My Urban Design and the Creation of Videogame Cities talk from Develop: Brighton 2017 now handily available in the form of a slightly edited and abridged article. Huzzah! The video complete with Q&A --for those who attended the 2017 conference-- is available here. More information on the subject and on my work on virtual urbanism can be found over at my game-cities.com site.
talk article, which I've tried to fill with as much useful information as possible, is an attempt to provide game designers and world builders with a spectrum of tools, ideas, and ways of thinking that will improve their game cities and interactive urban environments.
An attempt to help things move even further from the simplistic banality of ancient Police Quest-esque cities.
Urbanism, you see, both actual and imaginary, is incredibly complex, but I do hope that enough ground and most of the basics will get covered.
Unlike the majority of artists who get to do most of the interpretation of cities for their audiences themselves, in games we can only do part of this interpretation. The rest is up to the players; it is they who get to directly experience our virtual environments, and assign value to them.
Players get to decide whether our old fashioned, working class fish shops are brutal, backwards, noble or picturesque. They get to frame them, play with them, hide behind their benches, or totally ignore them while shooting things. We can only provide them with something lovely and convincing. Something believable. Something that even when barely registering will feel real.
To achieve this, what we actually have to do is provide players with places that --first and foremost -- make sense. And this is actually the most important rule to designing a good game city:
Game cities must always make sense.
Realism, you see, leads to believability, which in turn leads to immersion. A feeling of presence. Of being there in space.
Also, as I kept telling urban planning students for years, and am bound to keep on telling world builders, please, do not think like architects. That's the wrong scale when approaching settlements. Think like urban planners. Even better think like geographers and planners.
And, while you are at it, do let your city tell its story. Just like Los Angeles starred in the classic noir movies, your game's city can and should influence both your game's story and how it plays. Besides, researching your setting and trying to get to grips with its urbanism is bound to give you a lot of fresh ideas and solutions.
A city, you see, is not merely the sum of its walls, roads, buildings, and infrastructure. It is much more than its facades and landmarks. It is even much more than its roads and land use plans.
A city is the humans living in and using it. The animals too. A city is everything that happens within it. Everything that shapes and is shaped by it. A city is a huge, complex, dynamic stage for human life. For activity, and drama.
A city is its people, and how they dress. It is its economy, its rumours, hills, monuments, sounds, sky, traffic, buildings, power structures, the ways people get access to drinking water, religion, and a myriad other things, but above all, a city is its functions.
Now, I will of course be returning to the crucial subjects of urban functions, but first I'd rather briefly discuss some of gaming's most intriguing and inspiring cities.
The first one has to be Bioshock's Rapture. Not only is it brilliantly imaginative, utterly exotic, and unique in its underwater beauty, but I also consider Rapture an iconic place realistically organized around its guiding ideology. It is, admittedly, neither a complete nor a perfect design --no artificial city can ever be-- and its illusion of coherent urbanism is thoroughly helped by a plot that does away with the need for a proper city life, but it really works. And the Art Deco architectural choices are truly apt too.
On the other hand, Bioshock Infinite wasn't as successful, mostly due to unrealistic crowds, and its deeply flawed roller-coaster inspired transportation option. The latter in particular was what broke immersion for me, though I was happy to see several reviewers mention it too.
Interestingly quite a few reviews also picked up the importance of Dunwall's economy in Dishonored.
Like Bioshock Infinite, Dishonored doesn't take place in an empty city. Unlike Infinite, Dishonored's Dunwall is a living, breathing place. It's got some wonderful architecture, and an all influencing and original whale-based economy around which much of the city, and its everyday life have been constructed. Also unlike Infinite, Dishonored wisely sets its battles in places mostly populated by guards, safely outside the uncanny valley.
What both games do really well, is providing us with a sense of stark, spatially organized class divisions, and exciting (if hateful) civic politics.
Another clever choice is that the huge scale of Dunwall is mostly implied, and not modeled in detail in its complete and very expensive entirety. It is beautifully shown off from afar while travelling in boats or while looking from the shores of the river.
