SimCity: Welcome to the Glassbox Simulation Engine

[JaceTwice 115]

I wonder though ... whether an ability (regardless of introduced by a mod or present as a feature) to turn off the agent system would remove what kind of limits on the city size and the general size of the simulation. Think of it in terms of sacrificing "depth" for "width" of scope.

I'd imagine you would break the system entirely if you tried to do that. Resources couldn't move at all and you'd end up with nothing more than empty roads and zones.

[marksman1416 12]

I wonder though ... whether an ability (regardless of introduced by a mod or present as a feature) to turn off the agent system would remove what kind of limits on the city size and the general size of the simulation. Think of it in terms of sacrificing "depth" for "width" of scope.

I'd imagine you would break the system entirely if you tried to do that. Resources couldn't move at all and you'd end up with nothing more than empty roads and zones.

Agents are what control everything like power and water.

[macvirt 361]

I wonder though ... whether an ability (regardless of introduced by a mod or present as a feature) to turn off the agent system would remove what kind of limits on the city size and the general size of the simulation. Think of it in terms of sacrificing "depth" for "width" of scope.

I'd imagine you would break the system entirely if you tried to do that. Resources couldn't move at all and you'd end up with nothing more than empty roads and zones.

Agents are what control everything like power and water.

Perhaps. I do wonder though, GBE is a graphics engine first and a data abstraction model second. We've seen that before in the industry, in most such projects it turned out that what was presented and perceived as a data agent system was essentially a seperate engine (so to speak, a math model tied to handlers for abstraction and visualisation). It can indeed be the case that the agent system is the motivator element of what the user works with, it can however also be the case that it is just simply used as a means of selective enablement and access.

In simple terms, the agent system can very well be a snapshot mechanism allowing the user to interface with the visualisation of data and a transaction mechanism across sessions or states (the multiplayer element and/or regional mapping). It does not have to be more than that, in terms of game design. Which is basically why such mechanisms typically aren't.

Some people might see that as a "cheap" angle, it is not however. It is a valid method within both game design and the project management. Look at it this way, you could invest a disproportionately large part of budget to develop the perfect agent and transaction mechanism. Or you could develop for specific requirements. I could imagine the approach where the agent and transaction systems are present and configured for the requirement of multiplayer interaction. Which, considering the design of online play, is a prime requirement for the scope of the design. If Maxis made it more than that, there is potential for turning that off as a cost efficient means towards solo and/or offline play. If Maxis only went for the prime requirements, that will simply not be possible.

[Ocram's Razr 964]

I wonder though ... whether an ability (regardless of introduced by a mod or present as a feature) to turn off the agent system would remove what kind of limits on the city size and the general size of the simulation. Think of it in terms of sacrificing "depth" for "width" of scope.

I'd imagine you would break the system entirely if you tried to do that. Resources couldn't move at all and you'd end up with nothing more than empty roads and zones.

Agents are what control everything like power and water.

Perhaps. I do wonder though, GBE is a graphics engine first and a data abstraction model second. We've seen that before in the industry, in most such projects it turned out that what was presented and perceived as a data agent system was essentially a seperate engine (so to speak, a math model tied to handlers for abstraction and visualisation). It can indeed be the case that the agent system is the motivator element of what the user works with, it can however also be the case that it is just simply used as a means of selective enablement and access.

In simple terms, the agent system can very well be a snapshot mechanism allowing the user to interface with the visualisation of data and a transaction mechanism across sessions or states (the multiplayer element and/or regional mapping). It does not have to be more than that, in terms of game design. Which is basically why such mechanisms typically aren't.

Some people might see that as a "cheap" angle, it is not however. It is a valid method within both game design and the project management. Look at it this way, you could invest a disproportionately large part of budget to develop the perfect agent and transaction mechanism. Or you could develop for specific requirements. I could imagine the approach where the agent and transaction systems are present and configured for the requirement of multiplayer interaction. Which, considering the design of online play, is a prime requirement for the scope of the design. If Maxis made it more than that, there is potential for turning that off as a cost efficient means towards solo and/or offline play. If Maxis only went for the prime requirements, that will simply not be possible.

You are simply misinformed. The GlassBox engine, according to the earliest videos from GDC and available on the SimCity website, is solely about buildings that run rules (consuming and/or producing resources), paths (roads) that transport agents, and agents that carry resources. These are all intertwined in the Glass Box engine.

--Ocram

[macvirt 361]

You are simply misinformed. The GlassBox engine, according to the earliest videos from GDC and available on the SimCity website, is solely about buildings that run rules (consuming and/or producing resources), paths (roads) that transport agents, and agents that carry resources. These are all intertwined in the Glass Box engine.

--Ocram

As I said, a transaction engine tied to a graphics engine, the amalgam of hooks & handlers for interpretation and interaction (which is essentially the logical evolution of Glassbox from its first use in another product to its further development for this product). The question is to what degree the dependancies for data handling are configurable. Consider for example whether such dependancies are only on|off if cheats is enabled. Is cheats on|off the same as multiplayer on|off, or are there configurable combinations possible.

The same on the element of data collection / handling / tracking / presentation. Is the agent transation system is a snapshot mechanism, or (for example) is the span of it dependant on system capabilities, or are there hard values / variables.

