Simcity 4 - The 'Technological Advancement' of Modding

[Kiiroitori 55]

Have you guys ever notice that as Simcity 4 progresses, modding seems to be higher quality, range and complexity then ever before. I mean I guess this has been due to the ability of tools available by Maxis, then later the community it self. But I've noticed over the years myself, since I  first started encountered Simcity 4 in late 2005, the complexity and quality of the content has significantly advanced and new ways that we never thought were possible are being discovered enabling things, hence producing higher quality or more complex mods, despite the hard coding limitations that we have come across. Who knows what advancements in modding could hold? Imagine all the possibilities if Maxis/EA were to put the game into open source, the limitations would be lifted and the breakthroughs will be inspiring. 

 

P.S I like to thank everyone in the community, past and present for contributing to this great game to what it is today.   

 

So what do you guys think? 

[A Nonny Moose 8,261]

The impact of releasing the source on EAs revenues might be serious, what with the current flop and client disenchantment.  It is likely that players of the new edition will quickly get bored with it and either return or adventure into SC4.  Don't hold your breath on an abandonware situation.

 

Also, if SC4 became abandonware great care beyond fixing some of the bugs would have to be taken to avoid invalidating the current enormous set of custom content.  Essentially, a fixed up kernel might finally eliminate the problems with the old bus, but what would a 64-bit decor do to some of the existing mods?  Wouldn't take much of a tweak to invalidate the NAM.

 

Best let sleeping dogs lie.

[Tarkus 6,932]

Indeed, it is amazing to think how far we've come with this game, and it's really neat to see it maintains an active, breathing content creation community after almost 11 years.

 

Who knows what advancements in modding could hold? Imagine all the possibilities if Maxis/EA were to put the game into open source, the limitations would be lifted and the breakthroughs will be inspiring. 

 

 

This may sound like an unusual response, coming from a long-time mod creator, but I would actually prefer it if the source code was not released.  The chances of it happening, as my cervine friend above has stated, are very slim.  The game's actually still turning a profit and putting up surprisingly decent sales figures for its age.

 

Beyond that, it'd open up a Pandora's box (well, actually, the historical record has shown that Pandora had a jar, not a box, but that's another matter).  There is a common assumption among the general userbase that the types of things we've done with the game would easily translate to mucking around in the source code, but the fact of the matter is that it's a completely different animal than what we've been doing for the past decade.  The source code is in C++, whereas a lot of the stuff we've done is either using SC4-specific script formats and file formats, or raw hexadecimal.  While there are a few of us on the NAM Team with some degree of programming language knowledge, there's not very many of us with C++ skills, let alone at the level needed to pick apart a commercially-released simulation game, so most of the existing advanced modding community would be of little use in dealing with the source code.  We've actually had the ability for some time now to be able to produce DLL plugins for SC4, which are built off C++ code, and would theoretically be able to tap into similar types of things, but very few have been made, and most are very rudimentary.

 

There's also still a ton we can do without opening that can of worms.

 

-Tarkus

[Kiiroitori 55]

Hmm, you do have a point. Well knowing that its EA. of course they wouldn't let this game go rough seeing how far this game has gone, and in turn continue to generates profits to some degree. Even for its age. However, it is true that there is a clear distinguishing line between modding, and well the hard source C++. Although perhaps IF they did abandon the game (Which much of you has said that the possibility is very limited to an extent), then perhaps I wonder where the game go from there? (If the community was still intact of course)

[Fatherted 129]

The community would shatter as source is butchered then released by amateur nobodies. Not good.

[Cyclone Boom 17,016]

It's fascinating how the quality and range of content has improved. Just to think, it all started 10 years ago with plain old re-lotted Maxis buildings. At the time, this was considered revolutionary! Once the BAT was released, it allowed serious new additions for the first time. Unique content is what maintains people's interest in the game. We started to see recreations of many real-world landmarks, leading to City Journals. Nowadays with new software such as 3ds Max, the quality of BATs has been improved. With DarkNite models we now have more realistic nightlights. It always amazes me how the bar seems to be raised every year.

