit to leaving things half-way done in some situations, such as when two buildings need to link together, or two objects strongly interact with each other. But I would otherwise sum each object up before moving on, fully animated, vocalized, textured, and ready to go. I also usually draw scenes out in 2D images before I begin working on them, so I don't need to step back and look.

Quote:

Similarly, even just content creation itself should be done on a parallel basis rather than a serial basis. Testing that different pieces of artwork actually work together, or that the level designs are actually playable, or ensuring that you have the right level of detail on your models, should be done before you have spent a load of time polishing details that could end up thrown away or not even noticed by a player.

For me, finishing and polishing are two different things.

I've gone both ways on this...

I've polished content up until the rest of the game looked like a turd. But I get discouraged as if I spent 25 hours on one game object, I know I'm going to have to invest almost 25 hours on every other piece. I get discouraged.

I've used nasty looking place holders which mean I'm not very happy showing off the game in any format whatsoever - not even demo. Alas, this is my preferred style. Only because in the middle of making some content, I get a bit inspired and may try something new out that I didn't plan for... or, by the time I get around to those tasks of fleshing out it might be months later for me... I may feel a bit differently about them.

Then again, I don't work with anyone, and I'm not under any time constraint. If I were then I probably would go with the polishing up what I can as I could then tick it off the "Things completed list" and prove work accomplished.