I'll start by asking my question. My question is... would it be practical for a game to contain some sort of basic engineering drafting system (like Autodesk Inventor) within the game in place of a crafting system that runs FULL physics simulations on the item being crafted?

As a degreed engineer, I've always had a great deal of interest in sandbox games with in depth crafting systems and an engineering approach. Why not take the crafting system to the extreme by being able to design an item from the ground up, like an internal combustion engine for example, within a drafting software and implement it into the game's sandbox world. The simulator would take into account the temperature, pressure, strain on materials, chemical reactions, etc.

The answer to that question is easy to answer: there is no such computer powerful enough to handle real time simulations with that many calculations. We're just not there yet.

But what if it wasn't real time? What if a sandbox game had a built in drafting and design software where you could design an internal combustion engine, for example, from the ground up. You then define the interaction with the machine (input and output), which in this case the input would be fuel and the output would be rotational power. The software then runs a full simulation considering all variables and derives a function of the output relative to the input (this may take some time). In this case the more fuel the more rotational power there will be. You can then place the machine in the game's world where it has a graphic and a simple function of input verses output, thus greatly reducing it's computational demand. No real time simulation is taking place, just a simple function.

I also understand that such a simulation would be strenuous even on some of the most powerful computers made today, especially when dealing with a system that has multiple inputs and outputs. But perhaps that could be solved by making the game constantly online so that the calculations during the drafting/crafting portion could take place on a powerful server elsewhere.

What would be the computational challenges of a game like this?

This is just an honest question coming from someone with little knowledge of game design, so please respond with that in mind.

For one thing, you'll be opening yourself up to the wonderful world of combinatorial explosion. The more options you have that can interact, the more options exist. The more options exist, the harder it is to develop and test. You'll get chemists complaining about how when they mix three specific chemicals it behaves wrong. You'll get electrical engineers complaining that when they crafted something with a specific trace pattern it should make some cool electronic effect. You'll get all kinds of other topical experts complaining that their topic is done incorrectly.

For another thing, you'll quickly leave the realm of "fun game" and enter the realm of "tedious real life". You mention constructing an internal combustion engine. Imagine if in your relative haste (you didn't want to spend 3 real-world years doing it, so you only spent six real-world months building it) one of your piston shafts is designed slightly less than perfect with a slight taper, or something blocked the lubricant flow, or a segment of the crankshaft was weak, ultimately causing the engine to break the first time you use it. You just lost a bunch of your life to a pointless aspect of a game. Games are meant to be fun, to be an entertaining distraction of reality. The less of that you've got, the less fun the game usually is.

There is much more in game design about the topic. Depth and complexity are only slightly related. Depth comes from players being able to understand and use options available to them. Complexity by itself does not add depth to a game. Complexity can add options, but it does not mean those options important to the game, nor does it mean the options are able to fit in the player's mind, nor does complexity make the game more fun. Complexity generally makes games less accessible. The best games work hard at removing complexity, providing a very small number of simple mechanics which interconnect in powerful ways. Simple rules, simple mechanics, which combine to create depth. Not complex mechanics, those tend to result in boring games, chaos, or systems the players cannot understand.

If your crafting system requires an advanced degree in mechanical engineering, or requires a level of expertise usually reserved for masters of the craft with years of experience in real life, you probably won't have many players.

I see the problem as such:

You seem to be very knowledgable about a topic you seem to have a high passion for. The first point is why you most probably could handle such a complex system and really create something from it that would work. The second point means the actual process of doing so does sound like fun to you.

How many people playing your game are mechanical engineers (or at leas know enough about the topic to be able to use the fully blown system), and how many of these actually share your passion for the topic to such an extent, that when they come home from their work (as mechanical engineers, maybe?), they want to sit down and do something that could also be defined as "more work"?

Your niche suddenly became very narrow. To all other players, either the system is optional and will certainly not be used by them, or it will be one big barrier to their entry into your game world.

If you tone it down sufficiently, you could make it work. But now you take on one of the biggest challenges in game design in finding the right level of abstraction ("reducing to the max", as you could say), you still will face the problem frob mentioned (every combination SHOULD be tested, else you will have cries of overpowerdness in your game), and you will still have to fight to make the system both fun, accessible, and make sure it does matter in your game.

