Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[nil]

Actually, that shows what I was talking about fairly well - the problem isn't so much that there are houses out there, but the problem is the roads.

The roads in the Toady mockup are a grid, as if the city was already being planned out over in the fields. 

The roads in the map just stretch past the small clusters of farmhouses and head directly to town.  It looks more like a spiderweb, with roads radiating out fo a central point and a cluster of cross-hatched roads in the middle, and then there is just a couple ring roads out beyond that one.

Might the easiest was to solve this be to have some sort of mechanism that identifies whether or not roads would be used, and destroy them if they are not?  Legit question, I'm no programmer.

[monk12]

The problem is iirc. that the temperature of the magma is uniform thus each tile has the same temperature. This would lead to entire layers melting.

Easy peasy.  Don't have igneous layer stone that isn't magma safe, but ore could very well be.  If there is a layer stone that isn't magma safe yet is spawned near magma, it shouldn't melt when you start the map.  Rather, it should never generate in that position in the first place!  Simple enough to cull during world-gen, I should think.

The problem arising when you pump the magma to the surface and pour it out onto the battlefield, where it begins melting its way back down to the mantle.

Magma should not do this. Think about it. Real world magma is gradually losing heat to the surrounding rock as it approaches the surface. Yes, it melts some rock, but not much--It does far more pushing than actual melting. Unless you want your magma to gradual cool and harden into pumice and lava rock, don't ask for it to melt the walls (even if it did, it wouldn't happen instantly. Melting solid rock takes a much longer time than melting a tinfoil hat. less exposed surface area.)

Exactly the point I was making- that its not something easy to just slap in because it requires extra resources to be realistic. To get an accurate, workable model for magma you need to model heat loss, which means modeling the heat of specific parts of magma, which is not something the current fluid simulation is terribly equipped to do, at least not efficiently.

Actually, that shows what I was talking about fairly well - the problem isn't so much that there are houses out there, but the problem is the roads.

The roads in the Toady mockup are a grid, as if the city was already being planned out over in the fields. 

The roads in the map just stretch past the small clusters of farmhouses and head directly to town.  It looks more like a spiderweb, with roads radiating out fo a central point and a cluster of cross-hatched roads in the middle, and then there is just a couple ring roads out beyond that one.

Might the easiest was to solve this be to have some sort of mechanism that identifies whether or not roads would be used, and destroy them if they are not?  Legit question, I'm no programmer.

That actually sounds fairly reasonable- set up the network of roads to provide many different paths, then simulate traffic from each homestead to the city gates- it has the added bonus of not only dropping unused roads, but expanding well traveled roads, and then there will be just the little dirt wagon ruts out by the farms to show where they drive every week to get to market. Presumably only the major highways built in worldgen will be paved- all of this helps with navigation as well, since the player knows the larger roads lead to population centers and the smaller roads to farms.

[NW_Kohaku]

I'm imagining that, at present, cities claim a certain region of land for "this is my site" early on.  External fields are new sites, added as needed.  Hence the preplanned look of that city; it has to be, because land is claimed in advance.

Once sites can expand over time, I'm sure cities will do better there.

That said, it's almost unimaginable to me that cities would not be generated during worldgen, but would instead be retroactively created later.  That's asking for trouble, because it means a city you've visited as an adventurer will likely develop in a different way over time (once that's implemented) than one you have never visited.

Remember that it's ideal that people will create a world and then play LOTS AND LOTS in that one world, without needing to gen a new one.

Unfortunately, the game is simply not very well equipped for playing in the same world over and over again right now.  In adventurer mode right now, pretty much the only activity is to kill everything in the world.  Yeah, sure, you, as an adventurer will get more things, but the basic problem with the rest of the world is still the same.

Something I said in another thread:

As it stands right now, as I remember it, one of Toady's games as an adventurer involved him running away from an alligator to go visit the king of a castle, and get a quest, and when he came back from the quest, the king had been eaten by the alligator, so he recruited all the guards he could from the castle, and went looking for a new castle to get his quests from, because there would never be another king in that castle again.  Once something dies, the function it fulfilled is gone forever.  The game is built on a massive amount of entropic decay.  Societies cannot function, they can only collapse.

...That is really dark.
I am so glad that that's not what the real world is like.

In the real world, behavior like that in DF's AI would never have led to civilization in the first place.  It has its ups and downs, but real-life cultures have always generally striven to build and advance.  DF cultures do not.  They cannot advance, there is no "upward pressure" to innovate or build, and so the "downward pressures" of destruction are completely unopposed, leading to inevitable global extinction.

Killing every living thing (besides vermin) in the entire region is not something epic in this game, it's just hastening the inevitable, completely unopposed entropic decay of a DF world.