But most mods owe a big thanks to iLive's Reader. This tool has been fundamental to creating everything we know as "game changing". From the NAM to the CAM to the SPAM. To think, this is what allowed us to view and interpret the game data files, and create modifications.

For the source code, I also see no reason why Maxis would release it in the near future. To put things in perspective, it took until 2008 before the original SimCity (1989) was made open source. But I agree, since SC4 is still collecting sales and still popular, there isn't really a logical reason to do this. And yes, an open source game has the potential to be a mad dog without a lead. There are limitations to any game. Finding ways to get around them would not be easy, even with all the code at hand. Many of the things (e.g. the grid, single threaded architecture) may require superior knowledge to find a solution.

Even so, changes to the source could essentially make all mods incompatible. In this case, there would be no real benefit. We'd have to start again from scratch.

[zahrul3 472]

As a BATter, alot of the advancements came through with BATters using the BAT as a tool of art, often taking several months just to release one(but extremely detailed/large) BAT. This is in contrast to early BATters like Simgoober and the NDEX, they came up with okay BATs every other week, but none were very detailed(often with manipulated textures) and considering the limitations of the average Pentium 4/Athlon, they simply weren't able to. A lot of the prop packs that came with it were very grainy and simple. Compare that with CP's and Mattb's recent prop packs, they are much more detailed than their previous ones.

 

Then comes BAT4Max and the HD renders that came along with it. Beforehand, BATters utilizing 3dsmax would have to import it into gMax beforehand. You could notice the difference in Ill_Tonkso's pre-2008 and post-2008 work, especially in night renders. This was before the method to producing excellent reflections was widespread, so his BATs simply reflected the sky, despite that by logic, from SC4's camera, it shouldn't reflect the sky.. Most excellent BATs nowadays have reflections that realistically face the ground after reflections became widespread knowledge.

 

Then came the HD renders and textures. Introduced into BAT4Max, it allows small props that would otherwise be shown as grainy in game be finely detailed. Some larger buildings are rendered HD. HD textures came independently from HD BAts, but they came from advances in computer VGAs that could handle them. Having a city full of HD props, BATs and ground textures would be troubling for most 2013 computers - especially those that use integrated video. It has developed nicely into somewhat of a "Simcity 5"

 

 

I sometimes imagine, what if Maxis devs divide the grids into even smaller grids; like say, a 8x8. While you can't develop anything with that gridsize, it should allow more realistic looking city scenes and the ability to make better curves without wasting too much space.

[AcronShinra 60]

 EA were to put the game into open source

It's EA you talk about; they'd rather resell their old games as overpriced DLCs or make'em freeware (like it happened with glorious C&C:RA, just before EA started their DRM madness) than open sourced it

[Kiiroitori 55]

Well perhaps you could say that the EA/Maxis trademark and the limitations of the hard coding gives the community some sort of organised standardisation in terms of modding and community. In other words simply things like how mods are installed and how they work and files required to install (.sc4bat, .sc4lot .sc4desc) is all the same with everyone and every game. Which I think is good, more in fact great because it gives the commnuity here some sort of organised structure and everyone can help each other out.

 

But perhaps the community could work together (Or the big people could plan it out and could ask if there is anyone that has any knowledge of C++ language) and we could tweak a few things (with experimenting of course and Alpha's, Beta labels) that wouldn't entirely affect this structure and improve the outdated graphics (such as adding actual movement to the water and shaders perhaps?), and if these changes are to hard to run on older PCs, then they could revert back to the original (or if there were options to turn down graphics or something) without affecting their ability of using custom content.

[jeffryfisher 2,176]

there's not very many of us with C++ skills, let alone at the level needed to pick apart a commercially-released simulation game...

