Print

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

[MrWiggles]

I like the idea of auto support construction for the mine shafts. Though I see it only working in a few applications.

In a single z level enclosed shaft. I dont see how it can work for multiple z level open areas though. As in the single z level area, that can be taken care of without annoyance to the player, being invisible supports more or less.

[Quatch]

Look, whatever system is used for determining the cave-ins is irrelevant (how complex it is, and I favour complex, but for computation reasons, a simpler method is probably fair. And to computations, this is something that would only have to be calculated when digging occurs. You can precalculate the map at embark).

What we need is a good way of informing the player of the threat. I propose the following:

1) Inspecting a tile with k shows a stress value, eg, a percent. When that gets low, you know it is going to fail. You don't care how it gets low, but you know it is related to removing surrounding rock, so when it does, you stop.

2) At a certain threshold the tile should visually notify you that it is weakened. Much like with hot or damp rock, this should flash when you are designating. A difficulty here is when you are digging on a layer above or below, so you might need 2 kinds of flash (to make things easy on the player): (a)Weakened[will collapse if support removed]. (b)Supporting [removing this tile will lead to a collapse elsewhere].

3) At another threshold (larger than the first), you receive digging cancellations based on removing support.

4) Prior to collapse, you receive a paused notification indicating that collapse will occur at this site, and dwarves should now flee.

5) Collapse occurs.

I think this level of graduated response will train the player to the mechanics, rather than force the player to understand the algorithm prior to digging. For kicks, you could put the 2 thresholds in the init, or let the second threshold vary based on the skill of the miner who is about to dig.

Of course, the real complexity is when enemies dig.

(irritating legal note: this idea is in the public domain)

[sweitx]

Look, whatever system is used for determining the cave-ins is irrelevant (how complex it is, and I favour complex, but for computation reasons, a simpler method is probably fair. And to computations, this is something that would only have to be calculated when digging occurs. You can precalculate the map at embark).

What we need is a good way of informing the player of the threat. I propose the following:

1) Inspecting a tile with k shows a stress value, eg, a percent. When that gets low, you know it is going to fail. You don't care how it gets low, but you know it is related to removing surrounding rock, so when it does, you stop.

2) At a certain threshold the tile should visually notify you that it is weakened. Much like with hot or damp rock, this should flash when you are designating. A difficulty here is when you are digging on a layer above or below, so you might need 2 kinds of flash (to make things easy on the player): (a)Weakened[will collapse if support removed]. (b)Supporting [removing this tile will lead to a collapse elsewhere].

3) At another threshold (larger than the first), you receive digging cancellations based on removing support.

4) Prior to collapse, you receive a paused notification indicating that collapse will occur at this site, and dwarves should now flee.

5) Collapse occurs.

I think this level of graduated response will train the player to the mechanics, rather than force the player to understand the algorithm prior to digging. For kicks, you could put the 2 thresholds in the init, or let the second threshold vary based on the skill of the miner who is about to dig.

Of course, the real complexity is when enemies dig.

(irritating legal note: this idea is in the public domain)

The difficult part is HOW do you calculate a stress value?
There're 3 main type of stress (technically more, but for DF I only see 3 possible).
Compression, Sheer, and Tension.
The main difficulty is to compute how much each material actually support.
A starting point can be this.
1. Given a N x N room and M supports, compute the amount of compression stress on each support.
2. Compute the amount of sheer stress on each floor.
Add a z-dimension above, things gets complicated very quickly.

[azrael4h]

Before adding 'thin' supports, as mentioned (which would require tons more wood), the trees need to be rewritten into a form of creature. This way, when you cut them down (butcher them) you get more or less amounts of wood from each tree. A modder could probably make something similar work now, but they wouldn't be working as trees until Toady coded it in.

Before this happens, we need to be able to stack more than 1 unit of wood (or stone) per tile in a stock pile. Which, in turn, will eliminate the need for massive stockpiles to support the charcoal and coke making processes. Maybe cap each tile to 50-100 units of wood, and 25-50 units of stone/raw ore.

Then add in cave ins all you want. I like the 2D implementation; too many tiles of rock unsupported would result in a cave in being imminent. Fewer tiles, but same rules for soil. Each soil/rock type can use different hardness or whatever to determine how large an area can be built unsupported.

