Print

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

[Dante]

Silverionmox, I like your ideas. I only disagree with the last (balance), in that Toady can try to balance it somewhat as he goes along. In fact, he needs to, because if the game is unplayable while it's in alpha - which it will be for the forseeable future - then he loses players and their donations, so he can no longer keep developing the game.

Way too detailed

Too Detailed for Dwarf Fortress? 

I don't think the level of detail that Silverionmox brought up needs to go in. But I do have to say, the "we don't need farming more detailed" argument in general doesn't work when you have realistic systems such as steelmaking requiring pig iron, iron, flux and coke; or things like various ways of animal trapping and extracting which people basically never do, except for the novelty.

Somebody has a toadyquote in their sig with Toady talking about 'flaws in the current eyelid system'. Yeah, that's what we have to contend with in terms of realism.

[Draco18s]

Somebody has a toadyquote in their sig with Toady talking about 'flaws in the current eyelid system'. Yeah, that's what we have to contend with in terms of realism.

Its a flaw that happens on the dwarf level though.  For the player it doesn't really impact them, which is fine.

The same level of detail for farming is what we're looking for really: make farmers actually have to spend time working and increase the interaction of the player to a slight degree so its not "farm plump helmets forever" and be done with it.  Not busy-work (micromanagement), not difficulty per say, but sophistication.

[Footkerchief]

The way I'd put it is that farming should become an actual gameplay mechanic rather than a Free Stuff Button.

[CobaltKobold]

The way I'd put it is that farming should become an actual gameplay mechanic rather than a Free Stuff Button.

currently Free Stuff is bound to s, if I recall correctly: :3
[BIND:TRADE_SEIZE]
[KEY:s]

[G-Flex]

The way I'd put it is that farming should become an actual gameplay mechanic rather than a Free Stuff Button.

Cooking being so ridiculous doesn't help. Not only is farming easy, simple, and essentially zero-risk, but cooking is the only trade that can make a stack of fifty trade goods at once (with an insane profit margin, percentage-wise).

[Dante]

Pets not eating doesn't help either. You can wall yourself off from the world without farming as long as you have dogs and a well and a nice enough dining room to make up for the lack of booze.

[G-Flex]

Not having booze carries other penalties, though, considering that dwarves wind up acting much more slowly after being sober for too long.

Animals requiring food would be good, but one possible problem with that is: What food? In real life, you don't always give your animals the same food you'd have. Ruminant animals (like cows) will eat things you can't even digest, for instance, and a lot of pets would either find their own food (cats kept for pest control) to some degree, or be fed the stuff that you just plain don't want (table scraps and that sort of deal). So some differentiation there would be nice... of course, this would best go in when animal behavior in general gets revamped, so we can have stuff like predatory animals actually seeking out things to kill and eat, and herbivorous animals grazing and looking for fruit or whatever else in particular. Animals actually requiring maintenance opens the door for other things; since animals would actually have cost associated with them, it would make more sense to have animal power (where it makes sense), for instance.

Of course, that's a whole other topic in general, but you make a good point that animals-as-food is basically a free lunch (pun intended... yes, I went there), since they don't need food of their own.

[Dante]

I think a small set of tags governing what they can eat would be enough, and the actual code could be based (I'm guessing) almost entirely on the code that's already there for dwarfs. Something like
[GRAZER]
[SHRUB_EATER]
[VEGETABLE_EATER]
[MEAT_EATER]
[MEAT_SCRAPS_EATER]
[VERMIN_EATER]
would define whether they ate grass, shrubs, things from farm plots, meat, scraps of unbutcherable meat / chunks, or vermin. For example.

[G-Flex]

Well, you'd also probably want things that have to do with behavior as well. Some meat-eaters hunt in packs, others hunt alone, and other are scavengers, for instance.

I'm also not sure why MEAT_SCRAPS_EATER has to be some specific tag. I can't really think of an animal that would eat meat scraps but wouldn't eat meat, and most that eat meat would eat "scraps" too. Of course, this depends on what you consider "scraps" to be.

