Talk:Bourgeoisie

From Johnwiki

Contents 1 Businesses 1.1 AI Businesses 1.2 Management 2 Megacorporations 3 Property 4 Contracts 5 Skills/Workers/Consumers 6 Law enforcement 6.1 Business of Crime 7 Roads & Utilities 8 Education and Research 9 Design Draft 9.1 Workers and Leases 10 Multiple Governments, Warring States 11 Fineness of Detail and Micromanagement

Businesses

It may make the game very complex to make arbitrary businesses ownable by multiple players. In fact, it may be impossible for multiple players to cooperatively manage a business, since they may log on at different times, have differring ideas, etc... At least the way I saw it, you wouldn't get any ginormous businesses like M$ or IBM, you'd just have lots of little businesses that people can own, operate, sell, shutdown, etc... On the other hand, having large cooperatively owned businesses could be pretty cool. What if the game had some preset ones which you could by stock in - the people with the most stock would form the board of directors. People could vote for company policy by shares. Once policy is dictated, an AI would run the business - this would prevent a "too many cooks spoiling the soup" scenario.--Mike 08:31, 27 July 2006 (PDT)

Joint management is too complicated, you're right. What about a variant, where I own a business, but contract another player to act as CEO and manage the assets and bring me more profit?

Voting for company policy would overly complicate matters, I fear. They could instead vote for a CEO, and everybody would share in the profits. I guess I think the "management" responsibilities should always be controlled by exactly one player.

I don't like the idea of an AI-controlled business. Doesn't that take a lot of the fun out of the game? --John 08:48, 27 July 2006 (PDT)

AI Businesses

Having a CEO instead of an AI brings other problems, though. A CEO may wield a corporation as a personal tool to improve their assets. I think it would be very difficult to have AI determine whether or not they are doing that (and thus breaking the law). It would also be difficult to prove or disprove due dilligence with AI. --Mike 09:02, 27 July 2006 (PDT)

Couldn't the interests of the CEO be balanced against the will of the board? The operations of the company would be easily monitored by the board, so if they noticed too many deals made in the CEO's favor, they could vote him out. To keep in power, he has to keep a majority of the board members happy. If they can't decide on a leader, then the business must close. --John 09:07, 27 July 2006 (PDT)

AI-controlled businesses might be necessary in certain areas. Players might not be interested in running a business that, for example, supplies parts to a car manufacturer, but this is a necessary industry; so it could be taken care of by an AI. If at some point a player decides to enter this market after all, they will be in competition with the AI (or else will buy the AI's business). --Abe 14:24, 27 July 2006 (PDT)

I do like this idea, but the AI businesses should be passive stubs, helping the industries around them, but not too aggressive that an enterprising player can't easily displace them. Should they simulate players, or should they be state owned? --John 15:38, 27 July 2006 (PDT)

Yes, they should be passive stubs--no research or anything. What would the difference be between simulating players and state ownership? --Abe 16:18, 27 July 2006 (PDT)

I think the state will control things outside of players' hands, such as roads and utilities. Since these AI businesses will be displaced by players, they should be considered simulated players, not associated with the state. --John 17:04, 27 July 2006 (PDT)

I am hesitant to suggest any business be either AI controlled or programmed to be "passive" because it is simply unrealistic. Although I can appreciate concern that some industries may be less attractive than other in terms of "player interest", I think with proper programming the problem would solve itself. The cool thing about capitalistic competition is that there is only one goal: a high rate of profit. In this way I think that the monetary reward for both finding a necessary "niche industry" and filling it would more than compensate the barrier to entry (boredom). --Scott 16:33, 27 July 2006 (PDT)