In trying to set up a small world for a game I'm playing in right now, I had to cut back the amount of years spent in history to 120 years, just so that dwarves would survive worldgen.  They keep starving by year 27, but for a select few.  Those starve at year 200.

When you haven't visited a town as an adventurer yet, everything is cellophane-wrapped and pristine, but it only is a matter of time after you visit before everything collapses, because the random alligator you walked past will eat the king the next time you come by, and nobody will crown a new king, nobody will build anything or replace anything that gets broken, but everything can break. 

That, in a nutshell, is the overarching problem with the simulated fantasy world right now - nothing can be built, but everything can be destroyed. 

Unopposed entropic decay.

[Sowelu]

Still.  Just because that happens now, doesn't mean it will happen in the future.  Why write a fancy-pants retroactive town simulator when that entropic decay won't be an issue in the future?

[NW_Kohaku]

Still.  Just because that happens now, doesn't mean it will happen in the future.  Why write a fancy-pants retroactive town simulator when that entropic decay won't be an issue in the future?

Actually, I kind of split the response to your previous post in two, making the last post a complete post unto itself in order to help underline the problem by not letting it get mixed up with the alternatives.

I'm imagining that, at present, cities claim a certain region of land for "this is my site" early on.  External fields are new sites, added as needed.  Hence the preplanned look of that city; it has to be, because land is claimed in advance.

Once sites can expand over time, I'm sure cities will do better there.

That said, it's almost unimaginable to me that cities would not be generated during worldgen, but would instead be retroactively created later.  That's asking for trouble, because it means a city you've visited as an adventurer will likely develop in a different way over time (once that's implemented) than one you have never visited.

Remember that it's ideal that people will create a world and then play LOTS AND LOTS in that one world, without needing to gen a new one.

The reason why this is, however, is that it MASSIVELY cuts down on save bloat.  The save file of a world doubles in size on your hard drive the instant that you actually embark or visit a large site.  Just an empty 3x3 embark takes up several megs of space. 

When you play a game like, say, Oblivion (an example on my mind because of all this talk about Daggerfall) then you have all the unnamed characters and random caves simply reset every three days if you haven't actually looked at them.  (Meaning that a cave full of bandits respawns, and all its treasures can be looted again, and all its bandits can be killed again, but anything you left there has now disappeared.)  On the other hand, named characters are dead forever if they happen to be killed by a crab somewhere offscreen without you knowing about it, and any quest related to them can never be completed because of their death. 

To an extent, something like that has to be used just to prevent the fact that the player can mess with so many chests, random NPCs, and simply kick over so many chairs or push books off of counters whose positions need to be recorded individually that the save files would become even more horrifically bloated than they already are without them.  (Plus, the gameworld is small enough that you would eventually depopulate a pretty wide swath of the world.)

There have to be two ways of doing things -

• Players who don't visit a city for long enough might simply have that city removed from the saves altogether, and go back to being a sort of Schroedinger's Box after a set amount of time, like how Oblivion does it, but where the town can be built up or stripped down from its "pristine" condition if time passes.  This assumes some sort of period of time that passes will basically make whatever changes the player performed on the town glossed over by whoever visited the town afterwards.  If half the population was killed by the adventurer, then the rest of the survivors would just be assumed to have deconstructed old buildings for materials, and rebuilt the town in a huddled-together mass.
• Players who stay in one city for long enough are basically going to actually SEE changes to the city taking place.  If an adventurer sticks around a city for ten game years, then odds are, the player's going to have to witness some births to go along with all those deaths.  The player's going to need to see the dead king's heir take the throne.  The player's going to watch the new wall slowly be built up day after day.

This wouldn't be a problem in the case that the player actually spends the overwhelming majority of their time in one specific city (in which case, only that one city only has to be closely tracked), but the game has to be smart enough to recognize when it is one case and not the other.

The other thing is that towns being put back in Schroedinger's Box need to at least look superficially similar to what it looked the last time.  Some sort of seed or guidelines for construction does still need to be retained for when the city is rebuilt.

[Mel_Vixen]

The thing with forts is that you need the every bit of data in a well defined state. Towns on the other hand can be in very big parts in limbo.

I guess, you could give every "building" sub-seeds (for layout, inhabitants etc.) to break the town down into handleable chunks. Messing up one building wouldnt interfere with the entire town and killing half the population would just mean that you set a "not used" flag for the corresponding buildings.

Sure you would get an overhead but it beats saving everything. Any changes the player made gets saved in one or multiple "lists".

Also you could set up a (literal) clean-up routines by the population that lives in the place so that minor changes get gradualy removed. That and stuff geting destroyed by erosion and decomposing respective the building repopulated by new people etc. (depending on what happend) thus reseting a building to some predefined state (out of who knows many). After that you could remove stuff from said lists that became irrelevant.