[NW_Kohaku]

Sorry but I'm skipping pages 2-15 here, and pretty much responding to the original post, and to silverionmox.

I have a problem with "realism" being the (only) goal in making anything...  The objective should be to make something entertaining and engaging.  Simply making farms larger, and require a larger portion of your workforce doesn't make it more engaging.  Making a player sit and measure Ph levels in an obtuse UI, and try to force dwarves to use one specific kind of fertilizer as opposed to several dozen others is also a problem.  In fact, considering how often people just flat-out ignore work cancellation orders, it could be lethal to require too much micromanagement.

I think it best to work from the ending backwards - what level of involvement we players want in our farms, and then figure out what system would give that to us.

One of the complaints is that it just plain takes too little effort to set up a working farm - which is entirely valid.  A 5x5 plot can be set up within a minute of starting a new fort, and a single farmer can get your entire food supply nailed down for the first two years.  Making something like 20% of your fort farmers just to hit subsistance levels, however, is absurd, especially since that doesn't even leave room for farming to craft textiles.  (And my current fort has about 300 tiles farming, most of which are to dyes and cloth plants, since I have a very large textile mill (five each of the various textile workshops) as my primary income source... Call me a tradesdwarf.)

Basically, I don't think "throw more dwarves at the problem" is a good solution.  It just means you have more dwarves being tied up running longer laps hauling food than you do now, and it doesn't really involve anything more from the player, but giving him less dwarves to do other things that the player considers more important and engaging, like defense construction.

Something that I do think should be a goal, however, is some kind of "conservation of matter".  Farms right now produce food from nothing (a "fountain"), and the food dissapears into nothing when eaten (a "sink").  It's not exactly right, but at least it balances out in a way.  Metals and stones, on the other hand, are basically sinks without a fountain, unless you use obsidian farming, and excepting traders.  If we are to start using stones as fertilizer (not a bad idea to begin with), then we are going to be needing to mine stones at a certain pace just to be put into a process that ends in a sink.

Granted, I doubt many people ever hit a point where they've completely mined out their entire fortress site, but a goal of "realism" shouldn't produce a model that annihalates matter.

----

There are two things that I think would produce a good system... First, I agree with a system of requiring farmers to supply water to the farms.  Simply making them dump water with buckets may be silly-seeming, but probably better.  We don't want players setting alarm clocks for themselves just to remember to re-irrigate the fields with controlled floods, it's far too much micromanagement.  It would be best if we could trust dwarves to be able to irrigate their own fields, but let's face it, they're not smart enough to know to engrave the inside of a room instead of running outside when going outside is forbidden, straight into a seige, unless you actually lock them in the room until their work is done (and remember to let them back out).  In short, not holding their hands through a flooding is like making an engraved invitation for a watery doom.  Better, simply having a water source near the farm should be fairly fast and easy to track (especially if a well), and use something we are already fairly used to. 

Also, above-ground plots should benefit from rain... walled-in above-ground farms should be potentially as easy to manage as things are now... after all, real-life wheat farmers really did pretty much only work during planting and harvest times, and did largely nothing during the Winter.

----

The other idea is a simpler version of soil quality.  I think the best way to handle soil quality is to use a very simple metric - instead of having actual mineral levels in the soil, it would be best to keep this down to a metric of something like "red dirt", "brown dirt", and "yellow dirt", representing some general soil quality, without getting into chemistry.  After all, these are just middle-age-tech dwarves, soil color is basically the best indicator they have.

The way that this would require player intervention should be entirely front-loaded, so that there is no micromanagement once the ball gets rolling, unless you want to change crop loads.