But as a player I don't want to compete against computers, that's boring. I say the AI businesses should be passive in that they aren't aggressively pursuing profit. They should perform at the level of a very inexperienced player, enough to keep the industries around them going, but not profitably. This could be implemented by charging efficiency penalties. This can still be realistic, since it would appear they are run by incompetent management. An aggressive AI would do well in a market simulation, but this isn't a market simulation, its a game, so AI players should be minimized to the fringe of development and state-run businesses, places where they aren't making any profit. --John 16:43, 27 July 2006 (PDT)

Management

What I think y'all are trying to do is decide whether or not you want to make management of your sim-companies a function of the player or of the AI and I recommend that you consider an important inevitability of real-life corporate hierarchy when making this decision. Namely the schism between the interests of the ownership of a company (Player) and its managerial class (AI or Player). This phenomena has been discussed by any number of organizational theorists (Chuck Perrow) and while the idea that a company would make decisions against its owner's interests seems silly, it is really rather commonplace. Consider the huge of amount of managerial effort that goes into perpetuating his own job importance and protecting his access to resources. This isn't done in the interest of profitability or efficiency but rather in the preservation of self. I think that by tearing down this divide and strongly associating the day to day operations of a business with its given player you will create a much more interesting and dynamic, albeit unrealistic, business "character". --Scott 16:33, 27 July 2006 (PDT)

I believe the schism between manager and owner will be modeled well by the manager's efficiency rating, where an owner can pay more to have loyal managers. --John 17:01, 27 July 2006 (PDT)

Megacorporations

What are the advantages of allowing a single business to operate in two distant buildings/markets versus mandating that it be split into two businesses? Perhaps we can allow megacorporations, but impose strict taxes on this kind of growth, which would match the taxes earned by the same business transactions between the corresponding split businesses. In other words, a progressive tax related to the sum total of distances between every pair of buildings occupied by a business. --John 08:54, 27 July 2006 (PDT)

I think it's essential to allow--in the game's physics, anyway; what the laws are in the sim-society is up for grabs--to allow a business to get workers from one area, and export goods to another area. Initially, they could be restricted to selling the goods to other businesses, and selling direct to consumers only in their own market, but soon enough someone will try to simulate a large corporation by having several businesses with exclusive trading agreements or some such arrangement, which could get unwieldy unless the game provides mechanisms to simplify it some. --Abe 15:04, 27 July 2006 (PDT)

The issue at question with megacorps is whether a single business should be allowed to operate (in buildings) in several cities at once. The advantage to the business owner may be to solve problems of uneven skill markets (lots of biotech workers in city 1, lots of engineers in city 2) or vertical integration to avoid transactional taxes. We want to avoid megacorps and encourge small businesses and frequent trading of businesses, because it'll be more fun that way. The proposed solution is to just tax megacorps heavily until the largest corporations are no longer profitable.

Regarding sales, I think consumers should only be able to buy nearby things, while any player can sell to another with a shipping tax that depends on the distance between the two businesses. --John 15:21, 27 July 2006 (PDT)

Shipping as a tax? Couldn't that be a service provided by a business controlled by a player? --Abe 17:34, 7 August 2006 (PDT)

Yeah, it could probably be modeled by player businesses - you could buy transportation trucks and stuff. However, remember that you probably want some developer hooks into the gameworld so that you can affect game mechanics without "cheating." This way, if you feel like there's too much money in the gameworld, you could just raise shipping rates. Then again, on the other hand, there are other ways to do that too - e.g. increase the cost of gas which will increase the expenses of shipping businesses. But anyway (though I ramble), keep in mind that there should be some hooks into the gameworld without foisting them all off to players. While true capitalism may be interesting, remember that it's just a game and fun games need some method of keeping them balanced. --Mike 22:30, 15 August 2006 (PDT)

Property

Perhaps a good way to ensure that land keeps turning over is to make it very worthwhile to develop and sell land. Say, someone could buy up a falling apart apartment building, demolish it, put up condos, and then sell them for a huge profit. Or people could sell factories to AI owned industry. For this to work - i.e. for players to be willing to part with land instead of building businesses on it - there would have to be significant investment in starting a business. This is not a bad model; it is close to real life. People could build capital by developing land and then use that capital to start businesses on other parcels (perhaps even predeveloped land that they buy). --Mike 08:31, 27 July 2006 (PDT)