 

I have the C++ knowledge (and O-O software design, project-management etc skills). Just convince Maxis/EA to hire me, and I'd be happy to patch/extend SC4. I'd also take care to either avoid clobbering the popular mods or else document changes well enough that the mod teams could version-up gracefully.

 

For starters, I want a variety of freight "commuter" types to cross city-sector boundaries in region play (e.g. I want food from farming sectors to show up in hungry residential sectors behind the connections).

[A Nonny Moose 8,261]

<snip>

 

For starters, I want a variety of freight "commuter" types to cross city-sector boundaries in region play (e.g. I want food from farming sectors to show up in hungry residential sectors behind the connections).

 

What you are really talking about is a full scale expansion pack into a different game that might be called SimCity 4 Freight.  Besides a thorough understanding of the existing code you would need to specify all this completely.

 

At the moment, freight is a distractor that is used to clutter the network.  What you are proposing is the notion of freight depots which have freight as destinations.  This would also give rise to large warehouse/distribution centres and a new automaton and transportation type called a delivery van for want of a better word.  These would originate at depots and end at commercial businesses of the CS type.  It means overhauling the path finder and the rest of the simulation to accommodate this.  Accounting for freight leads to an entirely different game from the present go-to-work aspect of the simulation. 

 

At that point you might as well make it a general Sim-ulation and take momma Sim shopping, the kids to school, and so on.  Stadium events could even be included including the flush out of the stadia at the end of the event with the resulting congestion.  In effect a general municipal simulation, including commuters moving through airports to move between city tiles.  This will need an order of magnitude change in processors speed, and the knee of the exponential curve will still move further to the left.

[glazert 89]

Now that 4K monitors (3840 x 2160 pixels) are becoming available, albeit at £3000 UK and also needing an even more expensive computer, it would be interesting to hear when someone gets the chance to try out SimCity 4 on such a set-up to find if it can still run at that resolution. At those prices it won't be me.

[A Nonny Moose 8,261]

At that resolution you'll need a magnifying glass to read the GUI.

[glazert 89]

At that resolution you'll need a magnifying glass to read the GUI.

 

The monitor I mentioned has a 31.5 inch diagonal, so no worse than my 1920*1080 laptop screen, though I accept that the GUI does look small, but readable, on that.

[TekindusT 3,809]

Back on topic, as soon as SC4 keeps being sold and turning profit, you can forget the idea of it becoming abandonware. I saw the game for sale last week for 10 €. And when the people who have embraced SC2013 want some more complez city-simulation stuff, they will turn their eyes to SC4 making the game to sell copies again. 

[SimCoug 5,973]

And when the people who have embraced SC2013 want some more complez city-simulation stuff, they will turn their eyes to SC4 making the game to sell copies again. 

 

That's been happening for a while - It's been constantly in the PC Game best sellers section at Amazon.com.  Currently it's at #36, right around other games like Diablo III, Skyrim and CivilizationV.  Not bad for a 10+ year old game   

 

...Oh yeah, and for those of you keeping score, SimCity(2013) is currently in position #57.  

[infamousjbe 593]

Back on topic, as soon as SC4 keeps being sold and turning profit, you can forget the idea of it becoming abandonware. I saw the game for sale last week for 10 €. And when the people who have embraced SC2013 want some more complez city-simulation stuff, they will turn their eyes to SC4 making the game to sell copies again. 

It's called brilliant marketing strategy. Only EA could make more money from an old game just from flopping the new one. BRILLIANT I TELL YA.....

[jeffryfisher 2,176]

<snip>

 

For starters, I want a variety of freight "commuter" types to cross city-sector boundaries in region play (e.g. I want food from farming sectors to show up in hungry residential sectors behind the connections).

What you are really talking about is a full scale expansion pack into a different game that might be called SimCity 4 Freight.

 

Maybe... Like the way Rush Hour was added to the even older SC4.

