Comment by andys627 | Hacker News Reader

Comment by andys627 | Hacker News Reader

It's crazy how much SimCity cities look like the cities we've built for the last 75 years. Our cities really can be simplified to a game "fun" for someone to dive in right away. My thesis is that cities are oversimplified and stratified into easily separable blocks because they're easier to "industrialize" that way. This is causing crazy transportation requirements and isolation. If we mixed uses up a bit we could save on transportation overhead and be exposed to more things. The driver of this SimCity pattern of development is undoubtedly the single occupancy automobile. Sprawl works best for the car and the car works best for sprawl. Denser, mixed up uses work best with walking, biking, and transit. Conspiracy theorist in me says the auto industry influenced city planning from the ground up because they could profit off of it. Who profits when you walk more? Well, everyone... healthier bodies, healthier minds, more efficient transportation, etc. etc. etc.

My biggest problem with both SimCity and City:Skylines (its spiritual successor) is that neither allows mixed-use zoning, forcing you to rely heavily on either cars or dense (expensive) transit networks. It's like they can't break the mold of the US suburban mindset.

In the latter, while you can't build a building zoned for multiple uses, you can certainly very much blend zoning very thoroughly, alternating commercial and residential buildings as you go down the street. Of course, the simulation likely isn't sophisticated enough to make full use of that (people will travel across the entire map to go shopping). I can also understand from a gameplay perspective why this is the case.

Maybe the shops in that part of town have better stuff.

There's the issue you mention, but IIRC there are also some big stat hits (happiness, noise, land value) from that sort of pattern. The game really pushes towards districts of distinct zoning as an optimized solution.

It's very much because of gameplay issues. Theoretically the computer can simulate whatever model you can dream up, but the much harder problem is the user understanding, visualizing and editing that model with an easy interface that feels fun like playing a game, not tedious like doing your taxes in Excel or rigging skeletons in 3D Studio Max.

Think of the number of extra user interface panels and gizmos and views and pop-ups and menus that users would have to wrangle in order to build, edit, query, visualize, and maintain 3D multiple story mixed use buildings in a city simulation, instead of having a simple 1:1 correspondence between space on the 2D map and zone type.

The closest thing SimCity comes to mixed use zoning is multi-layered Arcologies, and the interface to those is pretty complicated and unwieldy. Imagine if every building and skyscraper in the city were that complex!

Sim Tower supports mixed use zoning (with different discrete rooms in the same building), but just within one high rise tower, not a whole city. It's as much a game about pedestrian traffic, elevators, queuing and congestion, as an economic simulation.

The Sims is whole a game unto itself at a different level of abstraction, that lets you build multi-story residential and commercial buildings with walls, doors, windows, furniture, etc. But it doesn't simulate "urban dynamics" between zones, just "personal dynamics" between the people and the architecture and contents of the buildings. The Sims 3 was scaled up so you could seamlessly walk around small towns, but The Sims 4 went back to focusing on houses or apartments in individual lots or flats, instead of expanding to simulating entire cities.

I'd be interested in a city simulation game that is focused on realism and depth with no concern for having a modern UI or being pretty. Think Aurora 4x, Dwarf Fortress, or Nethack. "Doing your taxes in Excel" is, surprisingly, a real game genre.

It would be a very unpopular game, but there are people who would love it.

I've long had a dream of taking the source code for the original SimCity and changing it more to my liking. The newer version was nice for streets with curves, but inferior in almost every other way compared to the original. But I doubt I'll ever find the time to actually take a crack at it, so it will likely remain just a dream.

Take a look at Lincity, an open-source clone. https://en.wikipedia.org/wiki/Lincity

Thank you!

The original source is now available under the name Micropolis https://en.m.wikipedia.org/wiki/SimCity_(1989_video_game)#Mi...

Github repo:

https://github.com/SimHacker/micropolis

Couldn't we solve this with a percentage type for zoning, or adding a "mixed commerical/residential" zone? UI wise that shouldn't be any more complicated.

Or you can change the zoning into japanese zoning, where the UI can stay simple as you escalate the permission level of a zone. https://code.uberinternal.com/T2597239

Seems like you've accidentally posted the wrong link.

This. If Jane Jacobs could play Skylines today, she'd probably be on the phone to the devs asking them "why can't I have ground floor retail under my residential?"

So, it faithfully represents urban planning in the US then?

Exactly!

People have certainly built Cities: Skylines cities with mixed-use developments, quaint mountain towns, little enclaves built around a park or monument, car-less cities, pedestrian-only cities, modular cities connected by mass transit, etc.