Contracts

Great idea! Kudos! --Mike 08:31, 27 July 2006 (PDT)

Skills/Workers/Consumers

It may make it easier to simulate if we abstract out the people, silly as this sounds. We could have it that each residential tile would generate worker and consumer points, with the skills of the workers and wants of the consumers being dictated by where they are (we could have many different kinds of worker and consumer points such as blue collar points, food consumption points, etc...), what schools are in the area, etc... Then you could just think of it as a flow problem and let the workers/consumers be distributed throughout the city based on transportation systems. Then, whereever they are distributed, that's how many workers/consumers there are in an area to be employed/shop. If memory serves correctly, this is how SimCity 2000 modeled the problem. --Mike 08:31, 27 July 2006 (PDT)

I agree. It might work to have a business need a certain number of factory workers, and the nearby tiles are queried for available factory workers, who are then allocated to the job. --John 08:22, 27 July 2006 (PDT)

So I'm thinking that one of the benefits to this game would be inherent parallelization of tasks. Computing who goes where and how traffic, crime, pollution, and people ebb and flow throughout the city can all be handled on different machines, if needed. --Mike 08:31, 27 July 2006 (PDT)

This worked for SimCity because a whole-city simulation doesn't need very fine resolution; but if someone is running a small business with just a few employees, they might want to be picky about who they hire, so they might want to look at individuals, rather than aggregates. It's more computation-intensive, but it seems important to get the type of simulation we're going for here. --Abe 14:45, 27 July 2006 (PDT)

I see what you're saying about how the coarse resolution may make smaller businesses show less profit than they could. It is an issue that probably needs to be addressed. I do have a few points arguing for the other direction, though. 1) Should the game be about micromanaging small businesses? If you want to do that, wouldn't a single-player small business simulator be more appropriate? The heart of a multiplayer game is cooperation and competition between players. So abstract out the small details, like exactly whom you're hiring, and let the players get to wheeling and dealing with each other. 2) People who manage multiple businesses don't manage each and every one of them; they set overall policy and make the large decisions while letting other people oversee the details of management. 3) If this is to be an MMOG, cities will probably need to be very large and have perhaps hundreds to thousands of players. Algorithmic issues, where you need to trade detail for speed, are sure to crop up. I'm thinking that an iterative coarse grained algorithm, though not necessarily accurate, will probably at least be both smooth and close enough that players won't realize the approximations that are taking place.--Mike 15:14, 27 July 2006 (PDT)

To points 1 and 2, I say that while some people may prefer to build vast empires with many businesses, others will probably be interested in micromanagement, and it would be nice to cater to both styles, if possible. For point 3, maybe a good compromise would be to keep track of the averages in general; then when a player wants to hire someone, to generate individual candidates with traits based on these averages, but with random individual variation thrown in for spice. The player can then pick from among several such candidates, with the number and quality of candidates of course limited by the surrounding population. Those who aren't so concerned with micromanagement can get a little bonus over the averages, both to make up for the fact that micromanagers can be picky and to reward them for easing the computational burden on the servers.

Of course, that might be more trouble than it's worth, but I think it would make the game more interesting, especially if we want the focus to be on small businesses rather than on large enterprises. The micromanagement style of play might even allow for the possibility of individual worker personalities, which would be another whole dynamic--not to mention competition for workers as well as for customers. Watch out when players start trading goods and services for employees! --Abe 16:16, 27 July 2006 (PDT)