Each plant could have soil/nutrient requirements (depleting one color heavily, and a different one lightly), and potentially replinish one color of soil at a certain rate, as well.  There should also be some kind of "fallow plant" (like clover) that would help greatly replinish the soil, while at the same time being generally useless, except as animal feed.  Clover "plowed back in" by not being harvested would boost soil even more.  The idea should be to allow for a crop rotation schedule that would give a carefully scheduled farm better production rates than one simply set to "farm plump helmets forever", especially as soil depletion would give less and less return (although not to the point where farming would be useless... hopefully, this shouldn't be a punishment for anyone who doesn't want to read up on the system, but a bonus for those who do take steps to maximize their gains, while even a poorly-designed farm can produce enough food, if it were simply large enough, and well-staffed enough). 

To help in this, it might actually be better to create a "farming screen" somehwere in the interface (like from the z-status menu, where each farm plot can be scheduled individually, both for crops, and for fertilization), so that you can plan crops on a loop multiple years in advance, rather than the current, rather user-unfriendly system.

Fertilizer would still be useful, of course, but would need to be broken down into what "color" of soil they produced.  It would actually probably be best if "chunks" and "remains" decayed into an invisible, but collectable "soil" item.  These might produce, say, brown soil, while potash produced, say, yellow soil.  (I have no idea, really, what soil color really denotes, I'm just throwing out arbitrary examples.)  Since dwarves live in a world where the only excretion is vomit and blood, and I doubt we really want vomitoriums just for fertilization purposes, chunks are the best source of additional fertilizer from renewable animal sources.  Fertilization in this way could let you keep running one specific crop you really like/profit from more often. 

----

As a side effect of this sort of system, it should actually be possible to have a much better selection of crops in the game, and not just in terms of making more dyes.  Using dyes as an example, however, hide root is much maligned for its lack of value when dyes are useful only for adding value.  If hide root were to replenish more soil than it depleted, a low-value produce would still be viable.  Meanwhile, sunberries or other high-value crops might deplete heavily, requiring much fertilization support to make viable.

----

On a completely seperate note, I do think it would also be useful if all crops didn't die on a new year (for whatever reason they do this), so that there could be a potential for "orchard" farming.  Especially in warmer (non-freezing, that is) climates, having fruit trees that only require very occasional fertilizing and watering, but which take a year or so to raise would produce an interesting variety in the types of crops that one can produce.

----

I would also like to suggest (although this may be more suited to a seperate topic) making a "garden" - basically, useless flower crops or shrubberies or the like that simply raise room values when they aren't picked (or make flower vase furniture or decorations for other items if they are), or can be used to create a "plant garden" to go along with a statue garden.  It would especially give those darn elves something to do with their time, and it's partially inspired by my own dwarven civ's elf queen (whose only job beforehand was "farmer").

[Silverionmox]

Silverionmox, I like your ideas. I only disagree with the last (balance), in that Toady can try to balance it somewhat as he goes along. In fact, he needs to, because if the game is unplayable while it's in alpha - which it will be for the forseeable future - then he loses players and their donations, so he can no longer keep developing the game.

As it stands now, the game is on permanent "introduction" difficulty: food grows like manna from heaven, any dwarf that falls is replaced withingthe year by a horde of immigrants, and most skills are trained up to legendary within years. It's  these things that don't need to be addressed right away: fiddling with immigrant quantities, or the trade value of prepared food is just editing numbers in the raws and belongs in the beta or even final stages.
What is needed now is to experiment with new systems, and how they influence the game flow and the options the players have. Respecting or improving the existing balance is not. It's true that we should err on the side of abundance in the vanilla alpha, in order not to frustrate players. Let's test the weights first, risking to tip the scales for a while, before trying to balance it.

I have a problem with "realism" being the (only) goal in making anything...  The objective should be to make something entertaining and engaging.  Simply making farms larger, and require a larger portion of your workforce doesn't make it more engaging.  Making a player sit and measure Ph levels in an obtuse UI, and try to force dwarves to use one specific kind of fertilizer as opposed to several dozen others is also a problem.  In fact, considering how often people just flat-out ignore work cancellation orders, it could be lethal to require too much micromanagement.