 

 

At the moment, freight is a distractor that is used to clutter the network.  What you are proposing is the notion of freight depots which have freight as destinations.

I was thinking that some factories could demand two or three classes of raw materials that are produced by farms or mines. Other factories could demand the meta-products produced by dirty industry. Certain CS lots would demand the outputs of the M and HT factories (as well as some food).

 

  This would also give rise to large warehouse/distribution centres

Warehousing would be optional but possible as a medium-density industrial lot.

 

and a new automaton and transportation type called a delivery van

I was thinking that the same trucks now hauling only to ports and map-edges could seek demands instead. Those exiting via a connection would emerge on the other side to continue the quest.

 

Accounting for freight leads to an entirely different game from the present go-to-work aspect of the simulation. 

To the extent that vehicles are spawned by one building to seek another, freight just becomes another kind of commuter. It's as if some "people" of a freight class "lived" at a farm or factory and "went to work" at another factory or retail outlet -- albeit at different times of day than morning commute.

 

What would change the game is if surplus or shortage of a freight class could cause businesses to abandon.

 

At that point you might as well make it a general Sim-ulation and take momma Sim shopping, the kids to school, and so on. Stadium events could even be included including the flush out of the stadia at the end of the event with the resulting congestion.

Those would be interesting twists -- adding traffic stresses at odd times of day (or days of the week, if SC4 has such a thing). Mayors would discover why evening rush "hour" is so much more congested than morning (it's because there are additional types of traffic in the evening, and some aren't as disciplined as daily morning commuters).

 

This will need an order of magnitude change in processors speed...

SC4's primary CPU load comes from graphics rendering, not operating the underlying model (except for the expensive pathfinding algorithm -- but that can probably be fixed^*), so I don't think that adding a handful of "commuter" classes and passing them between cities in region-play would trouble anyone's CPU. However, if more CPU power would be needed, then we can tap into the last 10 years' improvement since SC4 was released.

 

* It's probably not necessary for every commuter to know an exact path to a known destination at every moment. All that should be necessary most of the time is to calculate how many (not which ones) of each class are potentially available at any segment of the transit net. I have a theory that this can be done by modelling a transit net as an electric circuit. If I'm right, then an exact supply/demand and traffic-density solution could be calculated by a modern PC in a miniscule fraction of the time now used to calculate all of those paths. It would become practical to do the calc each and every sim-day instead of only quarterly. Exact paths would only be worked out when a player wants to see some of them.

[A Nonny Moose 8,261]

<snip>

 

* It's probably not necessary for every commuter to know an exact path to a known destination at every moment. All that should be necessary most of the time is to calculate how many (not which ones) of each class are potentially available at any segment of the transit net. I have a theory that this can be done by modelling a transit net as an electric circuit. If I'm right, then an exact supply/demand and traffic-density solution could be calculated by a modern PC in a miniscule fraction of the time now used to calculate all of those paths. It would become practical to do the calc each and every sim-day instead of only quarterly. Exact paths would only be worked out when a player wants to see some of them.

 

You mean like what EA did with their latest version?  Was a real success, wasn't it?  You still have to sort out your classes to get poppa Sim to his house.

 

Also, I was referring to current CPU power.  Take a look at discrete simulations of this type as the universe of discourse grows into the millions of objects.  You think this kind of simulation doesn't take up CPU cycles?  I think you've only made a superficial look at the problem.  I used to teach simulation, and my friend, with proper graphics handling, the rendering is trivial compared to the simulation.  Don't forget that the display is really for input, and the output is pure eye candy. 

 

Tera-hertz range machines will be needed for such a venture.  They are not as far off as some people might think.

[z1 566]

Now that 4K monitors (3840 x 2160 pixels) are becoming available, albeit at £3000 UK and also needing an even more expensive computer, it would be interesting to hear when someone gets the chance to try out SimCity 4 on such a set-up to find if it can still run at that resolution. At those prices it won't be me.

