Print

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

[NW_Kohaku]

The problem with just fertility as a value is that the simpler the system is, the less a proper crop rotation system can be developed.

We currently have so few crops because everything is basically the same except for what the crop can be used for - there is no difference between a prickle berry or a strawberry or a sun berry when you are growing them, but the prickle berries are lower value crops, and sun berries are high value crops. 

When you add in Nitrogen, you have to balance nitrogen-heavy feeders with something that supplies nitrogen fixation, or supply fertilizer to balance out your depletion... but you still only have one dimension by which to measure your crops.

When you have Potassium, you have one more thing to balance, but since you don't get "Potassium Fixating" plants, you aren't going to have it so easy if you deplete potassium too recklessly.  This also makes plants follow another axis along which they can be differentiated.

Frankly, if we just keep adding MORE variables, it doesn't really make the system any more difficult to comprehend (how much harder is it to comprehend NPK than NK?  It's just one more thing you need to push "green") while allowing a much greater diversity of plantlife.  (And speaking of more variables, I personally added in biomass as a recommendation, as it means that mushrooms are not photosynthetic, and if we wanted more variables, explicit temperatures and soil drainage would be the two main ones...)

I may have started out being leary of a complex system, but the more I put myself into it, the more a complex system makes sense and gives the player real diversity and real choice. 

So what if it's hard to learn at first?  So is everything else in DF, but we wouldn't say that we should strip out the military system to be a simple matter of attrition just so that people don't have to learn how to mine and make metal weapons and armor...  We should strive to make the game as deep as possible for those who understand the game first, and then simply work out the best way to explain it to those who don't understand later. 

It's that final part of presentation that I actually was intending to talk about next: In arguing with Jiri, I looked back at what had actually been said about the Farm Overseer, and realized that I had largely only talked about what its function should be in passing, but never really talked about how to build it, or what it should look like.  As such, I'm going to try to put together a post on that, which hopefully doesn't get too mired down in distractions, the way the stuff on page 34 did.

[NW_Kohaku]

Well, first, I want to get a general outline of what the Farm Overseer/Agriculture Menu should allow you to do...

• Farm Overseer is an appointed position.  It should probably take a meager office, like a bookkeeper, for the reason that it requires keeping records.
• The agriculture menu should have all information relevant to long-term planning of your farms, except for the physical layout of those farms.  The military screen is the best model I can compare to, even if it is a currently imperfect model.
• The Farm Overseer should keep records of previous year's actions (at least back 5 to 10 years) and the resulting crop yield for all crops. (Including if farms were accidetally flooded, which would set the "water" nutrient straight to 255, and probably kill most crops.) He also records other important information, like spotted pests and their prevalence.  To add this to the record, he needs access to his desk and bookkeeping time, just like the bookkeeper position.
• The Farm Overseer should also have a menu presenting the information that your farmers have acquired about all the crops they know of (important if we start involving procedural crops, and have to learn about them from the game itself, and not necessarily the raws).
• Crops that your civ plants will have full knowledge available from the start, as they are theoretically something your civ has planted for generations.
• The herbalist who first picks a new plant uses their herbalist skill to learn as much as they can about a new plant.  After that, the Farm Overseer's farming skill is used to add to that knowledge base each time a crop is used until you have gained full knowledge about a plant.
• The knowledge you have about each crop must be displayed in a way that doesn't break game knowledge (I.E. You can't SAY "Nitrogen".)
• In another window of the Agriculture Menu, you control every seperate farm zone, and you can schedule up to five years in advance, and set this schedule to be repeated in a cycle of a selectable number of years up to five.  This schedule includes what crops to plant, as well as how the soil should be treated, such as by fertilizer and watering.
• In another window, the Farm Overseer may give some reccomendations for the immediate future on given farms, saying what crops will be suitable for the soil you have right now (although he probably won't have the ability to tell you a full sustainable crop cycle).  The Farm Overseer cannot give reccomendations on crops he doesn't know enough about, so experiment a little.
• Every farm should also have a note about what how many of its resources are still available to it - if the water source you are drawing from is getting low, how much fertilizer you have in your stockpiles, etc.
• In another window, an overview of pest sightings and dangers can be presented to the player.  A history of their appearances should be available, provided the Farm Overseer has been able to track everything.  This knowledge should obviously tell you what crops are attracting and are vulnerable to those pests.
• If we go the advanced route, and have anti-pest measures, such as developing pesticides, either through predator species (like garden insects that eat pest insects, or snakes that eat small herbivorous mammals, or funguses bred to kill dangerous crop diseases), then we need a sort of "research and development" screen for the Agricultural Lab to work on developing this.  Such a menu might also be useful in designating how much of a crop is to be taken for seed, and how to unnaturally select the seeds for better crop development if we are going the fully procedural route...