Another one of my favourite game cities would have to be Kane and Lynch 2's Shangai. A place that felt incredibly lived-in, dynamic, packed with detail, a strong sense of locality, and thus real. Palpable almost. Much was shown and even more was implied via carefully crafted routes through backstreets and informal markets, and by utilizing skyscrapers that could be glimpsed in the backgrounds, or by emphatically showing that the city was a place under heavy construction.
More ambitious and an actual open world, Assassin's Creed's London is a truly impressive, 100% walkable game city. It only models part of the of the 19th century metropolis of course, but manages to cram in all major characteristic districts types, as well as many recognizable landmarks. It is a fine and very expensive abstraction of a city that existed, and one that is vastly helped by incorporating a lovely to look at and relatively cheap assets-wise river. Assasin's Creed doesn't forget to include upper, middle, and working class areas, historically accurate industrial spaces located both inside and separately from poor residential quarters, pastoral suburbs and parks, and even utilities such as railways, hospitals etc.
I would have probably gone for slightly narrower roads, and a greater mixture of land uses myself, but, admittedly, this is quite an amazing city. Oh, I would also fog up the sky a bit more too! Like this:
The lovely if extremely compact Prague from Deus Ex: Mankind Divided is also worth a mention, as is CD Project's forthcoming and apparently terribly ambitious sci-fi city of Cyberpunk 2077.
But, what we really have to spend some time on is Half-Life 2 and City 17. A hugely successful urban environment to play in that, believe it or not, is already 13 years old.
And yet it remains one of the most believable, coherent, economically designed, and memorable urban environments ever crafted. The imposing alien tower, the Combine Citadel --a major, central landmark-- instantly sets the tone and always keeps players oriented. It makes the well designed city easier to navigate in combination with minor landmarks (such as the train station), and acts as a defining juxtaposition of Combine and Human architecture and space organization logic.
Said landmark also works as the latest addition to the historical layers of the city. City 17 has different tech-levels co-existing, buildings of various ages, regimes, and styles, decaying houses, alien constructs, and, importantly, everything placed in it feels as if it has been placed there to serve some personal or collective purpose. Either for residents past and present, or the occupiers. The fact that the urban environments have been based on obviously extensively researched cities such as Sofia and Belgrade obviously helped a ton too.
What's more, the urban layouts and spatial organizations selected for the place, serve well as battlegrounds, what with City 17's courtyards, canals, and streets of varied sizes. Even the sky and climate feel right and not just something randomly chosen. As for the plot, it supports not having too many 2004 tech taxing people around, while blocked off areas due to security reasons, and long canals, help imply a civic size that isn't fully modeled.
In a nutshell, here are the major successes of City 17:
- Major Landmark: Citadel
- Obvious Historical Layering
- Extensive Research & Use of Actual Urban Configurations
- Urban Layouts that Support Level Design
- Sky, Climate Were Obvious Considerations
- Believably Restricted Civic Life
Sadly though not all game cities are this successful... I can think of MMO capitals that are supposedly sprawling metropolises, and are yet nothing more than a few dozen mostly disjointed buildings populated by a mere 100 or so NPCs. I can also think of dozens of medieval-esque RPG fantasy towns where there are two inns for each house, but, well, I'd rather quickly mention Citadel City from the otherwise brilliant Mass Effect 2 instead.
I consider it a missed opportunity. Despite being wildly imaginative, and stunning on the concept art level, it simply got its scale and complexity all wrong. Felt far too much like theatrical scenery built around Shepard. Like hints of paper-thin places only constructed to let her do a dialog bit and then vanish.
But I don not like being too negative. Instead of going on and mentioning all the cities that were sort of off, I have compiled a list of urbanistic elements usually missing, or commonly done wrong.
A (far from complete) list of problems and omissions:
* First of all we have vehicular and human traffic which is sometimes completely forgotten, but more usually inadequately approached. Traffic is commonly very scarce or overtly simplified, even if it is an essential and occasionally imposing part of the urban imagery. Then again, most game cities even lack ranked road networks. That is, a hierarchy of roads from the local to the interurban.