 

Seiki has a 39" monitor with 4K resolution selling at many places for $699, shipping included.  It's sold as low as $498 at Amazon and $549 at TigerDirect.  The size of the GUI elements is the same as that of a 19.5" monitor at 1920x1080.  As I have a laptop with an 18.1" screen at that resolution and I have no problem reading the text, the Seiki would actually be an improvement in size, and quite workable.  Imagine having the equivalent of four 19.5" monitors in front of you.

 

The next time this monitor drops back down to $500, it will be very tempting indeed.

[A Nonny Moose 8,261]

Considering that I will purchase a new monitor when this one dies in a year or so, I hope the price for really big monitors will be even lower by then.

[MandelSoft 2,770]

About rebuilding the traffic simulator, just having a course about it on the university, I think it's a very bad idea due to the increase of CPU power you need to handle this simulation.

 

If you're rebuilding the freight simulator as you proposed, you are rebuilding the entire assignment. If you want to do assignment right, you will need a lot of CPU power. While the current simulator in SimCity 4 can handle traffic simulations under half a minute (or a few minutes when it's having a hard time) due to their incremental assignment of traffic only generated by commuters, adding new trip purposes to the mix will just blow up the cycle time of the simulator, which is already one of the slowest parts in the game...

 

Furthermore, if you want to have a more advanced type of assignment (like a DUE or even a SUE assignment), this will add up to even more calculation time, time that SimCity 4 doesn't have when you are building your city. With an interactive game like SimCity, keeping the cycle time short is of the essence.

 

And I'm not even talking about the risk of a CTD caused by the game not being able to complete its cycle. Big, prop filled cities tend to be more unstable than small ones. Adding a heavy traffic simulator will only make that worse...

 

Best,

Maarten

[A Nonny Moose 8,261]

Well, I see nothing has changed since I was teaching that over 10 years ago.  Thanks for the insight Maarten.  You can only beat the poor horse for so long before it dies.

 

However, no program, however slowed down by volume should ever crash because of that.  In fact, no well written program should ever crash.  Programs that crash are not debugged and probably haven't had a full Quality Assurance cycle.  This is usually the result of deadlines.

 

In doing Q.A. on a program, the first couple of tests I always do are:

• No input.  An immediate EOT.
• Garbage input completely outside the specification of the program.

I had a reputation for poisonous fingers when it came to program testing.  You have to put yourself in the position of a user who knows nothing about the program.  Using the white box technique is usually a good approach.

[jeffryfisher 2,176]

<snip>

 

* It's probably not necessary for every commuter to know an exact path to a known destination at every moment. All that should be necessary most of the time is to calculate how many (not which ones) of each class are potentially available at any segment of the transit net. I have a theory that this can be done by modelling a transit net as an electric circuit. If I'm right, then an exact supply/demand and traffic-density solution could be calculated by a modern PC in a miniscule fraction of the time now used to calculate all of those paths. It would become practical to do the calc each and every sim-day instead of only quarterly. Exact paths would only be worked out when a player wants to see some of them.

 

You mean like what EA did with their latest version?

I am not familiar with the 2013 version.

 

Also, I was referring to current CPU power.  Take a look at discrete simulations of this type as the universe of discourse grows into the millions of objects.  You think this kind of simulation doesn't take up CPU cycles?

On the contrary, I am sensing that for the needs of the game, it is not necessary to solve all of those problems most of the time. As long as we can solve for the values really needed, we might save orders of magnitude on calculations.

 

I think you've only made a superficial look at the problem.  I used to teach simulation, and my friend, with proper graphics handling, the rendering is trivial compared to the simulation. 

Granted, I have only recently approached the problem. If there's a reason to calculate which exact job is filled by which exact sim, I haven't seen it yet.

 

However, I do understand that the pathfinding calc explodes as a factorial, especially in "grid" cities with lots of closed loops (I can theorize that "tree" structured road layouts will suck down fewer CPU cycles).

