Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Aqizzar]

If only to prove that I can actually write crap, and that Josh doesn't have a monopoly on pie-in-the-sky game ideas, I threw this wall of words together about PC-game pitch/design/wish I've had brewing for a while - The Metropolis.

Spoiler: This is the right frame of mind. (click to show/hide)

There's not a whole lot to this in concept, but plenty of room for detail.  At it's heart, it's a tried and true genre - city building.  Or even nation building.  In short, you the player are the dictatorial ruler of a gargantuan retro-futuristic arcology, responsible for building and maintaining (mostly maintaining) a vertical city of millions.

Dimension
The big change in dimension obviously is that you build a city vertically, not just along the ground.  Not in a sky-scraper sense, but like Dwarf Fortress, where you need full movement in three dimensions for infrastructure and such (more on that later).  The setting would encourage building upwards instead of out (since that's the point): the ground is thoroughly polluted and unstable, plagued with raiders, and other dystopian crap.  For architectural reasons if nothing else, it's always cheaper and easier to build on top of existing structures instead of adding new foundations, at least until the whole building is so tall and heavy that it risks collapsing of it's own accord.

Of course, the problem with building in three dimensions is being able to see anything inside.  That's easy enough to fix by rotating through parts and systems that highlight while the rest of the building wireframes out of the way.

Scale
I'm not really sure how you could convey the scope of a nation-in-a-building.  Spreadsheets describing how many thousands of people were eaten by mutants gets dry without a sense of scale, but zooming in for a street view anywhere is technically taxing and still doesn't do much.  The nature of the constructions would work with the building requirements, like giant water mains, flying cars, train-sized elevators, Soviet-big factories, and other things the player could see animating and working.

Distance
Some of my favorite city-building games were Caesarand Tropico.  But the problem I had with them was that you were personally responsible for placing and staffing every industry and aspect of whatever.  At the opposite end of control you've got SimCity, where at most you draw zones and place city buildings.  There's a middle ground.  As the City Overlord, you could place just about anything by edict, especially civic stuff like infrastructure and the arcology frame itself.

However, even that is too much to ask of a player without sacrificing the scale of the building.  Instead, anything but the largest constructions would (usually but optionally) be handled by managers appointed by the player.  This ties in with other areas, so I'll frequently mention it again, but throughout the game the player will rarely control building or activities directly.

Where it pertains to Distance, the player can lay down their general idea of what an area should look like with utilities and zoning, but the actual details of what's in an area would be the computer's responsibility.  This extends to non-civic functions; if the player wants more cybernetics factories you can specify tax-breaks or bonuses, or threats, to encourage more to move in, but you can't just mandate their construction.  If you want an autogyro dealership in an area, you could lay down an incentive-marker, Majesty-style, to draw one there.

You could also just build state-run factories and facilities for any product or purpose if you think its important, but eventually that would build up too much micromanagement for the player to handle, and by design.  Managers as well would have an elastic limit to how much stuff they can control over how big an area, and as their attention and yours is stretched, problems would start popping up with inefficiency and crisis-management.

Automation
To repeat the obvious, there will and should always be way too many decisions for any human player to control, and ideally the simulation shouldn't even be pausable, so the game keeps running while you're controlling it.  Most of the player's responsibility will be in attracting, appointing, and controlling managers for areas and functions.  This wouldn't be as simple as just clicking the Governor option; managers would have requirements, specialties, and problems of their own, and their own set of goals and relationships.

Exactly how far down the player could control, and how far down the managers could control, should be different and opaque.  For instance, you appoint the Thought Police Chief in the South Spire, he then appoints his own managers of different departments and neighborhoods (maybe with player approval/rejection, to preserve the lower-level control), sets salaries and policies, and stations squads at problem areas the managers perceives under his rubric of threat assessments and priorities.

While the player could conceivably issue Executive Orders for everything from water junction priority to uniform codes, realistically and technically speaking the player would never approach such low-level decisions.  But that doesn't mean they shouldn't be there.

Interaction
That kind of detail should still be present, even if it's generated on the spot when the player looks at it.  The important point is, just because the player doesn't control something, that thing should still be there and still be visible, if however remotely.

On a completely different subject, and for me the real heart of interaction, is how the player relates to the people living in his sky-scraping-slum.  Obviously the population isn't as monolithic as their home, and should break up along at least two or three lines.  Obviously, economic status would be the huge one, drawing on class-warfare inspiration as old as H.G. Wells.  Actual neighborhoods (not necessarily the same as the player-defined manager-regions, but based more on architectural access) would have their own AI and character details, so you could have inter-area gang wars between groups of similar status.  More axes are always possible (religion? culture? corporate employment?), but I wouldn't want to muddy the water.