I already build my great halls and larger chambers with columns designated for support, even though it's unnecessary. In fact, I'm planning a fort with two 3x3 columns acting as vertical pipes for the waterworks' upper lines. One for a feed line into the lower waterworks, one a return line to the upper waterworks. This is because of a few things that the waterworks will be doing.

[MrWiggles]

The 2d cave in model though, as Toady said can't be done the 3d version.

[Quatch]

Look, whatever system is used for determining the cave-ins is irrelevant (how complex it is, and I favour complex, but for computation reasons, a simpler method is probably fair. And to computations, this is something that would only have to be calculated when digging occurs. You can precalculate the map at embark).

The difficult part is HOW do you calculate a stress value?
There're 3 main type of stress (technically more, but for DF I only see 3 possible).
Compression, Sheer, and Tension.
The main difficulty is to compute how much each material actually support.
A starting point can be this.
1. Given a N x N room and M supports, compute the amount of compression stress on each support.
2. Compute the amount of sheer stress on each floor.
Add a z-dimension above, things gets complicated very quickly.

I disagree. There are obviously two completly separate problems: an efficent algorith, and an intuitive/simple display. I think starting with a simplified algorithm (ie what we have now, or something one step more realistic) and adding more with time and the avaliabilty of algorithms is sufficient, but the lack of a way to inform the player without overwhelming is the current bottleneck.

[LordDemon]

I disagree. There are obviously two completly separate problems: an efficent algorith, and an intuitive/simple display. I think starting with a simplified algorithm (ie what we have now, or something one step more realistic) and adding more with time and the avaliabilty of algorithms is sufficient, but the lack of a way to inform the player without overwhelming is the current bottleneck.

How about something like the depot access view (for general information) combined with the hot/damp stone warnings for crucial supports. Maybe even use multiple colors in the separate view, that should give player a good overview of the stresses in their fortress. And if they forget to look at it, the warnings would give newcomers some leeway by telling them they are about to bury themselves under tons of rock.

[Kogan Loloklam]

When Toady discussed this, he didn't say anything about computational power or difficulty coding. His one and only concern for implemanting this (according to Dwarf Talk #8) was the players. If he could make it so players can learn this without great difficulty, then it would be something he could consider doing.

That is in response to everyone who keeps saying the mantra of "efficient algorithm" or "coding time"

I'm not saying those concerns aren't there, just that to Toady, they aren't obstacles as to why we won't see a more impressive cave-in system eventually.

I could see "stressed" kind of warnings similar to a Depot access map working pretty well though. Most people understand stresses better than they think (IE people get it pretty close to intuitive if it has some element of realism), and a simple graphical representation of stresses could easily resolve the issue of people not grasping it. Actually something similar was done in my structural engineering class. A video was shown of I think it was a model of Notre Dame. The model was put into an oven, and when it came out different colored light was reflected from portions of it depending on the kinds of stresses that were acting on it. Very few people didn't get it after that (There was one lady from L.A., but I discount her because she couldn't understand why we built foundations the way we did out here. Anyone who gets that far in an engineering degree and doesn't understand that ground that freezes needs different foundations than ground that shakes isn't playing with a full deck. In hindsight, it was probably the fact that this wasn't intuitive to her as to why she could never figure that out.)

[BloodBeard]

I'm a big fan of realistic cave-ins because my style of play revolves around interesting architecture. It would add a whole new challenge to megaprojects and fortress design.

I've pondered about how it can be implimented, and it's pretty hard to explain so i'll try and show a a graphic:

This is a side view of a mountain with a cavern dug out. Assume there are only the 2 sides to the mountain and it's it's only 1 tile wide. It's just easier that way:

############
############
############
############
############
############
############
############
############
###______###
##________##
##________##
@@@@@@@@@@@@

# = rock
_ = empty space (not floors)
@ = ground with infinite load limit.

So, my vision is that each mineral/soil type is assigned a number. This number determines how strong it is/how much load it can withstand without collapsing. Should it reach 0 or below, it collapses.