Considering what Toady said about not making farms buildings anymore, I think it might be best to have a zoning system similar to the burrows system that puts any oddly-shaped area into a single farm, which recieves attention from farmers in the same way (so that seeds and fertilizer is spread among it equally... even if you haven't actually set up irrigation to any but a small portion of the farm, so that's on you if you don't do it right). 

Irrigation systems, like the "sprinkler" that I talked about using hydrostatic pressure and small pipes with holes in them hanging over the farms connected to a dwarf-operable valve, or even just cisterns, could also be put in the same zone so that you can designate specifically how to get the water (for the purpose of letting the scheduler know where to send your dwarves).

[Hammurabi]

So what if it's hard to learn at first?  So is everything else in DF, but we wouldn't say that we should strip out the military system to be a simple matter of attrition just so that people don't have to learn how to mine and make metal weapons and armor...  We should strive to make the game as deep as possible for those who understand the game first, and then simply work out the best way to explain it to those who don't understand later. 

There's a saying: A moment to learn, a lifetime to master ...

Do you agree or disagree with that statement? 

[Jiri Petru]

Just got an urgent idea, sorry for ignoring your post, had to write this down. I think I came up with a solution to present even a very complicated system quite simply and gameplay friendly. I think most of your system would fit in nicely, Kohaku (I haven't studied it in detail), perhaps with some omissions.

This is the framework:

(I don't know if I chose a proper image. If the common consensus is to use a different model - like perhaps NKH -, imagine that one instead of this image. The only important part is that it divides things into broad categories, so you can say something like "this is acidic clay" instead of "this has acidity 8, salinity 7, moistness 6" or whatever.)


These soils are tracked during worldgen and occur on geologically realistic places. Each plant prefers different soils, leading to varied biomes in worldgen. So far so good.

The actual Dwarf Mode implementation:
The player sees only the general category ("This is a clay loam", "this is sandy loam"), not more avoid information overload. They don't need any more information anyway. When you build a farmplot on any land, the plot menu shows these information and lists the plants best suited for this soil (the game should be clever enough to draw them from the raws). Something like:

FARMING PLOT

Soil type:     Silty Clay Loam

Recommended crops:
Lowland Grass
Bloated Tubers

Not recommended:
Prickle Berries
Strawberries

Click for more (when you click this, it shows a list of all crops and their suitability, expressed in percents)

(Insert normal farm controls here)

You build a field and select a crop to be grown, taking advice from the recommended crops. Voilá! You have a functional farm! (insert my previous automation/defaults suggestion here).

My personal preference: The soil components are not tracked in Dwarf Mode. They were tracked in worldgen, but not in fortress mode. Honestly, I can't see the added value in that, for me it's only headaches. Yeah, this means that I'm loosing the ability for crops to change the soil type. I don't really miss it, quite on the contrary: soil changing invisibly during play might be too confusing and hard to communicate to the player.

Optional: The soil components are tracked. Events like diseases or snail goo can change the soil properties. Plants can change them too, even though I would advice against it - changing soil type should be something rare and "big" that doesn't happen very often. These soil properties have absolutely no effect, until the soil moves to a different category. So eg. as long as you are in the "clay loam" category, all rules are the same, regardless of whether the soil has 50 or 40 percent sand. Once you accumulate more silt, for example, the "clay loam" changes to "silty clay loam". At this point the farming menu changes and shows the recommended crops for the new soil type.
In short: while components are tracked, the don't have a direct effect on crops. Instead, they can sometimes cause the broad category to change, at which point there is an effect on crops.

The pros of this are that the amount of information you need to show to the player are very low - basically just the soil name and then the percentual suitability of crops. (Perhaps adjusted by the skill of your best farmer? Something like the broker's impreciseness). The categories shouldn't be many, I'd say perhaps 10 - this is enough to allow lots of variety without being too overwhelming.


-----------------

ON TOP OF ALL OF THIS add some general fertility rating, which should be just a single abstracted variable. You increase fertility by fertilising, you decrease it by crowing crops. That's all. Basically what nil said here.