The effects for the non-agronomist players of DF would be experienced as follows: 1. Players attempting to order the planting of a certain plant in a certain biome would notice that it has some penalties (interface suggestion: colour the plant's name appropriately, eg. greener for better modifiers, redder for worse) in eg. marsh, but not elsewhere. 2. Players attempting to cultivate acidifying plants in the same spot for years would notice their yields diminish. 3. An option would exist in the plant zone menu that says "Lime the soil", with a neat help option that explains what it's all about. So even if they ignore pH, they'll be warned about it's effects beforehand, or will be able to predict them.

I think it best to work from the ending backwards - what level of involvement we players want in our farms, and then figure out what system would give that to us.

Different players want different things, hence my remarks about balancing. Only want to dedicate 10 instead of 20 dwarves to your food supply? Double plant yields, or halve labour requirements, or halve food requirements in the appropriate raws.

Basically, I don't think "throw more dwarves at the problem" is a good solution.  It just means you have more dwarves being tied up running longer laps hauling food than you do now, and it doesn't really involve anything more from the player, but giving him less dwarves to do other things that the player considers more important and engaging, like defense construction.

If the jobs are predictable, they can be automated, with very little cpu effort. Farmers would likely just run back and forth on the same old paths in their little burrows, requiring almost no attention from the player, if he is content with agriculture that has a high labour productivity, but a low space productivity.

Something that I do think should be a goal, however, is some kind of "conservation of matter".  Farms right now produce food from nothing (a "fountain"), and the food dissapears into nothing when eaten (a "sink").

That's my intention with nutrient and energy tracking.

It's not exactly right, but at least it balances out in a way.  Metals and stones, on the other hand, are basically sinks without a fountain, unless you use obsidian farming, and excepting traders.  If we are to start using stones as fertilizer (not a bad idea to begin with), then we are going to be needing to mine stones at a certain pace just to be put into a process that ends in a sink.

Realistic mining would produce rock dust and grindable rubble. Stacking/hauling is a prerequisite though.

There are two things that I think would produce a good system... First, I agree with a system of requiring farmers to supply water to the farms.[...] 

Plant zones would do that nicely: "water plants", either on schedule or on demand does that. The on demand efficacy would depend on the farming skill (just generating a hauling job for a bucket of water after moisture drops below the desired level; better skill=faster response). Irrigation moats or rain or buckets or whatever would all be treated the same because they all affect the moisture variable.

The other idea is a simpler version of soil quality.  I think the best way to handle soil quality is to use a very simple metric - instead of having actual mineral levels in the soil, it would be best to keep this down to a metric of something like "red dirt", "brown dirt", and "yellow dirt", representing some general soil quality, without getting into chemistry.  After all, these are just middle-age-tech dwarves, soil color is basically the best indicator they have.

Three variables are three variables, names are just that. I can attest that medieval farmers and landowners knew pretty damn well which crops required how much and which fertilizer. It's about the only thing they put on their lease contracts between the lords and the tenants.

The way that this would require player intervention should be entirely front-loaded, so that there is no micromanagement once the ball gets rolling, unless you want to change crop loads.

Completely agree. All the modifiers ought to be counted together and indicated by appr. colouring the plant's name in the menu.

Each plant could have soil/nutrient requirements (depleting one color heavily, and a different one lightly), and potentially replinish one color of soil at a certain rate, as well.  There should also be some kind of "fallow plant" (like clover) that would help greatly replinish the soil, while at the same time being generally useless, except as animal feed.  Clover "plowed back in" by not being harvested would boost soil even more.

That sounds good for the standard set of surface crops. Why limit it to these? Just let the nutrient requirements be described numerically rather than be picked from a few hardcoded ones.

The idea should be to allow for a crop rotation schedule that would give a carefully scheduled farm better production rates than one simply set to "farm plump helmets forever", especially as soil depletion would give less and less return (although not to the point where farming would be useless... hopefully, this shouldn't be a punishment for anyone who doesn't want to read up on the system, but a bonus for those who do take steps to maximize their gains, while even a poorly-designed farm can produce enough food, if it were simply large enough, and well-staffed enough).

