Manifesto

[Deed_Poll 32]

I appreciate that the mods might think this is in the wrong forum section. However, I would ask you to hear out my case. Primarily I think that delivering on the admittedly high expectations people had of SC2013 is a good starting point for such a discussion. Also as a product designer myself, much of the research into developing a new product is the identification of weaknesses in existing products to improve upon them. This being said, I would encourage posters to have a constructive and positive attitude in this thread focusing on what is achievable, at the expense of the many valid criticisms that we are all now familiar with.

I've read that a few have tried and failed to get a kickstarter city simulator off the ground. I lack the programming skill and experience in the industry to make my own attempt at this, but I would like to set out what I'd be looking to see in a large scale, 'blue skies' reinvention of the genre - and I'd be really interested to hear ideas from you all as well.

I think competition in the marketplace is essential to guarantee progress in quality. Games like Gran Turismo and Pro Evolution Soccer stagnated until Forza Motorsport and FIFA 10 respectively took the fight back to these franchises and reinvigorated progress in these genres. SC2013 is the best current example of a monopoly gone stale.

I'll start with my vision for such a game.

As I work with a lot of rendering software in my job as a designer, I would approach the visuals in a different way.

Instead of a fully interactive 3D view I would have the player choose 2 or 3 angles to switch between.

Each New building would be generated as a 3D model and then photorealistically rendered from each of these angles, at different times of day and weather conditions, so that it could be implemented in the game as a 2D object. A very simplified box model of this building would be used to update shadows on surrounding buildings as they grow and vice versa.

Buildings themselves would be algorithmically generated in accordance with:

- style of surrounding buildings

- plot shape / size

- desirability / wealth level of prospective inhabitants

- overall city technology level

- proximity of parks, waterways etc.

This would ensure that each building would be unique and communicate to the player the relative affluence and desirability of certain areas.

An algorithmic model would also allow buildings to gradually change; to grow, shrink and change textures naturally. This would be another important element in the uniqueness of each building as anyone who knows fractal geometry and chaos theory knows that iteration is key - having systems both as inputs and outputs will guarantee complex and interesting results even from simple foundations.

With this system you would be able to see the history of each building, how it has grown, shrunk, become enriched or impoverished as time goes on. Every 5 years the simulation would take a compressed snapshot of the building and pop it in the archive.

Additionally, parametric modelling like this would suit repurposing of buildings - from industrial to residential for instance. Chimneys would become lift shafts, lumber yards would become carparks or communal gardens. User input could inform the simulation, weeding out the genetic material of buildings you choose to demolish, and conversely favouring elements of those you make historic for future building generation.

In addition to the 2 or 3 camera angles you choose to view the game (which would of course be 'isometric', or at least would have infinite 'lens length', i.e. no perspective) you would also be able to make 'districts' in the city where a point is placed in the centre

and allows a panoramic, perspective view from the centre of that point. This would behave like a city view on Civilisation, where you would be able to tweak that region's spending on health, education, police and other services. Of course, each building would have to be rendered separately for this view as well - again, just to make up a 2d image, just one which could be looked around like an old point and click like Zork or Myst.

I would prefer a statistically based model rather than an agent based model, to aid scalability.

I think public land like parks and plazas should be grown and not placed, better to fit in with their surroundings. They should still cost money to run. I like the idea that a park zone in an affluent suburb could turn into a skatepark or swimming pool, if you place a higher monthly cost limit, or a basketball court, playground or barbeque pit in a poorer area.