The fertility variable can internally be measured in any detail (say a range from 0 to 1000), but for display purposes it would again be grouped into a few colour-coded categories. Something like: very high, high, normal, low, very low (perhaps a bit more, but no more than 10). In the end, a farm menu would look like something like this:

FARMING PLOT

Soil type:     Silty Clay Loam
Fertility:       Very low

Recommended crops:
Lowland Grass
Bloated Tubers

Not recommended:
Prickle Berries
Strawberries

Click for more (when you click this, it shows a list of all crops and their suitability, expressed in percents)

(Insert normal farm controls here)

----

When this is implemented along with the defaults and automation I mentioned, I think we have a nice, flexible, and yet easy-to-learn system.

 

[NW_Kohaku]

I think most of your system would fit in nicely, Kohaku (I haven't studied it in detail), perhaps with some omissions.

... please at least read it before assuming you know what it is I am proposing, especially if you think it is insufficient...

It would do quite a bit more for your case, as well, if you could get "NPK" straight.

Look, Jiri, we COULD do some sort of tracking of relative values of clay, silt, and sand (for whatever that means) plus some "fertility" value.  But then we'd be tracking four variables of soil quality plus water.

This is "Simpler" than my method of tracking... five variables of soil quality plus water.

You're throwing verisimilartude and realism out the window, as well as most of the really interesting aspects of the system for no benefit.  Why would we want to do this?

There's a saying: A moment to learn, a lifetime to master ...

Do you agree or disagree with that statement? 

Disagree.

Mastery is merely learning to completion.  In the form of learning you are talking about, and as applied to this game, they are essentially equivalent concepts. 

The entire saying is simply vacuous.

[Zalminen]

I do agree with NW_Kohaku that the system should clearly encourage crop rotation and Jiri's latest idea is too simplified.

There's a saying: A moment to learn, a lifetime to master ...

Mastery is merely learning to completion.  In the form of learning you are talking about, and as applied to this game, they are essentially equivalent concepts. 

The entire saying is simply vacuous.

Well, no. In this context, a moment to learn just means learning enough to be able to use.
Compare to something like Go - you can learn the rules and be able to play in a couple of minutes but even after 10 years you're still learning to play better.

In DF terms it should be very simple to just plop down a simple farm that will produce food but there should still be enough variation, complexity and challenges to keep things interesting for veterans as well.

[Soralin]

In DF terms it should be very simple to just plop down a simple farm that will produce food but there should still be enough variation, complexity and challenges to keep things interesting for veterans as well.

At the same time though, if you can just put down a simple farm that can produce food enough for your dwarves to survive, easily and without management of it, then why bother to learn any of that variation, or complexity, since it's not necessary to do so?  As for 'challenges', easy and challenging are mutually exclusive, you can't have challenges to keep your dwarves from starving to death, if it's always easy to keep your dwarves from starving to death.

[Fieari]

I really like the full nutrient system described here (NPK+Ph+Biomass+Water) as it creates a lot of divergent gameplay. The trick, as has been mentioned before, is communicating this to the player, and making it both meaningful yet not all-consuming, game wise. This isn't a farming sim after all, but farming is important.

I think the solution revolves around managers, auto-rotation, and dwarven free will.  Basically, what I mean is that you can put down a farm, and have your farmer dwarves worry about all the details so you don't have to.  On the plus side, you don't have to worry about a thing!  On the downside, your farmer dwarves may make decisions you don't want-- and suddenly you aren't producing as much food as you need.  This is FUN(tm), and in my opinion, the GOOD kind of fun.

Set to auto-grow, the crop chosen should be selected based on certain criteria. The dwarf will basically take the full list of every plant that exists, and narrow down the choices until he gets the one to pick. If at any point, there is only one plant left, that's the plant to chose. The selection culling would be as follows:

Step 0) Based on farming skill, estimate the soil nutrients. High skill means the dwarf knows the true value of the soil, no skill means he guesses, making shit up, middle skill means he has a pretty good estimate (random +/- a certain amount). Dwarves should also estimate plant requirements based on the same skill. The farmer would also fertilize as he sees fit, based on his skill knowledge of the plant needs and the soil values. (An unskilled dwarf might never fertilize, or might over fertilize, based on ignorance, but he'd learn).