Don't players of most games micromanage in order to optimize, the theory being that the best players are better the automated process? Wouldn't players be angry if micromanagers were punished? I think the averages in each case (micromanagement vs. automation) should be the same at each level of decision making (budget, employment goals by skill, and hiring). I do like the idea of selecting your employees from workers generated ad hoc from the labor pool, though. But once you hired them, they would have to lose their personalities and just be rolled into your company's general employment profile. I don't think employees should be sold as slaves, but a business could control the labor pool by hiring all workers of one skill or skill level (think Google and PhDs). --John 16:29, 27 July 2006 (PDT)

Well, it's not that a player couldn't do better by micromanaging, it's that those who micromanage will have to be relatively skilled at it to benefit from it. A micromanager can reject poor applicants, so, similarly, the automated manager shouldn't just accept the first randomly generated applicant that comes its way; to simulate the screening process, the random numbers can just be biased a little high.

The general logic of this argument can be applied if aggregates aren't used, too; and I have argued in "Workers and Leases" that they shouldn't be. --Abe 17:19, 7 August 2006 (PDT)

Another thought: The bias could be affected by the quality of the managers. Hire good middle managers, and you don't need to micromanage so much. --Abe 17:42, 7 August 2006 (PDT)

Law enforcement

The approach to law enforcement isn't completely specified in the article, so I want to suggest a particular approach to it that grew out of a side-discussion with John.

The basic idea is to have the police function as a business of sorts, subject to some of the same forces as other business. Its revenue comes from taxes, so if there isn't much business being transacted the police might become underfunded and have trouble enforcing laws. Rival "protection services" (a.k.a. the Mafia), run by players, could then spring up to replace the police.

The police don't have infinite resources, so crimes only have a certain probability of being noticed; thus, black-market players can get by without paying bribes in areas with weak law enforcement. Furthermore, officers may be corrupt to varying degrees, so some would be virtually unbribeable while others would be cheap. The price would depend on the officer's ethics (a "skill" that could be trained by the police, as any employer might offer or require job training) as well as other factors, such as the officer's income (someone making a lot of money already has more to lose by accepting a bribe) and oversight (self-policing--after all, the officer who accepts a bribe is himself committing a crime). Allocating more resources to officer compensation and internal investigations could cut down on corruption, but would limit the resources available for other aspects of enforcement.

I offer this as an alternative to John's original idea, in which the police have infinite resources (catch all crimes) but are infinitely corrupt (will always accept bribes). My goal was to come up with a setup that would make the black market more interesting, and allow for organized crime. In fact, if implemented thoroughly, this system might make it possible for some upstart protection service to replace the actual government; the original government would become increasingly impotent, as failure to collect taxes led to lack of income, leading to a continuing failure to collect taxes, and meanwhile an initially Mafia-like organization would become more and more respectable and effective at keeping peace, and "protection money" replaces taxes. At that point a revolution has occurred--whether for good or for bad, only time will tell.

Importantly, this shouldn't be much harder to implement than the infinite-resources idea, and might be easier--after all, the police will be just another business, albeit with an odd way of collecting revenue; all the mechanics devised for other aspects of the game will be readily applicable to law enforcement.

John raised the issue of conflicts of interest, and this is a good point, since if someone owns the police as well as another business, they might choose to let that business get away with certain things. However, conflict of interest is a fact of real life as well, and might be a good thing to simulate. A political system allowing elections of new leaders might prevent the most obvious conflicts, but if democracy doesn't work, then it can be replaced in a revolution. --Abe 14:21, 27 July 2006 (PDT)

The problem, that I see, with allowing rampant corruption is that it skews the game. It makes it easier for the rich to get richer and the poor to get poorer. Somebody who's just entering the game may get all of his/her businesses torched by mafiosos because he/she couldn't meet the huge protection fees they were charging. This is not a good way to keep people involved in the game. If you have a fairly competent (but not infallible) police force, it protects the weak from the strong, since it's not worthwhile to tangle with the law just to kill some small fish. However, the risk of coming up against the police may be more justified in a coup between two rival players with huge fortunes. To sum it up crime should not pay, on average. --Mike 14:41, 27 July 2006 (PDT)

I agree that a rampant Mafia could cause problems for new players, but it should be the responsibility of the police to protect their monopoly on the use of force. Hopefully, if the universe is large enough, there will always be some place or other with a well-run police force of some kind, and business in such a place will tend to boom, at least for a while.