 

I still suspect that if the problems one usually needs to solve are "How many (not which) workers reach this factory?" and "How many (not which) vehicles use this road segment?", then it is not necessary to identify each one individually (any more than it is necessary to know which electron goes where when calculating the voltages across the various legs of a multipath electric circuit.

 

Freed from a whole level of detail, the complexity of calculation could be reduced by orders of magnitude. What I am suggesting is that the simulation is calculating way more information than it needs to produce the results that it needs most of the time. What's more, if my electronics analogy can be translated into an actual model, then it can be solved with off-the-shelf algorithms.

 

Of course, Maxis is unlikely to re-open development, so ours is almost certainly a purely academic debate. We're almost certainly stuck with what we have. Even so, I find academic debate stimulating, so if you'd like to explore my theory further, just ask me to spawn a new thread for it.

 

Don't forget that the display is really for input, and the output is pure eye candy. 

 

Natch

 

Tera-hertz range machines will be needed for such a venture.

Different venture.

 

As always, it is a pleasure "locking horns" with you Mr Anonymous... er "A Nonny Moose" 

[APSMS 1,893]

Interestingly, if you see the gameplay videos of early Sim City 2013 pathfinding, you will see the immense benefit of having jobs assigned to specific workers. In fact, the vanilla pathfinder of the game allows for a behavior that is similar to the kind seen in SC2013, but is still substantially better.

 

Now, while it is true that the game technically doesn't simulate individual sims, it does account for them, and specifically as well. Often Sims from one house (there may be a number of job-holders living in a single house) will continue to commute to the same job location provided their wealth and education stay relatively constant. The route occasionally changes based on traffic patterns and infrastructure updates, but usually the destination point will remain relatively stable (this is, of course from my own observations, which are not necessarily as scientific as some of the NAM team members', so I may be wrong on some of these points).

 

Now the an interesting point about SC4 is that it actually, on a simulation level, only calculates how many travel-types use a certain road, and how many workers are employed at a certain factory. The interesting thing, however, is that in order for it to know these things, it must also determine whether the Sim can get there, and if it can get there in time. Z has a lot of documentation on the importance of the traffic simulator, which I won't attempt to butcher here, but again, what I've observed in my own game is that traffic that has a specific start and end point, as well as the qualifier that traffic needs to use the fastest method possible to reach that destination requires a specific route to be known. While Sims are viewed by the game more as data points which fill certain requirements (e.g. health, ed level, wealth, etc.), they also have abit of pseudo individuality, which is what allows you to use the route query tool to determine with relative accuracy the job and commute of virtually every Sim. with obvious exceptions and limitations.

 

Now, while a number of these problems with regard to CPU time could be resolved with multi-threading, it does involve a large amount of source code editing and/or revising that would take an immense amount of time and effort, especially to be spent on a project that would be done without compensation. Multi-threading in impletation would be simple enough, but setting the game up to run this way (and then trying to support the project after it got released into the wild) would be a headache far worse.

 

Also, Nonny, I read an interesting post that pointed out that in the last ten years, available hard drive space has increased around 100 to 1000 times, 2 to 3 orders of magnitude. I have a computer still that has a Pentium II with a 5Gb hard drive. My laptop currently has 100 times this and I could potentially put even more space in. Floppy A disks were 1.44 Mb, and they now make thumb drives with almost 144 Gb, which is 5 orders of magnitude in around the same time.

 

The speed of that particular Pentium 2, however, is around 333 to 500 MHz (I haven't booted it up in years, since the HDD is out of space and won't let even MS Word run) in terms of cycle speed. My current laptop? 2.4 GHz base and 3.4 GHz tops, which is an improvement of a single order of magnitude, at best. Even supercomputers use standard Intel and AMD chips, which mean that their individual speed tops around 5GHz for a well maintained and cooled unit. Their primary benefit is from multi-threading (if I'm not mistaken; of course, I'm saying this all to a veteran computer programmer, so please understand that I say this with all due respect; I could be easily out of my league here, even if I like to think that I'm not), which is easier to come by when you have somewhere around 200 hyperthreaded cores.