Lets assign # to a strength of 8 and take a look at the mountain again:

888888888888
888887788888
888886688888
888875578888
888864468888
888753357888
888642246888
887531135788
886420024688
885______588
88________88
88________88
@@@@@@@@@@@@

The rock maintains a strength of 8 when it's touching solid ground, with all rock directly above maintaining full strength as well. However, if a rock wall has an empty space below it, it's load limit equals the load limit of the rock beside it minus one. It furthur reduces if there is a rock load above it. Load limit is then transfered vertically, where if a rock wall has a load limit of 5, the rock above it has a load limit of 6 and so on untill it's at it's maximum. The updated graphic shows the strength of each rock relative to each other due to the empty cavern.

Notice that two rocks have reached 0. They would then collapse, and result in the mountain looking like this:

888888888888
888887788888
888886688888
888875578888
888864468888
888753357888
888642246888
887531135788
88642__24688
885______588
88________88
88___88___88
@@@@@@@@@@@@

Sorry if I lost anyone, but yeah, that's a rough version of my vision. Here's another graphic where an outside archway is only barely standing:

~~~~~~~~~~~~~~~
~~~543212345~~~
~~64~~~~~~~46~~
~85~~~~~~~~~58~
~8~~~~~~~~~~~8~
@@@@@@@@@@@@@@@

(~ is sky)

Now imagine how this kind of feature can be implimented with different materials. Sand would have 0 strength while metals have extraordinary strength. 

[Retro]

I can understand others' arguments about it being more of a challenge and yet another Fun thing for players to have to watch out for, even from those who focus on fortress architecture. However, the idea of realistic cave-ins completely makes fantasy architecture - the kind of architecture I enjoy - completely impossible. One of my favourite - scratch that, my absolute favourite part of the game is being able to create highly stylized structures otherwise found only in paintings, animation, and computer-generated imagery. If I want to carve a giant cavern, I don't want to have to worry about it being chock full of pillars from head to toe; if I want to create a winding passageway up a spire without ground beneath the suspended walkway, I don't want it to collapse halfway through; if I want to dig away at a mountain, I don't want it fall down and crush my miners.

Towers and aboveground structures are another matter entirely - how would one even build a tower using any sort of realistic gravity? Trying to build to any significant height would be out of the question, especially as DF does not understand foundations nor can suspension wire be tied between a tower and the ground (even trying to fake one out of constructed rock would cause a cave-in partway up). And what of the interface? Building structures or digging caverns with any amount of open space greater than perhaps 3 z-levels becomes exceedingly difficult to keep track of with a 2D view of a 3D world regardless of annoying flashing displays. I hate to bring my own work into a debate, but I have dug a cavern nearly 50z tall with an asymmetrical naturally-curving variably-high ceiling spanning an area of around 50x100. Even if I could hypothetically dig such a thing out with cave-in realism it would be impossible to keep track of what areas of the ceiling or walls were strong/weak, even with dedicating every single one of my hotkeys to different parts of it.

Furthermore, we the players would have to learn everything about every single construction and building in regards to support. If I dug out a 100x100x100 square chamber using only up/down stairs, would it support itself? What about fortifications? Or even doors and support pillars? Constructed vs. natural walls? Would floors provide any support? And gods forbid having to learn the individual weight of every building and construction in the game (let alone material weights) to know whether or not you're allowed to build another wall atop your fortress barricade. Even creating a large room with a vaulted ceiling for support would be impossible, whether dug-out or constructed - it would work upon completion, but getting it completed really would be impossible to do - it'd collapse as you built it up before you could connect the curves, and to dig out a vaulted room you'd have to start at the top and dig down, adjusting the curve with every new level you dug (and how would one even get back up if you needed to adjust the curve once you'd dug down three more layers?). I dread having to learn the logic behind what grade curves create a support-giving vaulted ceiling, and above all I dread having to teach this to new players. Dwarf Fortress should not be restricted only to structural engineers and architects.

The option to turn off cave-ins altogether is similarly unfavourable. I want things hanging unsuspended in the air to fall down. I want to have to be careful when digging. However I above all want to be able to dig. If this feature was ever to be implemented, I would require a third toggle to have the current fantasy-style cave-in system currently implemented, and I don't know if Toady would be able to completely rewrite how gravity and support work while keeping the current attachment-is-support system and having unattached land/structures still fall.