Maybe the solution is just to have certain areas where the police cheat in some way to stay ahead of the organized crime rings, ensuring that newcomers can test the waters before facing the sharks; but there should be something, such as high tax rates in the safe zones, to encourage people to move into the wider world. This way we simulate political as well as economic systems. --Abe 15:00, 27 July 2006 (PDT)

I don't think the owner of a police station would have the "powers" of the police, so it is unlikely he could focus his police powers to or away from any nearby business. Instead, he could only affect the general level of policing in an area. This would be advantageous to a drug dealer, mafia boss, or real estate agent (to drive prices up or down); basically anyone interested in seeing more or less crime in the area. In my opinion, this is what bribes are for. Putting a player in charge of the police station just elevates one member of the neighborhood over the rest in terms of controlling the general police presence. Would this make the game any better? --John 15:29, 27 July 2006 (PDT)

Business of Crime

Should crime be profitable? I think it should be, but only for the most experienced players. In other words, police should prefer to protect businesses than accept bribes from criminals. --John 14:49, 28 July 2006 (PDT)

My thoughts are the crime should be a tool that, if wielded with skill, yields advantage in rare circumstances. Allowing crime to be profitable, in general, removes the law and order of the business world and replaces it with chaos. --Mike 00:17, 30 July 2006 (PDT)

Roads & Utilities

The state will decide where to put roads, but players could operate contracting companies to build them--companies which other players could always hire to build private roads, of course. And utilities could be deregulated too, so we could see if the results are as horrific as in real life. --Abe 08:14, 28 July 2006 (PDT)

I agree that the state should contract local construction companies to build roads, lay pipe, so on. Perhaps a player can bribe/coerce officials to win contracts. I don't know about utility companies though. Perhaps we can base our model on the real-life practice where by people can contribute to the power grid and get cash from power companies. Would this work on a large scale with electricity, power, and telecommunication suppliers? We could have a commodity market for utilities, so instead of buying power from one business, you buy from all busineses equally at the same price. This would allow utility companies to compete based on production efficiency. --John 12:54, 28 July 2006 (PDT)

We need to distinguish between what will be possible in the game, and what will actually happen in any given city. It might or might not be interesting to deregulate utilities, but the game should allow for it so that some cities can try it out. This is similar to what I suggested in the "law enforcement" thread, that some regions should have permanent governments supported by the game engine, while others should allow for revolutions to occur. Some people will want stable energy prices and a stable polity, while others will want to try to game the energy market and start revolutions. The game engine should allow for all kinds of tastes. There could even be embargoes between unfriendly cities/regions, and smugglers breaking the embargoes to feed a hungry market; why not? That's capitalism. --Abe 13:10, 28 July 2006 (PDT)

That is an idea that Scott suggested to me last night. Basically we should parametrize the cities enough to allow very different modes of gameplay in each. Then, instead of deciding between the mutually exclusive X or Y, we allow a city to use either. Requires more abstraction, but I'm in favor of this. --John 13:22, 28 July 2006 (PDT)

Education and Research

It seems to me that the most fundamental issue for this simulation, which hasn't been addressed yet, is that of education and research. What skills will be available? How will new skills be added? How will citizens gain skills, how will they maintain them, and how will they be limited in the amount of skill they have in any area, or the total number of skills they have? What sorts of products will businesses be able to make and sell, and how will new products be created? These are tough questions. If a product can be improved endlessly in any particular direction--say, if enough research can make cars as fuel-efficient as you care for them to be, or if enough research can eliminate all side-effects of a drug--then a lot of realism will be lost; but if every new product or improvement to an existing product has to be approved by some central body, or if all possible products and inventions have to be planned out ahead of time, innovation will be stifled and the game won't be as much fun as it could be.