In my system, setting enough fallow periods would allow one to get away with bloody murder, agriculturally.

To help in this, it might actually be better to create a "farming screen" somehwere in the interface (like from the z-status menu, where each farm plot can be scheduled individually, both for crops, and for fertilization), so that you can plan crops on a loop multiple years in advance, rather than the current, rather user-unfriendly system.

That'll be useful, although I'd rather keep the focus of the player on the map, where the farm plots are, to remind him of the local environment.

As a side effect of this sort of system, it should actually be possible to have a much better selection of crops in the game [...][
/quote]More variety will come with combining more variables (like, eg. pH).

On a completely seperate note, I do think it would also be useful if all crops didn't die on a new year [...] orchards

Doing it through a temperature variable wouldn't only allow the player to cultivate palm trees in the subtropics, but also near magma pipes or caged fire demons or such. Having the growth timing depend on a combination of plant speed and environment variables would eliminate death-by-calendar.

I would also like to suggest (although this may be more suited to a seperate topic) making a "garden" - [...]

Unifying the interface to mess with plants keeps it simple, I think.

[NW_Kohaku]

I'm not going to quote things from that, Silverionmox, since I think we're largely agreeing on the overall nature of system, but disagreeing more on how abstract or exact the system should be.

I think that, rather than having a "saturation level" of the soil, to replicate a plant's water tolerance level, a farming plot tile should, instead, have a boolean "wet" and "dry" inidication, with a timer, potentially based upon each plant's needs, for how long it can go without needing water again. 

Actually, thinking specifically about implimentation, it might actually be best to work with a system where any given plant has a "target growth number".  (It might actually help if there were some kind of visible marker for how far along a plant's growth was, instead of "planted" and "ready to harvest".)  Every certain number of turns (possibly going by the tick of days on the calendar), a growth modifier would be recalculated, if need be, and the growth modifier would be added to the running "current growth" value, until it hit the targetted growth amount. 

Soil conditions would multiply this amount... there would be certain brackets of soil quality.  Soil quality below all other brackets would give a growth multiplier of 1.  Poor soil quality would multiply by 2.  Average soil quality would be worth 3 or 4.  Superlative soil quality could be worth between 5 and 10.

In cases where multiple soil types are depleted by a crop, these may be best just added together before multiplying this factor with the other factors.  My first reaction is to say "averaged together", but that is actually generally a bad idea in a situation like this, as averaging factors tends to make efforts to up a single factor (in this case, a single type of soil) less valuable.  That is, if you want to add, just to use my previous example soil kinds, red fertilizer to a farm plot that needs red and yellow soil qualities, then you would only be getting half the benefit from a single bag of fertilizer that you would get from adding it to a red-soil-only farm plot.

Soil quality would be ticked down with every check of the system, if it were a depleting crop.  This means that a crop might be well-fertilized early in the growing season, then move down to just average levels as the days pass.

Anyway, next factor would be watering.  Water could be a little simpler: When watered, a number is added onto a "soil wetness" value for that farm tile.  This number could be different depending on how much water a plant needs, so that "thirsty" plants would have a smaller number each time it gets watered. This ticks down after every day/check of the tile.  When it gets to a low value (let's say 3 or 5 or so), the water multiplier hits 1.  Before that, the multiplier should be 2.  Hitting this low growth state should also trigger a farmer to come back and water it.  If water ever runs out, the water multiplier on growth hits zero, and plants will stop growing without water.

The final variable could be a "tending" variable, based on the skill of the dwarf doing the farming, rolled only at planting, as it is now, since we don't really need to have the farmers running out and babysitting every crop.  This might actually be a relatively large number (that is, something where a legendary+5 could roll as high as 25, while a dabbler could roll between 1 and 9 or so).