[A Nonny Moose 8,261]

^ Quite an interesting article.

 

Just looking at the physics of current CPU cores shows that they've just about reached the limit on this technology.  A quantum leap in circuitry is needed to get more oomph out of processors.  The introduction of multiple cores and multitasking was Intel's answer to go faster because the speed depends on component coupling problems when you decrease the spacing very much more.

 

There is work being done on a 'quantum' computer which is supposedly faster.  Funny if it turned out to be slower.

 

I am aware of some classified work that was being done on photonic circuits using photons instead of electrons, but that work seems to have petered out.  This was over 10 years ago, and nothing has shown up.  Things start getting rather relativistic when you try for things operating at light speed so the power curve may have killed this one.

 

I think that digital processing will eventually give way to massive parallel analogue processing simply because it is the best we have found so far.  The operating system of the human brain is this way, and no computer I've seen so far comes even close.  It isn't particularly fast, but it is massively parallel.  So expect the next leap in processors to be on the side of the 'positronic' brain of Isaac Asimov (who was a biochemist by the way).

[jeffryfisher 2,176]

Hmmm... If exact paths have more value than I've seen thus far, then opportunities for optimizing the pathfinding programming are more limited. However, I still have some ideas.

 

For one thing, calculating each Sim in a "vacuum" means that many calculations are repeated again and again as similar paths encounter the same grid. If memory is cheap, then I would first analyze the grid, creating a tables for the fastest way to go from one node to another via each transit net. Then when each sim is being analyzed, many of the factorial explosions would already have been reduced to table look-ups that can be used again and again.

[z1 566]

For one thing, calculating each Sim in a "vacuum" means that many calculations are repeated again and again as similar paths encounter the same grid. If memory is cheap, then I would first analyze the grid, creating a tables for the fastest way to go from one node to another via each transit net. Then when each sim is being analyzed, many of the factorial explosions would already have been reduced to table look-ups that can be used again and again.

 

I believe that the traffic simulator already does something very much like this.  From what I recall of what Maxis said, there's a limitation on what is considering a "node" by the traffic simulator, not only for memory reasons, but for reasons of processing time also.  The main point is to get the main routes calculated and stored, so they don't have to be recalculated for every Sim that takes them.

[A Nonny Moose 8,261]

Quite right, I believe.  Hash tables are best used for this kind of work.  You only have to recalculate if some feature of a route is changed.  For simplicity's sake, I generally try to constrain routes to work once the traffic is clear of the residential maze.  Generally a main drag associated with the residential pods does the trick for main routing.

[z1 566]

 

Now that 4K monitors (3840 x 2160 pixels) are becoming available, albeit at £3000 UK and also needing an even more expensive computer, it would be interesting to hear when someone gets the chance to try out SimCity 4 on such a set-up to find if it can still run at that resolution. At those prices it won't be me.

 

Seiki has a 39" monitor with 4K resolution selling at many places for $699, shipping included.  It's sold as low as $498 at Amazon and $549 at TigerDirect.  The size of the GUI elements is the same as that of a 19.5" monitor at 1920x1080.  As I have a laptop with an 18.1" screen at that resolution and I have no problem reading the text, the Seiki would actually be an improvement in size, and quite workable.  Imagine having the equivalent of four 19.5" monitors in front of you.

 

The next time this monitor drops back down to $500, it will be very tempting indeed.

 

 

Since I wrote this, this monitor has dropped from $699 at Amazon to $485, an all-time low for any dealer.  Shipping is free to the continental U.S. There's a price war going on over this monitor now in $5 increments between Amazon, Sears, and TigerDirect, so I think that the dealer's cost can't be lower than $450 or $460, including shipping costs.  This would seem to be the monitor to get for SC4.