Roads in particular, those important elements, are too often done incredibly wrong. I know that Syndicate is an old game, but you'd be surprised as to how many recent, even forthcoming releases have roads this angular. Now, I do understand why this happens, especially for smaller indie studios using tile tools, but just rounding the corners of the pavements will instantly make things look better. Then you can start thinking about modeling your crowds and vehicles.
* The lack of in-city topography or of anything interesting from a natural geography perspective also hurts game cities. Topography doesn't only help with orientation. It has also historically shaped how cities were build. Besides, hills can be incredibly characterful landmarks, differences in elevation can be associated with class differences, and Rio de Janeiro just wouldn't look this amazing if it were flat.
* Believable city life. Or any city life for that matter, though usually there are some animatronic style interactions that keep on repeating themselves, as well as a few people randomly moving around. Or just standing there like some sort of prop. Even worse, eternally re-enacting the same little scene at the exact same place.
How could this be fixed? Well, first of all, do flag your urban life vignettes! Seeing them once is much better than seeing them a dozen times. Maybe, have a pool of interactions and just make certain they are all seen by, say, forcing them on the players paths. Or have simple, sensible scripts governing the movements and actions of people around cities. Have workers commuting towards the factories in the morning and then back. Have guards patrolling the walls, and cooks walking around the markets looking at ingredients.
Also, keep in mind that not everything has to be extraordinary. Mundane images of city life are essential. Just look at those screencaps from Ghost In The Shell -- the mundane can be beautiful.
* Similarly, the scale of things is often all wrong. Size, functions, and variety are correlated you see. Though it would make sense in a village, being in a supposed metropolis and telling the players to go to the one book shop just doesn't feel right. And all it needs to get this fixed are a few words. Do not tell them to go to the butcher, but to John's -- the best place to grab meat in all the city. Or direct them to the only open place at this time at night in the butcher's district. It's OK for Thimbleweed Park to have a single factory, but it wouldn't be OK for GTA Vice City to sport a single bar.
Other commonly forgotten elements include:
* the evidence of functioning infrastructure
* public spaces
* street furniture including benches, signs, dumpsters, lights, etc
* and class and land use divisions in space, and realistically mapped land uses
* Dynamism and history. Particularly irritating is the incredibly common lack of buildings under construction. It would be nice see a palace or a cathedral still under construction -- these things took centuries to complete. Also, cranes and building sites are too common a sight in each and every contemporary city, and can help provide a virtual place with a sense of history in the making.
And where are the wells or aqueducts in our medieval fantasy cities? They were super important social spaces back in the Middle Ages, and we all still do need water, but I do believe that even the amazing urban centers of Witcher 3 missed those. And if they do not exist, shouldn't we have substitutes? Magical water catchers perhaps? Or merchants monopolizing the water trade? Regardless of the way it is done, being able to sustain the life of its inhabitants is a core urban function.
Then again, Thimbleweed Park, which I adored, actually did away with the whole urban function of residence. So, yes, games do miss the concept of urban functions --of the core things that cities are meant to do, allow or provide with-- mainly because they don't have them in mind, and this is the source of many of their problems.
This of course leads us to the very basics of imaginary city building and geography, which have a lot in common with the very basics of urban planning and real life city geography. For, as already mentioned, things in virtual space have to make sense even if conforming to the most alien of internal logics.
The very basics of imaginary (and actual) city building:
* First of all, we have the crucial subject of urban functions. The functions that are the reasons cities exist, the functions that differentiate a village from a city, the absolute core of urbanism and city geography. Cities are built to facilitate and support said functions and in turn evolve as those functions themselves evolve.
To give you an example... Contemporary urban functions, core urban functions, would be commerce, production, consumption, human reproduction, transportation, culture, ideology, the circulation of capital, and housing. Reproduction itself would include further functions that an urban area has to provide such as health services, shelter, entertainment, schools, and access to food and water.