I don't have any very concrete suggestions to make regarding this issue, I just thought it was something that needed to be brought up. --Abe 13:32, 28 July 2006 (PDT)

I see research producing patents or secrets (special kinds of products) that affect the efficiency of a company (production technologies such as chip fabrication) and qualities of the product (consumer technologies such as transportation). This should be easy to handle. The problem you describe breaks down into how to define products and skills. The only idea I've had is to explicity define sets of skills and products for each game instance. Afterwards, skills and products can be manipulated by admins to keep things balanced and fun, but the maintenance should be minimal. Can anyone offer a method to allow players to create skills and products without making the game too complex? If I want to start selling widgets, how do I define the properties of a widget? What can it do? I think it'd be easier to make products flexible enough to allow one company's products to be distinct from anothers, but they are still the same abstract product. Likewise for skills. --John 22:51, 28 July 2006 (PDT)

Design Draft

This is my attempt at mocking up some of the relationships envisioned so far. The black arrows point from children to parents, so a worker is a child of the building he works in. The red arrows point from subclasses to superclasses, so a worker is a consumer.

Obvious deficiencies: many-to-many relationships between businesses and owners, businesses and buildings.

Glaring flaw: a consumer may not own his residence, since by this definition a residence is also a place of business. The fix would probably be to split buildings into workplaces (held by businesses) and buildings (held by properties) and allow consumers to be owners (even if they don't own anything). But that would make my arrows too messy.

I am writing this up in python, too. --John 22:50, 28 July 2006 (PDT)

Two things. Firstly, what are you writing up in python? Secondly, I think a simpler way to do things would be as follows (and I'm not gonna make a pretty picture so deal with it): Developments (of which buildings are a type) reside on land. Land can be owned by either players or by the computer - there's no real need to differentiate who the land is owned by if a player doesn't own it. Land has properties, such as number of people who live there, crime, pollution, urban blight, etc... These properties are influenced by the surrounding land and any developments on the land. Businesses can be owned by either the player or the computer. A business resides at between 1 & n plots of land. Businesses receive products as inputs and produce products or services as outputs. Am I missing anything? --Mike 00:25, 30 July 2006 (PDT)

Okay, I definitely appreciate those points. I originally intended a Worker to be a group of employees with the same skills, but I see now that my model of Worker, Cosumer, and Residence is far too complex and unwieldy. I still think my Owner, Property, Business, Building model is pretty good though. The question that is bothering me is how should we model the availability and skills of the labor pool of a locality (Property) and the labor used by a business? For instance, the 100 low-skill factory workers living on this property are split 60-40 between two factories. What is the best way to model that? (I'm using Python because it's great for prototyping and its list processing can't be beat.) --John 11:32, 30 July 2006 (PDT)

No, no. Not why are you using Python. What are you writing in Python? A simulation? The only issue I take with the Owner, Property, Business, Building model is that it still doesn't allow accurate modeling of relationships. First, I think it makes much sense to just simplify things so that only one business can reside per property. That's just a point of taste and aesthetics, however. My real issue is dealing with buildings/properties that aren't being used by businesses. Why can't players rent out buildings (be they apartments or office buildings) to computer businesses. Like, I'm not sure how much point there is in making players start commercial businesses - why not just let them put up commercial buildings and then just rent them out to businesses. Also, buildings could be vacant, they could be condemned, etc... The model just doesn't seem to handle those issues very well. --Mike 13:02, 30 July 2006 (PDT)