1) Select only seeds you actually have. (Obvious)
2) Select only seeds that will grow at all in this location (within the TOLERATES range)
3) Select personal favorite if available
4) Select only seeds that will thrive in this location (with the IDEAL range)
5) If and Only If (IFF) the dwarf thinks the soil has dropped below a certain level (level determined by CAUTIOUSNESS personality trait), select crops that will replenish it.
6) Select based on highest Value or Crop yield depending on dwarf personality. I'd think adventurous, achievement seeking, risk taking, greedy dwarves would go for value, for instance, while cooperative, altruistic dwarves would go for yield.
7) If there's STILL more than one crop on the list (ties?) then select randomly.

Now, of course, this is just if you leave the dwarves to their own devices, which most newbies probably will. It's good enough for starters, is reasonably intelligent (if you have skilled dwarves), and pretty self sustaining in most circumstances.

To make things more transparent, there could be a new Thought Report system, just for farming, like the announcements we currently have, except only accessible when viewing the farm in question. Here, you would see the farmer's choices, and reasons for those choices... "The soil is becoming depleted. I'd better get some Gubber Beans planted this season," or "Sun Berries will increase our fortress wealth!", or "I *love* Whip Vines, gonna plant some more."

Which is when you find out that your starting farmer is a greedy entrepreneur who throws caution to the wind. He's depleting your soil with cash crops that aren't even feeding all the new immigrants coming in. That's when you take matters into your own hand, and set a Farm Guild leader to oversee things for you. With a skilled Guild Master, it would show you the soil levels (as far as the guild master knows them) and preferably give you a nice graph of how things are predicted to go based on what plants you put in. In this interface, you would schedule the crop rotation yourself, using the PREDICTED values as your guide. The manager would even give you sorting suggestions based on the above criteria. Min-maxers rejoice!

Or, you find that you've got plenty of food, and want to start growing some cloth. Or market conditions have changed, and the neighboring human kingdom REALLY wants to import a particular type of food. Or your dwarves start getting unhappy thoughts because they're only eating one thing. All kinds of reasons could make you WANT to manually set your crops. And the nutrient system interacts with all that, adding a layer of DEPTH. And because of the automated system, there's no real complexity that would turn off a newbie! Yay!


Thoughts?

[Saint]

insane stuff

20% of the population? 8 plots per dwarf? Just so you know, 200 dwarves times 8 means you need 1600 plots with a large fortress. And 56 squares just for your starting seven.

It seems that you want the main challenge of the game to be farming. With this system, it would be unbelievably difficult to feed your dwarves. This means that nobody gets any time to do anything except farm.

I agree that farming needs to be made a little harder, but this is just extreme.

1 migrant load (lets use 21) requires 168 plots. that's the size of a small fortress.
Meaning just 28 dwarfs (not even enough for a mayor yet*) requires over 200 plots.
*I think, don't quite remember.

[Jiri Petru]

Ok, scratch my last post, it wasn't good. It didn't really allow for crop rotation, now I see it. I'll try something else... Considering you want crop rotation, there are many ways to implement it, not just one. Let me demonstrate why I'm worried about the NPK+Ph+Biomass+Water system.

This is how I understand it, correct me if I'm wrong:

You could let the AI handle it automatically (like in Fieari's example, a nice post by the way!), but you some of you seem to dislike the notion. So let's do everything manually.

There are several soil properties being tracked (6 if I count correctly). All of these continually change (depending on what you grow and what's happening to the soil) and affect the crops. Let me repeat it: you have six separate, continually changing variables for something as critical as farming. This sounds like trouble. It means the player will regularly have to check their farms to see if something important has changed... having six variables, it's very probably that something will have changed every few weeks or months. The player needs to step in and order more water, fertiliser, perhaps change crops or do something special. Every few months. Translation: we have added a menial uninteresting task to the game that needs to be done regularly and can't be skipped. Awesome.

OK BUT... the player can set the system once and then let it work indefinitely by itself, right? So they can, for example, spend 30 minutes browsing through the nutrient numbers, game wiki, farming tables, crop properties, etc. and after a lot of difficult calculations come up with a clever system that will keep all the six variables on desired levels without them drifting away in any direction... but only if they hadn't made any mistake in any of the million calculations. If they had, the soil properties will change after a year or two and whole the process must begin anew.

BUT! Let's say I'm very clever, I'm willing to spend a day learning all about the nutrient system, then I'm willing to to spend half a day calculating the numbers AND I don't make any mistake. In the evening, after a hard day's work, I have a nice small wheat farm that's able to grow tasty grain for years, and with crop rotation set so that the soil properties won't change. Victory!

Or not?