Functions do of course change, and dialectically evolve throughout history. Religion for example would be at the absolute center of a holy city, shelter via walls would be important during the medieval era, and a modern metropolis, an essentially new urban formation type, has variety as an essential new function. Obviously it incorporates many more functions than a village. In a nutshell, functions are what a city does.
* To start building an imaginary city though we have to first ask (and answer) three questions: where, when, and how big? Answering these defines a set of central functions, and also helps us approach crucial elements, like topography, history, and economy. You don't obviously have to answer the questions in this order, but keep in mind that what you choose to ask and answer first, is probably what you find to be the most important thing in your city. For example: let's do something in an ancient town or let's do a huge sci-fi metropolis.
Obviously if climate and topography are to define your settlement, then you'll have to start with where. And this where can and will influence local attitudes towards the sea, nature, and even the building materials used.
* On to the production and dialectics of space. Always keep in mind that space produces and is produced. Space for example is produced when we plan and build a road from the city to the airport, and then this produced space in turn produces its own space by attracting malls and gas stations around it.
* Everything in every city, town, or settlement of any size has been put in its place for a purpose. That is, everything has in a way been planned. Even if not by a central authority, or in recent years, everything that exists has been put there by someone for a purpose. That is why we expect spatial elements to have been designed, and why certain spaces in games work even better when their function is instantly recognizable and understood. When buildings, public spaces, or street furniture lack a sense of function, they fail to convey a meaningful sense of civic illusion, and believability suffers.
* The biggest and most prominent buildings can and should emphasize the dominant urban ideology. Are they cathedrals? Corporate HQs? Military installations? The Parliament? The halls of the People's Assembly? We have to show them off.
* Cities are always works in progress, always dynamic. Created by hugely complicated histories, they are the battlegrounds of classes and social groups, and the places of countless human interactions . This dynamism has to be reflected on the built environment too. Mix old and new buildings, have your majestic central temple be under construction, remember the majority of buildings in New York don't get to be 25 years old, that factories open, go bankrupt, and are reused as art spaces, and that even a square might radically change itself for a carnival or the Sunday market.
Change and dynamism are such important elements of city life that even board games like Arkham Horror tried to incorporate them. Arkham has its shops closing down, and people moving out as the game progresses and the horror escalates.
* All cities are internally differentiated, with sub-areas either strictly or porously defined. The main divisions are those according to class (rich and poor areas for example), prevalent or exclusive land use (entertainment, residence or industry), topography, proximity to the center or centers, and any combinations thereof. We can for example have industry in residential areas, marketplaces inside posh districts, the poor by the harbour, and guild halls next to Papal residencies.
* All cities have history. Even if this is brief, and especially if we've set our game in something as old as Cairo, this history has to be researched, possibly made up, and definitely layered. If you visit Geneva's Cathedral you'll have a fine chance of observing the physical layers of the city. From the primitive, to the Roman, to the Medieval, to the Renaissance one.
* Each city is unique, with its own distinct feel and internal logic.
Keeping those absolute basics in mind, let's move on to some spot rules.
* First. Always try and have a fully functioning city in mind. If possible sketch it out, and always allow your thinking to move from the general city to specifics to details.
* Embrace huge, dramatic differences and changes within your city.
* When informed by history make sure your historic choices match. Be influenced by what actual cities were influenced by. Keep things consistent.
* Consider the wider geography and hinterland of your city.
* Never forget to also think about urban infrastructure. From today's complex telecom networks, to medieval wells, to the Roman roads, infrastructure is the skeleton of cities.
* Another thing you should be considering is which parts of your city are public and which private. What are their relations and how strong are the borders. In the UK for example you have private parks, which would be pretty unthinkable in Greece, where --in many villages-- house yards are essentially used as public space.
* Street activity is very important. If possible show people doing everyday stuff: butchering, praying, laughing, playing, looking at the sky, carrying things, going places. The power that little vignettes have in evoking city life cannot be underestimated.