Minor stuff like a book left on the table, or footprints left on the floor should expire within a day or two if the building is used. Food left out should be gone within a Month or two.   

A town is fuzzy for lack of words. To any time its just a seed with a list of events attached that determine the growth and changes. If it appears "on screen" you go from planting the intial stuff over all the tiny historical things like wars. Presto your town in the pre-adventurer state. Then you consult the formaly mentioned list to apply the player made changes and that are still around. Its like an incremental backup, i would say.   

[G-Flex]

To an extent, something like that has to be used just to prevent the fact that the player can mess with so many chests, random NPCs, and simply kick over so many chairs or push books off of counters whose positions need to be recorded individually that the save files would become even more horrifically bloated than they already are without them.  (Plus, the gameworld is small enough that you would eventually depopulate a pretty wide swath of the world.)

Then what about, say, Morrowind? You could store anything pretty much anywhere and it would stay put.

[NW_Kohaku]

To an extent, something like that has to be used just to prevent the fact that the player can mess with so many chests, random NPCs, and simply kick over so many chairs or push books off of counters whose positions need to be recorded individually that the save files would become even more horrifically bloated than they already are without them.  (Plus, the gameworld is small enough that you would eventually depopulate a pretty wide swath of the world.)

Then what about, say, Morrowind? You could store anything pretty much anywhere and it would stay put.

Morrowind was also problematic - there was a major bug that meant the game would be unplayable after certain periods of time because certain characters would gradually "shift" in their positions over time, eventually moving through walls in interior cells where they would become inaccessible.  See: http://www.uesp.net/wiki/Morrowind:Glitches#Moving_NPCs

Also, there are problems with that game's saves - go too long without saving and reloading, and the game crashes because of a memory overflow.  Also, if you save too many times, the game starts having load errors.  See: http://www.uesp.net/wiki/Tes3Mod:Engine_Bugs#Savegame-related_bugs

Beyond that, Oblivion simply was a game with more things to save, IIRC.  It used the game engine that let it throw random bits and pieces of objects like carrots and books and cups all over the floor.  Tracking the facing and location of each individual radish and fork on random farmer's dining room table adds up to a lot of data in a larger gameworld than Morrowind was tracking.

Plus, I think it was a bit of a design decision to let you hunt vampires in the same cave again and again if you just kept playing the game - you get to keep playing even after you have finished the main quest and every side-quest.  At that point, modding and adventuring back into the same bunch of vampire caves are pretty much all the game you have left.  If all you want is a fight and a crack at some loot, you can do those until the end of time if you just go in a circuit of dungeons every three days.

Plus, it's kind of required for the whole rubberbanding the level of the enemies and loot thing they were doing in Oblivion, although that has no real bearing on DF, obviously.

[G-Flex]

Morrowind was also problematic - there was a major bug that meant the game would be unplayable after certain periods of time because certain characters would gradually "shift" in their positions over time, eventually moving through walls in interior cells where they would become inaccessible.  See: http://www.uesp.net/wiki/Morrowind:Glitches#Moving_NPCs

Haha, I remember this. Not sure why that's a necessary aspect of the system, though... sounds like just a weirdass solvable bug to me.

Also, there are problems with that game's saves - go too long without saving and reloading, and the game crashes because of a memory overflow.  Also, if you save too many times, the game starts having load errors.  See: http://www.uesp.net/wiki/Tes3Mod:Engine_Bugs#Savegame-related_bugs

Also sounds solvable; it's a buffer overflow problem.

Beyond that, Oblivion simply was a game with more things to save, IIRC.  It used the game engine that let it throw random bits and pieces of objects like carrots and books and cups all over the floor.  Tracking the facing and location of each individual radish and fork on random farmer's dining room table adds up to a lot of data in a larger gameworld than Morrowind was tracking.

You could move just as much stuff in Morrowind, with facing and orientation and stuff... of course, there was probably still a lot less granularity, since everything was still plastered to surfaces.

[Jiri Petru]

Btw. as this map shows the these lone buildings out in the fields did exist. If you compare that to the two maps i linked earlyer you can see how the city grows and that toadys estimate isnt that wrong. I think the problem with toadys map is that the divide between town and farmland is a bit to hard.

Actually, the map you have posted is from 1750, which is way too modern. I doubt those outlying farms were present in the middle ages. What you see are most probably "country manors" of rich burghers (most probably traders, guildmasters, lower nobility etc.) who grew tired of the filth of the city and moved outside. They could afford carriages to go back to the town on a daily basis, and generally were nothing like medieval peasants (notice the fancy gardens). You can even see that all the roads are in fact avenues with trees which means they're made for rich people. Before the modern ages this "middle class" didn't really exist and these avenues and country houses couldn't exist. In fact the emergence of the middle class is what started the modern era and made it so different (and turbulent).