Where population-groups and opinions become important, and where the gameplay should really come together, is the player's ability to directly interact with them.  Namely, with proclamations and messages, and all the ways that your stated intent or thoughts, the population-groups belief of your intent and thoughts, and the actual actions that take place.

For several examples:  Workers strike down in the sludge pits, so you can threaten to turn off the oxygen flow if they don't get back to work.  If they believe you, they'll stop rioting; but if they don't, now you have to cut off the air, at least until they get the message, or they'll know from then on that they can push you, and that knowledge could spread to other downtrodden areas.  If the air turns off by accident to the middle-class decks, and you say "hey, sorry about that", they shouldn't get as pissed as otherwise, but if it happens a lot no apology will make them not mad.  If the nobility get pissed about your industrial taxes, tell them you'll withdraw the riot police; when they scoff, you could do so, only to find that by not paying their taxes the nobles have enough money to hire their own security anyway, and now they're really confident without your management.

There's a metric fuckton of AI relationships and judgment infrastructure underlying those relatively simple examples, but the important idea is that the game should track what you say you're going to do across time and economic/mechanical activity.

Randomness
This last point builds back into the Distance ideas, but to reiterate, I like to see management games that create not just their own problems, but their own flavor.  And not everything should be forseen and perfectly knowable by the player.

In many games, the wear-and-tear present on a structure is knowable to the percentage, and remediable without worry.  Structural integrity should be an opaque guess as good as the last inspection team, and even then you never know when an accident or hidden defect might break something.  When something does wear or break, and needs to be repaired, it won't actually look or function the same way it did before, complete with patchwork and resettling.

Likewise, just because the player wants something built, and should be able to specify for instance what a bridge will look like and where it will span, doesn't mean you'll have complete control over every kink and panel.  Every construction will have some amount of random corner-cutting and redirection to match the specifications, available space, the manager's personality and team, and the skill of the construction crew hired/coerced/forced to build it.

So for a large example: Over time, the megastructure would take on a jumbled, organic look without the player ever trying for it.  Building plans would go awry or over budget and wind up wholly different from their plans.  Quakes and collapses would send whole chunks clattering down to lower levels, while the whole structures slowly compresses into a twisted mass over time, forming it's own foundation of what used to be spires.


That's nowhere near everything I could possibly say on the matter.  I used to think that wasn't enough to start a conversation, but I know now that people here like spitballing queries and concepts for this stuff.  There's lot of other things to talk about, like politics, war, inter-city trade, the environment around the city, the cultural inspirations and references, and all that other good stuff.  This is just me jotting down ideas and throwing them up for shits.

[IronyOwl]

Looks neat. I'd say it could be the next SimCity.

Personally, I love the notion that everything would be inefficient and nonsensical simply because the player didn't have time to go in and fix it- that some janitors would be wearing orange wool suits while others got paid as much as middle management and were required to have a degree in the arts. Both realistic and ensures there's always SOMETHING for the player to do, if they've got a mind for it.

The diplomacy notion is also interesting, and could also feature something I'm not sure I've ever seen in a game before- tone. Usually there's a certain tone at least implied in various dialogue options, but I think it'd be pretty awesome to be able to range between "I'm really really sorry, that was my fault and it won't happen again" and "Yeah whoops, kinda busy here, stuff happens." It'd also probably be one of the more innocuous ways to immerse players, and add another layer to management in general without just having a single Reputation stat. Citizens could learn that when the boss says he's really sorry, he means it'll be fixed Tuesday and happen again Wednesday, or the elite could be flattered that someone who's never sorry for anything is genuinely apologizing to them.

[alway]

I've been thinking about doing a social simulation based with AI using neural networks and genetic algorithms for quite a while. And I think this idea would actually be fun to play and fit well with such a simulation; it has me interested at least. But, alas, I have my current project to finish... Not to mention the fact that I have yet to get my cubes to render properly. 
You have definately given me something to think about though!

[Armok]

To bad this will never be done, it would be awesome.

[Ghazkull]

It somehow reminds me of Warhammer 40k. The whole inefficient
gargantuan city thing, with upper classes in the higher spires and the poor industry workers at the bottom of them having to deal with bad healthcare,food, mutations, Mutants, Gangs, Parts falling down from above.

Actually it would be fun to introduce something like second class(if the industry workers weren't already second class... ) humans, namely mutants. They are living in the deepest pits of the spires and are living from the leftovers, of the leftovers, of the leftovers...instead of paying taxes or having a right of living in the city they would be treated like vermin and hunted(if Religion introduced) by some radical groups. Now you could either deal with the problem by either sending headhunters/security forces/armed mobs down and start something like the "Annual Mutant Hunt" to get rid of most of them.