I really do understand the desire among some players from realism in this aspect of gameplay, but it derives from my ability to do what I want to do. It's true that digging right now is 'too easy' and has few risks, but unlike other similar complaints such as farming being too easy, mining can not be made more difficult while keeping all its abilities. If farming becomes more difficult, I can still farm. If mining becomes more difficult, I can't dig, and the casual tone with which some people appear to be considering the matter very deeply worries me.

I see realistic cave-ins as a step backwards in DF, deeply taking away from the formerly vast field of architectural possibilities a player can create. The day that Dwarf Fortress starts imposing limits on me as a player is the day I stop playing it.

[Jordix]

I'd love to see a realistic method like that implemented, but the processing power needed would probably make the game unplayable.

Here's a very simplified alternative: for every unsupported "floor" tile (either a constructed floor, with nothing above, or the ceiling of a cave, with rock above), the game checks for a supported tile in up to, say, 5 tiles in every direction. If there's none, it colapses. Maybe an announement if there's one just right at the limit.

So in pratice, you'd be able to make a 10x10 room safely (or 200x10, for that matter), but a 11x11 would collapse. Make the game check for this just whenever a square is digged, in the area where it's happening (say 10 squares in each direction, for the z-level being digged out and the level above) and it shouldn't need too much processing power.

A cool side effect of this: you could possibly larger rooms, if you made a 'domed' ceiling, which is realistic and makes for challenging desigs. For example, you could make a 15x15 room like this:

##################
##################
#######          ######
##                              #
##                              #
            ^    ^
             2    1

The ceiling in the middle of the room (1) won't fall even though it's more than 5 squares away from the wall, because it's supported by (2), wich is in turn supported by the wall.

[Kogan Loloklam]

I'd love to see a realistic method like that implemented, but the processing power needed would probably make the game unplayable.

Nope.
Retro's Argument sums up the only problem with realistic stuff. How do you make it simple for players? Toady has stated that that is the only real concern when it comes to doing this stuff. Listen to Dwarf Fortress Talk #8 for more information.

Retro,
I disagree with not being able to build towers without some kind of odd cables since this was not only possible, but happened quite frequently (The white tower, built in 1066, is 90 feet high. In 537 Emperor Justinian I commissioned a fairly famous building that was significantly larger than that.) The largest outdoor dwarf fortress structure I made that wasn't some kind of statue was only 5 floors tall. Even building statues, with a realistic model as long as you use the right material for your "core" (HFS Metals, for example) then the whole thing can stand up pretty easily. This does make some megaprojects harder, but also it gives each megaproject that much more value in my opinion.

But I see what you are saying, in that learning all the materials would be necessary, but would this be true if there were a simple "overlay" you could activate that explains the likeliness of collapse of your structures (similar to the display that activates when you turn to view the Depot)
Say something that makes every "supported and stable" tile green, stressed tiles (ones that if you do a lot more with it, they might risk collapse) yellow, ones that are barely stable as red. It'd be fairly simple then to see that something is unstable and at risk of collapse that way.

[Jordix]

How do you make it simple for players?

I may have expressed myself poorly, but don't want a complex scientific model either - if I had to pull out a calculator every time I dig a room, I might as well be working. I'm just saying there should be a way to stop you from just designating a whole z-level to escavate, minus a couple squares for support, and forget about it for a year or so until your miners reach legendary.

It should work a bit like the model for hidrostatic pressure, I think: Not an exact scientific model, just an aproximation that's simple to understand but keeps people from doing absurd stuff, while providing alternatives to make it work.

[BloodBeard]

I see no reason why giant 100 z-level towers and even sizeable underground caverns can't still be possible with the right system, even mine. It's when an entire mountain can be supported by a single floor tile made of sand is where the need for a system is argued.

Real caverns can be surprisingly large. A lenient cave-in system that won't effect all but the most outlandish setups would still be realistic I think.

[Grimshot]

I'm all for realistic cave-ins. As for those against it, most likely there will be a option to turn it off so it operates like it does now.