Oh, I haven't done much. I'm just trying to build a usefuly prototype implementation of the model. Once that's done I'll probably try a simulation. How does this sound for a new model of businesses. Properties and businesses are held by owners. Buildings are held by properties. (So far no differences.) Then, a Lease, a subset of a building's space, is held by a building and a business. In a trivial Lease, the entire building would be owned by the business (the company's owner). In a non-trivial or complex Lease, the entire building or parts of a building would be allocated to businesses owned by other players. This allows me to rent out my buildings to other companies and to spread my company out through many buildings (allowing growth). Plus we can also model condemned buildings (no leases) and undeveloped properties (no buildings). Properties without owners make as much sense as buildings without properties, but we can just say an abandoned property is owned by the state. The only limitiation to this model I see is that you can't own a building on someone else's property, but doesn't that seem unrealistic? Also, a little logic will be required to make sure the parts of a building are disjoint. What do you think? --John 15:53, 30 July 2006 (PDT)

The problem of many-to-many relationships and a worker owning his own residence could be solved by letting a business have a reference to its owner, and an owner have a list of references to businesses; letting a residence have references to residents, and residents references to residences, etc., rather than making everything be data members of everything else. Thus, the worker who owns his own house has three capacities: Owner, Worker, and Consumer. Qua Owner, he has a reference to the house as a thing owned. Qua Worker, he has a reference to whatever business he works for. Qua Consumer, he resides in some building, which can as easily as not be the same one he owns qua Owner. The three capacities may well be different objects in memory, but they draw on the same bank account. If he has multiple jobs, maybe his Worker object has references to two different businesses, or maybe he has two Worker objects. --Abe 18:18, 7 August 2006 (PDT)

Workers and Leases

Here's my new model. I also had some thoughts on the problem representing labor pools. By property (or property cluster) record the total residential capacity (based on the residential buildings) and residential population (or vacancy). Then for each skill (or skill class), store the number of jobless workers. When a worker with skill S is created by the game, send that worker to live in a random property P with vacancy, with weights based on the demand for S in that area, the cost of living, etc., and add 1 to the number of jobless S at P. When a business wants to hire (fire) workers of skill S at lease L (a building-business pair), they need only decrement (increment) from the pool of jobless S in properties near L and increment (decrement) the number of S they employ at L. There will have to be some rules for non-workers, too. For instance, when workers move by their own volition, they take along k non-workers with them. The fun part is, at each stage (moving, hiring, firing), we only deal with aggregates instead of individuals. The ultimate lower class marginalization! --John 17:45, 30 July 2006 (PDT)

What about employee turnover? Employees may leave one business to go work for another so a new business that starts in an area which is already tapped out in terms of workers may be able to pull some away from other surrounding businesses. This could have some strategic consequences, like forcing most of the businesses in an area to not be profitable so one (or a consortium) can buy them all out. --Mike 20:09, 30 July 2006 (PDT)

What about workers with multiple skills? E.g., say I'm looking for someone with management experience and experience in food service. The simulation needs to know not only the proportion of workers with each skill, but the amount of overlap between the two groups, to determine how many people are competent to help me run my restaurant. Note that this can't just be calculated from the individual proportions. Say there are a lot of restaurants in the area, and a business school as well. Then there are lots of waiters and busboys, and lots of future managers, but the students at the B-school are wealthy and don't need to wait tables to make a living. Thus there isn't much overlap between the two groups. But to keep track of all the combinations--multiply out by a bajillion different skills, and that's a bajillion-factorial combinations.

A possible compromise would be to limit the number of distinct skills a job can require, but for a number S of different skills and a number R of skills any particular job can require, there are still S^R combinations to keep track of, which, for any reasonable values of S and R, is still astronomical. If there are 10 skills (way low), and any particular job can require up to 4 (reasonable, in my opinion), then that's (more than) 10,000 variables to keep track of--in other words, as long as the population is below a thousand, it's more efficient to keep track of individual workers than aggregates, even if each individual has a value for his ability in each skill. Make it 100 skills, and it's 100,000,000 variables; now it's more efficient to keep track of individuals even if there are a million of them. Furthermore, if the population is treated individually, R can grow arbitrarily large without taxing the system. In Big-O terms, the memory requirement of the aggregate method is exponential with R and (for R = 4) quartic with S, whereas that of the individual method is linear with S and invariant with R. In other words, never mind marginalizing the lower classes, treating workers in aggregate may limit the flexibility of the whole skill system. To put it cynically: The rich capitalists can make better use of the poor by treating them as individuals.