* The importance of graffiti, posters, signs, street art is also significant. Not only do these tell stories, but they've been around for quite a while, and can be safely considered as almost integral to the image of the city throughout history. From ancient graffiti, and 19th century posters, to terrifying street signs, and the hobo sign language, urban streets & walls should be intriguing.
* Little uniquely local touches can work wonders in conveying the particular sense of place an urban center has. It could be everything from thousands of little god statuettes wrapped in tinfoil that are everywhere to be seen, the common use of gestures when passing in front of a temple, a tendency for people to always wear something red, an overabundance of beggars, or even small open street restaurants or cooks.
Locality can also be further emphasized by the use of distinct materials, the climate, and one (or more) particular architectural styles. It also becomes apparent in idioms, customs, and myths.
* A city or town name is also important. City 17, Rapture, Liberty City, Raccoon City all set a tone and evoke initial thoughts, but do allow me to digress and ask you not to overuse the cyberpunk trope of Neo New York, Neo Tokyo, and New London. It's been done do death, and actual place naming never worked like that. Medieval York did not actually grow up to become New York.
* Do not forget that the vast majority of cities have important historic remnants and monuments from previous periods. Not all buildings are new, and not all buildings are contemporary. Cyberpunk settings do tend to miss this particular detail too.
On to some urban planning and design guidelines now, as, generally speaking, good urban design and planning practices can be applied to virtual cities. Frankly, I believe they should be.
Interestingly though, and especially when it comes to virtual and imaginary cities, the opposites of good city planning and design practices can also be applied; provided we are looking to evoke a specific feeling. An imposing city ignoring the human scale, utterly confusing labyrinths of narrow dark alleyways, and unnerving symmetries would never be considered good urbanism, but could be used to create evocative, unique cities for specific atmospheric needs.
Inverting the rules for comfortable streets, and safe and inviting public spaces, can be employed when we are, to give you an example, designing for horror. Dreamlike environments could employ spaces that would not work in reality, or completely ignore core functions.
As a rule of thumb I'd say you should generally let your city inform and enrich your level and game design, and in turn allow it to be shaped itself by the needs of gameplay.
On to some specifics:
* First of all, aesthetically speaking at least, a nice balance of order and chaos works wonderfully for cities, as both the sense of a planner's intelligence and forethought, and the elements of surprise make for good looking, intriguing places. Mixing things up usually works.
* It's all about the exciting dialectics of mystery versus orientation. The interplay of open spaces and enclosure. That sense we get when entering the medieval centre of a modern European city. Sometimes getting lost can be exhilarating.
* You should also go ahead and do some rudimentary but sensible land use and city planning. You could for example have the polluting, space-hungry factories in the suburbs, decide on the density gradient of your city (higher in the walkable center), and work all the way down to the more architectural level of city design.
There's another whole set of separate rules for city design, but let me just provide you with a couple examples. When designing a square that's supposed to be cozy keep it to up to 30-35 meters in width --that's the distance people are able to recognize each other from afar. And your sidewalks should allow for at least two citizens to walk abreast.
* Most of the times you will want to guide players through your city. There are of course many game and UI design tricks to help you achieve this: mini maps, in game GPS guides, big arrows, subtly lighting the right path, having NPCs provide with directions, or even giving players a character to follow.
Then there are also the --definitely more subtle than a big red arrow-- tricks urbanists, and city planners use to help actual humans navigate actual cities. From the very obvious and widely used street signs, road names, and numbered buildings, to ranked road networks, public "you are here maps", and networks of pedestrian only streets. Helping people find their ways is a very important real life concern.
In addition to these, legendary city planner Kevin Lynch, came up with certain tools we can use to make urban space legible and thus easier to navigate. Lynch decided there are five crucial elements people see and experience when moving around cities -- they are the same elements gamers can be provided with in order to create their own mental maps:
- Paths. Routes along which people move throughout the city. Like roads, highways, railways.