Thank you both for the practical feedback.

Sometimes I forget other people don't study engineering related topics in their spare time like I do... Yeah, I know, I'm weird. Your responses were exactly what I was looking for. I now agree that this idea would satisfy very few people.

On a separate note, I still like the idea of a game that introduces concepts of engineering in a very simplistic way. To me, it's one of the best ways for older kids discover if they have an interest in technical stuff. My dream would be to develop a game that makes players work together to design and operate a simplified complex system (like a power plant or something) without taking away the fun, even if it doesn't appeal to the majority of people. That topic doesn't really belong in these forums so I'll stop there.

Thanks again for the responses.

On a separate note, I still like the idea of a game that introduces concepts of engineering in a very simplistic way. To me, it's one of the best ways for older kids discover if they have an interest in technical stuff. My dream would be to develop a game that makes players work together to design and operate a simplified complex system (like a power plant or something) without taking away the fun, even if it doesn't appeal to the majority of people. That topic doesn't really belong in these forums so I'll stop there.

Why shouldn't it belong here? Because it could be implemented as a serious game?

I think your idea is very valid, and if you manage to simplify it enough, you will find that it will appeal not only to a small niche of players. there is a reason why crafting systems and games with indepth customisation and item generation systems like for example Borderlands have found a lot of fans.

Give players a deep crafting system without overwhelming them, maybe develop a system where the beginner can get something to work, but an expert will be able to create a better version, give them a very good tutorial / wiki system that will make sure the non-engineers get into it quickly, and everyone interested can reach expert level, and players will be hooked.

Just make sure you don't need an engineers degree to master your system.

collaborative building might also be a major draw for your game. Isn't something like this already possible with for example minecraft? You could take it to the next level for example by giving the player the choice of a job class, that defines what they can build. Wanna build a technical truck for a shooter game?

Now you need an engineer to plan the car and create the running parts (which define a lot of the technical specifications of how the car drives), a designer to create the outer hull (which defines much of the look, but also some technical stuff like armour, aerodynamics and so on), a gunsmith to create the gun on the back.

Maybe someone to create the leather interior?

Anyway, anything will work, as long as it is streamlined enough for the average player to understand.

You may want to check out Player Studio for Everquest Next Landmark. It features a very powerful voxel-based, metagame crafting system. You create a template in Landmark, and in the future that template could become available for purchase in EQN, based on voxel amount and peer review. Then again, this is a next-gen MMO and not some minor indie project.

As a general rule of thumb though, I think you should cap the steps of any crafting system to something like this (and usually within that one program and not extending it across software, unless you make each segment a playable thing onto itself within that parent software):

1-2 step recipes: 50-75% of all craftables. (smelting/refining and a subsequent crafting step with those materials)
3 step recipes: ~25-50% of all craftables. (using 2-step crafted items as part of the recipe)

4+ step recipes: 0-25% of all craftables (special high-end items)

Interpret this openly according to your own structural paradigm. But you don't really need to go miles outside the box in order to create a deep, robust crafting system that makes sense and is relevant for the player. It's enough to have representative crafting, and not forcing players to craft every single nut and bolt to exact millimetric specifications. Too many steps and people will quit your game from sheer boredom and frustration.

Additionally, instead of having 4+ direct crafting steps, a lot of games tend to revolve around (1) tiers (i.e. some resources have greater rarity, zone/biome restrictions, level requirements, etc) and (2) contextualized steps (e.g. brewing in Minecraft requires a brewing stand, whereas smelting requires a forge. Creating either items is done in just 1 extra step, although the brewing stand involves a higher-tier item whereas the furnace can be crafted within the first five minutes of play).

There's also the idea of having additional parameters like skill, financial costs and other circumstances (e.g. a full moon for certain fantasy weapons or a workshop with opening and closing hours). But be careful to differentiate between occasional and mass production. Having the odd super-complex recipe for something you craft only once can be refreshing and interesting, but not if the item is something the player will be producing en masse (unless that mass production is itself a one-time event late in the game and/or can be coupled with other activities that allows the player to fire-and-forget).

There's a lot being done already in games, and even more that is not being done but still fully doable (if you're designing a crafting-heavy system like, say, Eve Online manufacturing or similar).