Or you could give them rights, like basic human rights, right of housing(so that they can life at least in the lower spires), work rights (so that they can earn money legally) and so on until you finally give them Emancipation and allow them to even get in the noble spires. but this liberal way comes at an cost: the rest of the population, the religious people, later on even the nobles will start rebelling or rioting. For example lynch mobs will begin to slaughter mutants and even some government officials(for being mutant friendly), some racists will form extra-radical groups like a Mutant KKK or terrorist cells will bomb government buildings if you treat the mutants too kind. At best you can slowly but surly integrate the mutants into your society a worst it will end in apolitical disaster for you, sending the whole Hive into a awesome df-like tantrum spiral....


On the other hand if you do nothing about them they will get more and more and settle on potential bulding space, interrupt your workers, and being all around unproductive leeches forcing you to do something about them, or facing the possibility that yomeday they might rally and overrun your city...


wow that was much writing....

[JoshuaFH]

I just got around to reading your idea Aqizzar, and I really like it.

One question, would this be like in the Sims, where you have no set goal outside of build/manage/expand/etcetera into infinity or do you think there could be missions with there own set objectives?

[Aqizzar]

To bad this will never be done, it would be awesome.

Tell me about it.  I used to think that maybe, just maybe, I could bung together an all-text prototype, to at least demonstrate the idea.  Turns out I'm a terrible, lazy programmer.  Oh well.

It somehow reminds me of Warhammer 40k. The whole inefficient gargantuan city thing, with upper classes in the higher spires and the poor industry workers at the bottom of them having to deal with bad healthcare,food, mutations, Mutants, Gangs, Parts falling down from above.

I was certainly inspired by Warhammer's hive city kick.  That of course was itself a refinement of nearly a century of predictions of mega-cities and super-industrialization, going all the way back to H.G. Well's When the Sleeper Wakes.  Warhammer did flesh out and run with the idea to some interesting places, especially the ecology of arcologies, and the interplay of social levels and physical movement of places from the upper to lower reaches.

Your ideas about classes and "mutants" lets me expand on the distance/randomness idea.  Just because you zone an area to be used some way, doesn't mean it will be.  That just tells your law enforcement and infrastructure how to treat it.  Squatters moving into unwatched space and sapping resources from neighboring lines should be a constant problem, as another incentive to expand proper usage and build new areas.

One question, would this be like in the Sims, where you have no set goal outside of build/manage/expand/etcetera into infinity or do you think there could be missions with there own set objectives?

I'm not sure really.  Of course, goals are just ideas that you add to free-building, and there's always room for both.  An actual mission arrangement could come from inter-city politics, maybe even over-arching empires that demands your city to produce resources and raise armies, or face reprisal.  Likewise, you could be moved back and forth between controlling different cities facing different problems as they arise, both to utilize all your skills, and to prevent any one governor (like you) from building an entrenched power base in one city, the way Roman governors used to move around.  I haven't thought much at all about what the "setting" would look like and how that would shape gameplay, but the concept writes itself pretty easily.

One issue that does confuse me is how to convey time.  Obviously a building large enough to hold a city isn't built quickly, nor do the people who occupy it reproduce very fast.  Every piece of literature describing megacities either relies on superscience to explain where they came from, or assumes that they took uncounted generations to raise.  And any way you slice it raises the sticky question of who the player actually is and how you exist in the gameworld.  I can see three ways of presenting time vis a vis the player.

1) Time Compression.  Megastructural engineering is carried out at SimCity speeds of months and occasionally years, people move in from the "countryside" (which is even more blighted and dangerous than your superslum) or are already living in the crumbling, abandoned recesses and reveal themselves as the conditions improve.  You the player are a single ultra-nobleman appointed/elected/dueled/scrabbled into power, long lived maybe but mortal, who has a presence in the gameworld, however tiny and remote.

2) Realtime Immortality.  Mile high city-cubes take realistic decades or centuries to build and populate.  Maybe they were even built long before the game began and it's just your job to keep them from tottering over completely.  The player is still a single person, but rendered effectively undying somehow (short of a very powerful rebellion).  Certainly a superscience or supernatural explanation for the player-character's longevity could be extended to why anyone pays attention to you at all, the city being enraptured with mortal fear under their God-Mayor and administration.  Whether your managers would be similarly long lived is another question, but I think it would add to the experience if you had to continuously raise and replace your mayfly lieutenants.

3) Realtime Dynasty.  The construction would work the same way as above, but the "player-character" is not one character.  Instead, the player would represent the collective will of a dynastic ruling family/clan/organization/cult/whatever, in the famous manner of the Total War series.  None of the arrangements allow for democratic ruler change (without the game ending anyway) or conflicting personalities within the Mayor's office, but this system really brings that to light.  You'd have to have some control over which members of the ruling enclave hold what power and how they act with it, but you'd have to contend with their own attitudes and predilections somehow, such as their strengths and weaknesses in management or how the population groups react to them.