- Edges. Boundaries and breaks in continuity. Again these could be railways (if you are a pedestrian for example), city walls, waterfronts, etc.
- Districts. Areas characterized by common traits; internally coherent areas. This could be aesthetically, architecturally, or by activity. Say, the printers area or a red light district. Maybe even the red or multi-coloured district like Boca in Buenos Aires. Or the Old Town.
- Nodes. Important, strategic even focus points that help orientation such as squares and junctions.
- Landmarks. Landmarks, can be major or minor (even personal, player-centric places), and must always be highlighted.
Add to those the elements an old professor of mine suggested , scale and grade/slope, and my humble addition of sound, and you have all the elements an easy to navigate city needs. Mind you, the soundscape can really help show off such things as how close to the metalworks you are or how busy a particular area is.
* Finally, there's the matter of implying size and scale in cost-effective ways. Of conveying scale and complexity that really isn't there. And, truly, there is always the need to imply more for a game city. Even the huge virtual worlds of Grand Theft Auto are tiny when compared to real life urban formations.
First tip! Know that even a simple picture can imply scale. Look at the set of buildings above. Look at the obvious residential density. The signs of life. The amount of detail, and the number of windows. Could this be found anywhere but in a big city?
Or how about this scene from the Mouth of Madness movie? It feels instantly as if we are in a town, yet there are only 7 buildings we can see. Their placement and our almost instinctive knowledge of how settlements tend to look does the director's work.
Then again, and that's another tip, when you are looking to create long avenues and boulevards, you should probably avoid straight roads especially if they allow the horizon to be seen. Compare these two slides of a model I quickly did in Sketchup:
You can also block off areas as Grim Fandango brilliantly did.
Or just like Grim Fandango, and a ton of other games, have taller buildings, landmarks, or major infrastructure been shown in the background.
Better yet, if possible add more inaccessible buildings to your urban areas. And, yes, these do make sense. We do not after all expect to enter every building we see, and most doors are locked. Even when in a panicky situation just having to knock on 5 different doors to find the one that budges and allows you to hide can be incredibly tense, not to mention realistic.
Also do consider the solution of procedural buildings. Maybe even whole procedural sub-areas if they are to be shown from afar, or edited by hand at a later date.
A very brief city how-to:
So, to put everything together in a way, let me quickly discuss how could someone go about crafting a city for their game. It could of course be for any setting, be it fantasy, contemporary, sci-fi, or covering a specific historic era.
1. The first option would be to pick an existing city, and try to abstract it to a manageable level. A kind of cartographic generalization, the method, that is, that cartographers use to decide what goes into a map could be applied. It's a method for deriving a smaller-scale map from a larger-scale one. That is for example showing a city from a sum of buildings, to a sum of city blocks, to an outline, to a dot. You'll have to pick and choose what to show, what to model, then note down the main urban functions, and research and codify what makes the place unique. Assasin's Creed's London is a fine example, as admittedly is the original Gabriel Knight.
2. Of course you can always heavily modify and combine existing cities and urban configurations, following the example set by Half Life 2.
Famously, even Terry Pratchett's Ankh-Morpork was based on Tallinn, Prague, and had bits of 18th century London, 19th century Seattle, and early 20th century NY thrown in.
3. But, let's say you decide to create something from scratch. Well, start by asking where, when, and how big. Then make a list of your core functions in as detailed a way as possible. Start crafting and layering the place's history. Make sure you come up with at least a rough map of the whole city and its environs, and work from the general to the specific.
To tackle the task ahead you'll probably need to also do your research, and read quite a bit. The urban planner and geographer was always considered a dilettante and dabbler in all trades, and we could argue that the game designer is the ultimate dilettante. So read. Know things. Use a variety of tools.
My tools include sketchbooks and notepads, digital tablets, LEGO bricks, QGis, SketchUp, Autocad, and anything else from tile utilities to Photoshop. They have to change with each project and each city.
As for the books, well, to finally bring this article to an end, here's my suggested bibliography for you:
~ ~ ~ ~ ~ ~ ~ ~