Besides, I think it could be interesting to pull a worker out and look at his life history. How many rags-to-riches stories can we simulate? Will the rich just get richer, the poor poorer, or can we reduce the income gap? How old is this manager I just hired? Did he finish high school? Did he finish college? What was his degree in? Maybe I'd rather hire a smart young buck with lots of ambition than an older, more experienced guy who wants to spend time with his family and not necessarily grow the business--if the game simulates ambition. And so on. --Abe 16:21, 7 August 2006 (PDT)

You gave an example where two skills happened to have very little overlap. Can you offer an example where we should see skill overlap? How will preventing overlap hurt gameplay? --John 23:17, 16 August 2006 (PDT)

Multiple Governments, Warring States

There has been talk of 1. creating separate instances of the game and 2. characteristics of a state government. I suggest we commit to both of these concepts by defining an instance of the game as a nation-state with an autonomous government. Different game properties simulate the effects of culture and relgion. Players from different nations would be allowed to trade with each other, or the nations would set up trade agreements or embargos or make war depending on how the players affected the government through lobbies and bribes. This would allow situations similar to the MPAA threatening Sweden with WTO sanctions if they didn't shut down PirateBay. The United States may control the WTO, but the MPAA and RIAA control the US, or at least can issue empty threats through it. Plus, with military powers, nations could also be bribed to use militia to break up strikes, since usually the police aren't powerful enough to combat 10,000 organized workers.

This government would be artificial, for development simplicity, but we could roll out a companion government-simulation for the 2.0 release. (I think it should be called SuperPowers. Yes I'm already planning the sequel. Now back to reading Howard Zinn.) --John 19:27, 30 July 2006 (PDT)

Further talk with Mike has made it clear that "globalization" or "internationalism" is by no means necessary in an initial design. In addition, "striking" need not be modeled so complicatedly (or at all), so militias won't be needed at first either. I still think these are good ideas, though, and we should discuss them a little more and think about how to incorporate them later. --John 22:58, 2 August 2006 (PDT)

Striking could me modeled fairly simply, and with no extra features dedicated to it specifically. Just allow a worker to belong to more than one "business", make "intimidation" a type of skill, and voila--some people will form gangs, some will form unions. The unions might bully workers into joining--stranger things have happened--and then turn around and bully corporations into doing good things for those workers, like giving them raises. The unions might also become corrupt and join forces with the corporations. The point is that the union is, in an abstract sense, just another business--its income comes from dues, and its product is protection from evil employers. Strike-breaking mercenaries are another type of worker. Will the game model loyalty?

Actually, part of my point before about revolutions, etc., was that no special game mechanics would be necessary for that either. The same forces that allow people to form gangs and unions might also allow them to form governments.

...Well, okay, so unions would require some bargaining mechanics, which governments would use for a judicial and legislative system, but which businesses could get along without. So maybe 2.0 after all. Still, I think that sort of capability would be of benefit to regular businesses as well. --Abe 16:45, 7 August 2006 (PDT)

Fineness of Detail and Micromanagement

This concept has been brought up in several areas so I think we should try and abstract it out a bit. How many details can we/do we want to include in the game. This is a complex subject on two fronts: 1) Finer detail requires more computing power to simulate 2) Micromanagement is complex - those with vast empires won't be able to personally handle each business/property while, at the same time, those with smaller empires must have enough to do that they can spend enough time playing each day to have fun (we must have some concept of time; erecting a building may take as much as several days of real-world time and generating money from businesses/properties also needs to take real-world time). So, thoughts anybody? --Mike