[Supermikhail]

I'm completely out of tune, but some side of the concept, the picture in the beginning, and something from my own ideas, gave birth to this idea:

You are a single governor-tyrant, who has this city in his iron grip, and all the stuff you talk about, but, being a tyrant, you always have to fight rebels... or, gradually, in the course of the game, the rebellion appears and grows. You are represented by a game world person who lives and rules from his/her high castle, and the game ends, when the rebels reach you, and overthrow the tyranny, via the means of assassination.

I don't know if I understood the problem of nation-in-a-building correctly, but for some reason I had the visions of Bladerunner as a solution. I mean, any single building being explicitly multi-cultural and buzzing - one corner is occupied by Chinese restaurants, another side is all European, everywhere bright signboards, maybe in a different corner a war is going on...

[Shades]

My advice would be to start this as two connected projects with different goals.

The first is a fairly standard 3d engine with a physics component capable of accurately modelling both the large structures of the towers but also the stresses involved with traffic moving around it. This isn't a small task but there is lots of existing code to work from.

The second is an overly complex social AI engine. You have less code to work with here but can probably find bits and pieces. The system, from my point of view, would mostly rely on emergent behaviour from each of the people working in the city. However you might need some grouping to avoid requiring a supercomputer to run it on. I would start with a model of the flow of workers over the course of a day and see if you can build needs into the model to give it some variety.

Most important however is to just get something working so you can play with it and improve. There is obviously a lot more to it to get everything you want but those two leap out as the core problems to solve opportunities to take advantage of.

[Blacken]

The first is a fairly standard 3d engine with a physics component capable of accurately modelling both the large structures of the towers but also the stresses involved with traffic moving around it. This isn't a small task but there is lots of existing code to work from.

I'm not sure I'd even bother with a physics engine--or anything stepwise, for that matter. Abstract away individual behavior, or you'll be bringing computers to their knees from a ton of calculations even before you get to the scale a "dystopian city" would need. Dwarf Fortress can get away with a very low scale because, hell, who knows what a dwarf settlement looks like? Little harder for something closer to normative experience.

[Shades]

Abstract away individual behavior, or you'll be bringing computers to their knees from a ton of calculations even before you get to the scale a "dystopian city" would need.

A decent physics engine should be able to do that, but yes it's important not to model every vehicle and person for stress analysis.

[qwertyuiopas]

Maybe have the level of detail increase as the number of blocks visible decreases, and after a certain point you can see people. Rather than simulate them entirely, have their day simulated as a general list of things to do, toss in some variance, and have it alter itself periodically. When you go to watch them, it could check what they were doing and generate a plausible location for them. It should also add an invisible goal at the current time that records where they were so that it can consistantly place them where the player feels they should be. It wouldn't even need to do pathfinding until the player can see them, and the invisible node ensures that the path they take passes through where they were seen.

Such a list could also be recorded by security whenever a destination is near enough to a camera, so you could read the logs of citizens that appear to be rebels as decided by your underlings (Or you, if you feel like wading through tens of thoudands of
mostly normal files to find three deviants, most of them legitimately just trying something new that day.

[IronyOwl]

I'm thinking the best way to handle individuals would be to abstract most of them but create a few visible examples. So if your populace is 80% disheveled factory workers, 3% rebels, 10% wealthy elites, and 7% poorly disguised mutants, the game would occasionally spawn a handful of people, with those odds of being a given type. Said people would wander around a bit before- unless you did something to specifically keep track of them- merging back into the populace. So in your city of fifty thousand people, you might have fifty or a hundred of them wandering around at a given time, which gives you the ability to zoom in and see what your people are up to, without too much of a load on the computer. If "people view" were a separate perspective/mode, you could even generate them on demand when you entered that level of zoom and remove them when you left it.

This also generates the possibility for sampling errors, which I suspect nobody other than me cares about. You might examine two or three people and find they all like you, leading you to believe you're doing okay, when in fact you've just hit the 20% that don't want you gone three times in a row.

[Supermikhail]

An AI for population statistics like... in Civilization? But I'm not sure. Of course, even if I'm correct, a much more sophisticated AI.

[jplur]

I read a large part of this, really like the idea.

I think one of the basic concepts, the structure of the city, is very similar to an idea I've been working on for some time; http://www.bay12games.com/forum/index.php?topic=48617.0

Basically, legos.  Where in the basic blocks have some innate functionalities, that end up being combined and creating incredibly complex overall systems.  The part where you talk about cutting oxygen off to parts of the city really caught my attention.  It would be very hard to program that in directly, but if your building structures have some innate characteristics like volume and connection points, then all you have to do is extend a base object to act like a cut off valve.

Anyway good luck!