I think the problem - in SimCity, Cities: Skylines, other video games, and real life - is that if you're building for aesthetics, there are many right answers, but if you're building for efficiency, there's only one right answer. The reason so many mid-size American cities have converged upon 1/4 acre lots with a road hierarchy, 3-lane boulevards, commercial strip malls on the boulevards, office parks, and freeways to link suburbs is that that's the most efficient way to satisfy the constraints of 2-3 child nuclear families who want privacy and a patch of greenery that's all their own and have varied destinations that each support too little traffic to be worth building mass transit to. The reason many megacities (I'm thinking particularly of NYC + Asian megacities like Beijing, Taipei, Hong Kong, Shanghai, etc.) have converged upon high-rise condominiums with street-level retail taking up whole city blocks with abundant bus & mass transit routes is because at a certain population density the suburban model ceases to work, and this becomes the only feasible way to move people to work & shopping. Interestingly, both models show up in Cities: Skylines, and the transition point between them is usually a painful inflection point in your city's development.

This might be the one reason that makes the F2P Sim City worth mentioning, in which where I’ve found it much more challenging to achieve aesthetics while growing at a reasonable pace. Having grown up with the original Sim City series, I naturally have a dislike for the F2P version, but in the context of your post I appreciated the irony of F2P enforcing a certain realism as a side effect of, well, a realistic economic factor. Ok, now somebody take this and work in a proposal for a blockchain hybrid of the two.

Its not just a space optimization but a locality optimization. Segregated neighborhoods from 50 years ago are still likely to be largely minority and gated communities adjacent still largely white. A tremendous amount of the demand that goes into urban sprawl is the desire to live distant from those deposed by the society the homeowners often perpetuate and instigate.

The quarter acre of manicured grass, the per-working adult car, the maximum occupancy limits, the minimum square footage requirements, zoning in general all work to keep the "riff raff" out economically.

Nothing about sprawl housing is efficient. Its almost maximally inefficient - you get the worst of everything, the traffic and congestion problems of density without any of the actual density because the people centralize on an urban core and surround it in extraordinarily wide expanses of cloned stick houses. But that inefficiency is baked in by design to push out those who cannot afford to participate in such inefficiency. You don't need gates on your community if the area is unlivable for the working poor for miles around.

I don't doubt that's one consideration, but my point is that people's preferences - whether they be racism, greenery, privacy, autonomy, or whatever - function as constraints that the market then optimizes around. And given those constraints, the suburban tract neighborhoods with a road hierarchy and wide boulevards lined with strip malls optimize for efficiency.

I grew up in suburban New England. That's inefficiency - triangular acre lots, empty space with no road access, single-lane roads, three lefts that get you back to your starting location, driving to the next town (or New Hampshire, often) for the nearest commercial zone, the "Massachusetts left" because there's no protected left-turn lane at most intersections, major roads that terminate at a stop sign on a freeway where people routinely drive 85 mph. The suburban sprawl of Silicon Valley was a breath of fresh air by comparison - the lights on the 3-lane boulevards here are at least timed in rational ways with protected turn lanes.

Hell, they have traffic lights here - I grew up 2 houses down from the intersection of 3 state highways at a stop sign, and we fought for about a decade, fruitlessly, to get a traffic light and crosswalk there. It backs up for literally a mile in each direction during rush hour, and accidents are common.

I believe your opinion is based on very local observations. Cities in Europe are wildly different, and in many aspects, they are very successful. Even in the US, cities have changed a lot over time, as shown by the gentrification of inner cities.

This one efficient model you are talking about is merely the solution to one set of tradeoff decisions in your area, with other tradeoffs being chosen in other places.

There's no conspiracy needed, the push towards stealing roads from people and horses and taking them for cars is pretty well documented. Once we decided it was ok to tell kids they couldn't play in the street, our cities were lost.

https://www.citylab.com/transportation/2012/04/invention-jay...

I recall that one of the "optimal" designs for one version of SimCity (2000?) was to place street intersections--without connecting streets--at regular intervals, and build a connected subway station next to each.

Certain features were hard-coded specifically for surface street intersections. So you couldn't simulate a Kowloon Walled City level of density, unless you count the arcologies. The car assumption is as built in to the game as it is to real cities.

Power was supplied exclusively by perpetual-motion hydropower dams, enabled by the terraforming tools: flatten, raise land, add water, build dams on the falling-water squares.

Anyway, the point is that if city planners these days are simplifying their task anywhere near SimCity levels of abstraction, the ensuing disaster won't be one that can be selected from a pull-down menu.

The designer behind sim city (in the early 2010's) discusses how they just ignored parking because it makes cities ugly. https://humantransit.org/2013/05/how-sim-city-greenwashes-pa...

You are not the only one that believes that the auto industry worked to change urban transportation.

https://en.wikipedia.org/wiki/General_Motors_streetcar_consp...

The book you’re looking for is “Seeing Like A State”

Judge Doom[0] explains it.

[0]https://www.youtube.com/watch?v=z75lu0wKbX8

Someone else posted the "General Motors streetcar conspiracy" already.

Reply on news.ycombinator.com