Here's Görlitz in 1575. I couldn't find an older map. You can't see much of the countryside though. But I'd say there would be normal villages around, not suburbia.

By the way, ancient Romans did have suburbs of sorts - but again it was rich people building their villas out in the coutryside. The point here is that if you generate suburbs like these, it should be manors of rich people, not cottages of poor peasants. Peasants live in villages.

EDIT: By 1650, the suburbs had only just started appearing.

[LoSboccacc]

you should not take those drawings at face value, those were figurative, not informative

edit:

the same suburban area on the north of the latest map is to be found on the left of the previous drawing (it's quite hard to differentiate between trees and roofs)

[Mel_Vixen]

Well "way" to modern i wouldnt say. The 1750s may be in building style different but the way the land is settled is still the same because there werent any mass transportation systems. That map i did post because it shows how the town developed over time (compared to the 2 earlyer maps i posted) and i atleast think the underlying principle of town development still holds up until the industrialisation begins.

The "map" of 1575 actually is more like a elaborated postcard i would say and is "shot" from the (now) polish side of the river, didnt change much btw. thought we have a new bridge and the cityside gate doesnt exists anymore.

You make a point with the manors but these places didnt exist in a vacuum and they did exist since before the 1400s in the case of nearby Königshain for example (1298 created as Hunting lodge for the Kings wife). We had a healthy population of lower nobles around here too in fact many of the surrounding villages were originally Knight-manors.

That alleys and gardens you see are a symptom of the barock and i would say many of the manors etc. existed before. I give you thought that the density might have increased to a point that wasnt possible in the 1400s. 

edit: and before you say something the bastion in the west of the city (that rondel that sticks out from the citywall) was etablisched 1427 and got modernised over the time (as did the Gate-tower) so even if it looks barockish its foundations are "old". Just saying.

edit2: The oldest "map"/view of the town i have found online so far is from 1565 ... well maybe i should walk over into the museum for older stuff.

[Jiri Petru]

No, my point wasn't a specific picture or city or whatever, the point was that the suburbs aren't peasants' farms but rich guys' houses - and these of course are in a limited scope possible even in DF's timeframe (Romans had them too). In fact, Toady wants to add manors and it would be cool to have them in the countryside, but he hasn't done it yet so that means his current suburbs are peasant cottages. That's what I was saying.

[Mel_Vixen]

Ok a little misunderstanding then. Well i can see if i get into the town archiv in the next days for to fiddle the farmer/'rich guy' ratio out. Hehe well maybe not for that but for my own interest. By the way there is a archaeological estimate for the city plan (only the city) of the 13th century. The interresting thing is that the course of the "via regia" itself was changed to use the new bridge a bit south of the old ford.

Also for the sizecomparsion: Between the 1415 and the late 1700s the city maintained between 5000 (around the 1500s due to the plague and various other stuff)  to 8000 people. If you guys can set that in rlation to toadys maps we could estimate how big the populations should be to be realistic. 

  i think we have gone a bit astray from the original discussion.

[de5me7]

i find the debate on what information to retain for saves/entropy etc interesting.

The entropy situation has always been a major issue for me in certainly adventure mode (atm due to limited interaction its less noticable in fortmode). It really kills the persistent world aspect if it cannot maintain its self.

Ive always assumed that the game would just (or does) just save the significant information. Certainly for sites you havent visited it does. For sites you do visit i dont see it as a problem, if youve dropped something and next time you go to site X it isnt there (unless its a very signficiant item such as an artifact, or if its in a container that is somehow tagged as significant). In a city aslong as the important characters are tracked at a detailed level i dont think tracking the pesants matters much unless they become part of the story and there by important. If i visited a city twice, and found different named people living in different houses it wouldnt bother me. But i would expect main characters to be in their correct place. As for destruction of property and rebuilding of property i guess you could modulalise it or something. So when a city map is genned, the game could gen several alternate states for said city. It could gen 4 size expansion maps, and 4 ruined versions of the city (far more than this could be genned, this is simply an example). WHen a city goes past a threshold in population or wealth the next phase could simply be loaded from these original gens. This would be very un organic but it would be simple and presumably take less space than a truely organic system, where the state of every building/item is tracked. You could add detail to such a system by giving certain structures within the city several versions which they could move between depending on whether they had been upgraded, or sacked etc.

Can new cities be created after world gen in the current version? Perhaps the game could do a simulation update at the end of every X number of years and add new settlements if some prequisits were fulfilled. Presumably updating the simulation is what is needed for kings to be replaced etc, and entropy to be defeated.