There's two ways to go with how these target growths could work, however.  One is to make harvests come sooner if you get really large modifiers, but it raises questions of just how you determine harvest stack size.  (If stack size was simply inverse to the amount of time it took to hit the target growth number, you would have geometrically better returns the better your soil and grower's qualities were, as well as how hot you were on ensuring your plants are re-watered.) 

As such, I think it may be better to go for the current system, where once a tile's scheduled harvest time is up, the crop is marked for harvest, "ready or not".  (Probably more realistic this way, as well...) There would, again, be brackets for how well a crop grew, with a minimum bracket simply being crop failure, and with gradiated targets producing one more crop on the stack.

----

While it may not be in the spirit of "realism", it may be easier on the CPU, while we're throwing more calculations at it, to make an entire farm plot be considered a single cohesive unit.  Seeds are to be dumped into a central tile and "worked" until they have filled up the full amount of seeds that a plot of that given size would require.  Water buckets would hold x amount of water, and dumped, would share that water amongst all the tiles of the plot, rather than requiring bucket brigades to hit every single tile individually, and require tracking moisture loss on every tile, as well.  Bucket brigadeers could then be instructed to just keep watering to whatever has the lowest value of moisture.  Likewise, soil quality could be averaged out against its component tiles.

A problem arises when talking about tracking soil quality, however.  When we are talking about tracking only farm plot quality, there is the problem that farm plots can be un-designated.  Either we have to make those soil tiles remember its soil quality, or we have to disincintivize unconstructing and reconstructing farm plots (possibly by making it extremely labor-intensive).

----

This should really require an improved stacking feature, though, as it would really help in harvesting all these crops if people could pick more than one tile of crops at a time before going back to the warehouses.

[Dante]

Sorry but I'm skipping pages 2-15 here, and pretty much responding to the original post, and to silverionmox.

The original post was way, way old, and we've moved on a lot since then. I revived this thread to see where the thinking was at now. You should probably read the new debate from the last few pages, starting from my summary posts here, here and here.

[/shameless_self_promotion]

[Silverionmox]

NW_Kohaku:

Not quoting from that monster is understandable.

You seem to contradict yourself with regards to moisture, mentioning both binary and larger values. I picked a value (0-7) that would correspond with a non-soil tile, so it's easy to visualize and easy to grasp how much water it is for the player. And it's as simple as possible to predict how much water would be in a channel that would be dug in soil with a moisture 4, for example. (Additionally, when seeing a tile is 0 or 1, the player has no clue how long it will stay that way, nor how long a plant can go without water. A broader scale has the advantage that the player can see it diminish too quick for his taste, and take measures before it's too late.)

This is a picture of how they did it in Emperor: Rise of the Middle Kingdom, a game that's recommended for most DF players, by the way.

The generalized tending variable is a good idea, since there are lots of small tasks to do with the plants that aren't worth the bother (or pathfinding expense) to simulate separately.

I wouldn't try to separate a farming plot from its environment that way. There would be too many inconsistencies to address when Something happens on part of the farm plot, when part of a moat falls dry and only the farthest part of the plot ought to get problems, etc. Something like that will be necessary, however, for settlements that the player isn't involved with currently (off-map farming villages for example). Farming isn't a big fps threat, anyway. The biggest danger is the planter's pathfinding (although that will be highly predictable, allowing more efficient algorithms) and as a distant second the frequency of plant growth updates.

[NW_Kohaku]

(Oh, hey, Emporer! Back when you had to actually worry about pathing the bazaar walker, or the people in the house next door to the bazaar will starve because they can't find the market... good times, good times.)

IIRC, Emporer's water levels were extremely simple - just a pure measure of how many tiles away from a water tile the farm was built.

As for the boolean - gradiated thing, I can sometimes "think out loud", and change my mind in the middle of a post that long... Anyway, if it were the player's job to irrigate it, x/7 style moisture measurements would certainly be the correct way of doing things.  The difference is that the player has no direct control over when things get watered, so he really only needs a "there is a problem" and a "there is no problem" indicator.  That said, a x/7 moisture indicator would probably not hurt anything, so why not?