[Ocram's Razr 964]

I presume that commuter agents are similar to simplified Tropicans. In Tropico (3 and 4), electricity is similar to a data layer with radii from power plants and substations as well as maximum consumption and production (surplus) but Tropicans (citizens, commuters, and tourists) are persistent from birth/immigration until death/emigration and all people are always accounted for no matter what settings you have. That is one reason why the Tropico series has such high system requirements. There is a hard coded limit of 200 tourists and 1000 citizens in Tropico 3 and Tropico 4 and it is very difficult and dangerous to mod that out (and by doing so, you pretty much void your warranty, no more technical support) therefore, I presume that there will be no mod to increase or remove the limit on the number of agents in any given city though expansion packs (with higher minimum requirements) may increase the limit by a certain amount. The new SimCity 2013 has an ingenious system where agents are somewhere between persistent (they exist only as they travel between buildings along paths) and statistics (when people and resources are inside buildings, they exist only as statistical resources and are displayed as mere animations). I love the Glass Box engine and I hope that expansion packs and spinoffs (but no more sequels to SimCity too early after release, lest the servers for SC2013 go away) are released. Since I have been playing SimCity 4 and Cities XL, but mostly Tropico 3 in preparation to the release of SimCity 2013, I have realized how big of a challege free elections are to a city builder game. Today I played a sandbox game for several hours in Tropico 3 with free elections on (must hold elections every time demanded, cannot commit voter fraud) and a couple of times I had to desperately enact certain edicts that increased happiness and respect, even going so far as bribing people) and I had to choose wisely what type of speech I gave though in the end I amassed nearly $200,000 in my Swiss Bank Account. So I think that it would be a wise decision for the new SimCity 2013 to have an option to hold elections where the opponent is either an NPC or another player in the region and I think it would be pretty fun if "retirement funds" could be added as well.

--Ocram

[BLANKBLANK 4,414]

While I like the notions of resources and these new inter-city mechanices, I fear the Glass Box engine will be nowhere near as engaging, or complicated or as versatile as the SC4 one. The core of every simcity game has been RCI and SC4 perfected that, I just hope RCI doesn't become less important (or worse still simplified and dumbed down) and instead non-RCI buildings become far more important.

While in SC4 non-RCI were of course vital, they alone didn't determine what kind of buildings would be built in RCI zones. So please Maxis please retain the following: stage limits, demand caps, wealth types, population stage limit ratios, and all the other things which made the SC4 simulator so amazing. I certainly don't want to see skyscrapers rising up in a city of less than 10000 because that wouldn't be realistic and SC4 ensured such things would never happen.

[halicem 16]

It's been awhile since I last played and downloaded new content for sc4. Went to my Account section to see what else was updated and found the section which lists all my posts. One in particular stood out:

I post this here because I think Maxis did the right thing with GlassBox, it is the next evolutionary step to the simulation. I detailed in that post why I didn't think they'd be successful then, but they did build it - albeit they had to make it smaller than what most would have liked.

Reading through my old post, I found it chilling at how close it was to describing how the agents are supposed to work. I'll quote the closing section of the second paragraph:

Therein lies the problem, if everything were on-demand, and its a living and breathing city, millions of objects change status every second which ends up swamping the computer. Ever notice how on the latest games with the latest graphics, whenever big explosions happen, frame rate drops to pitiful! That's because all particles in the explosion is a separate object which the computer has to keep on recomputing where it's supposed to go and how it affects other particles in it's area. SC5 will be like that, only much worse because there are multiple cause and effect scenarios and multiple collections of collections of objects. And unlike explosions which could be turned down, there is no such setting in SC5 because you're talking about the logic of the city builder itself.

I think the GlassBox does have the capability to be really really complex, it should, in theory, surpass the complexity of SC4's calculations. Everything I've seen and heard so far suggests they were able to do just that. Now it just seems to be a matter of whether or not they've been able to unleash it to its full potential or will they be purposefully hold it back and wait to see how things go.

[iomex 24]

While I like the notions of resources and these new inter-city mechanices, I fear the Glass Box engine will be nowhere near as engaging, or complicated or as versatile as the SC4 one. The core of every simcity game has been RCI and SC4 perfected that, I just hope RCI doesn't become less important (or worse still simplified and dumbed down) and instead non-RCI buildings become far more important.

While in SC4 non-RCI were of course vital, they alone didn't determine what kind of buildings would be built in RCI zones. So please Maxis please retain the following: stage limits, demand caps, wealth types, population stage limit ratios, and all the other things which made the SC4 simulator so amazing. I certainly don't want to see skyscrapers rising up in a city of less than 10000 because that wouldn't be realistic and SC4 ensured such things would never happen.

I don't know where you're getting any of this from. RCI works in the same way as SC4. Infact, there's probably more in the way of sophistication here. There's even tech levels for RCI, buildings will not progress to the next development stage unless conditions are met, and so on.

[BLANKBLANK 4,414]

While I like the notions of resources and these new inter-city mechanices, I fear the Glass Box engine will be nowhere near as engaging, or complicated or as versatile as the SC4 one. The core of every simcity game has been RCI and SC4 perfected that, I just hope RCI doesn't become less important (or worse still simplified and dumbed down) and instead non-RCI buildings become far more important.

While in SC4 non-RCI were of course vital, they alone didn't determine what kind of buildings would be built in RCI zones. So please Maxis please retain the following: stage limits, demand caps, wealth types, population stage limit ratios, and all the other things which made the SC4 simulator so amazing. I certainly don't want to see skyscrapers rising up in a city of less than 10000 because that wouldn't be realistic and SC4 ensured such things would never happen.

I don't know where you're getting any of this from. RCI works in the same way as SC4. Infact, there's probably more in the way of sophistication here. There's even tech levels for RCI, buildings will not progress to the next development stage unless conditions are met, and so on.

It's all there in the SC4 Prima strategy guide, demand caps, stage limits and so forth are all the technical game terms which are vital to growth (or its restriction) in RCI, it's a lot to consider and that's what made SC4 so hard but in a good way. Decoding the simulation was half the fun of playing SC4 and I just hope the Glassbox Engine delivers the same experience...

[strokesboy21 11]

im looking forward to this day by day but noticed in the video if people are halfway at a crossing point the car just goes through them :s hope that gets worked out