Two years into the game I realise my wheat field is to small to feed my fortress and decide to create a new one on the opposite side of the river. But alas! The soil there has different properties! I can't use my former calculations because they would make the soil quality drift away. I need to calculate it from scratch. Fortunatelly, I'm very good at it and with a small aid of a calculator manage to build my field in mere 15 minutes.

All is well until another year into the game I decide I will simply import all my food and will grow rope reed insted. I switch both my farms to grow rope reed, only to find out rope reed has different effect on the soil and makes my farm become too phosphorous over time. I need to scratch my former calculations again, and begin anew. Twice, because I have two different fields.

Swearing never to change crops I again, I play on. Another year in the future, a plague comes and kills all my rope reeds. Which is fine, I have many in reserve. What's worse is that the plague changed two of the six properties, completely ruining all my former plans. I need to run all calculations again! Cursing heavily, I turn Dwarf Fortress of and decide never to play it any more.

----

So did I get it right?  I'd be very happy if you disprove my theory and explain why the NPK+Ph+Biomass+Water is much less difficult. Because I'd really hate to have the above in the game.

(EDIT: I also doubt historical people spent their evening in the company of NPK+Ph+Biomass+Water tables, trying to come up with a way to make the land less phosphorous or perhaps decrease its pH. Hell, they didn't even know there's something like phosphorous and pH. Their farming system was "the land is good", "the land could be better", "the land is bad, let's try planting some legumes and hope for the best". If they could use such a simple system, why can't I?)

(EDIT2: Let me also say I've grown up on a village myself. My grandma has a large garden and has been growing lots of different vegetables over her life. If I asked her what nitrogen, phosphorous, and potassium are, she wouldn't now. All she does is water her garden, weed it, and sometimes buy use a general "fertiliser". Yet! A miracle! The garden lives, the vegetables grow, and the soil isn't getting too acidic or whatever.)

As for 'challenges', easy and challenging are mutually exclusive, you can't have challenges to keep your dwarves from starving to death, if it's always easy to keep your dwarves from starving to death.

Oh... when will people learn and stop doing this logical leap? Easy controls =/= easy game with no challanges. Difficult controls =/= fun game.

[Fieari]

This is exactly why I believe automation is CRUCIAL for any farming system. There should be a set-it and forget-it method, but the automated system should be sub-optimal, encouraging the ability to do BETTER with manual tweaking.  In fact, the NEED to manually tweak things in certain circumstances, or else find food in other ways (trading, remember?)

Automating the system is not dumbing it down. You're still making strategic choices. Those strategic choices are still influenced by things like location. The full nutrient system will still have emergent impact on gameplay and fortress layout, even if the player never touches it at all.  Just making it so that there are locations where it's better to grow mushrooms and areas better for traditional crops will do a lot. And some of the systems are pretty intuitive, I think-- water for instance. Everyone knows plants need water! Automated sprinkler systems will be new megaprojects, and will contribute to having sewers, which regardless of your view on night soil management, everyone wants just because sewers are awesome to go adventuring in. And where else are you going to store your alligator ranches?

With autopilot, more complexity can be reasonably managed by the player, yet each of those variables can still mean something and still influence choices.

[Draco18s]

1 migrant load (lets use 21) requires 168 plots. that's the size of a small fortress.
Meaning just 28 dwarfs (not even enough for a mayor yet*) requires over 200 plots.
*I think, don't quite remember.

Did you really mean to quote and reply to a section of thread over two years old?

[Jiri Petru]

To prevent a misunderstanding: the point above was not "Screw your system!". It was: "Screw it, unless you can automate it". Or to put it better way: "Automation is crucial, not 'dumbing down' as some of you replied to me."

This leads my back to my farming wishes:

• I designate a field
• I turn on farming labours
• I tell the dwarves: "grow this crop"! (or better yet: "grow anything!")
  
• It works!

Anything else is just a bad design. 

EDIT: By the way, automation will have to go in. Remember the farms we've seen in the latest dev_news? They won't work without complete automation. Your servants in adventure mode won't be able to handle your fields without complete automation. Once you have it, you can pretty much allow it in dwarf mode.

[Soralin]

So did I get it right?  I'd be very happy if you disprove my theory and explain why the NPK+Ph+Biomass+Water is much less difficult. Because I'd really hate to have the above in the game.

One thing you might be missing, is that it doesn't have to be optimal to work, in fact, it may not be possible for it to be optimal in the way that you have here.  Look at the raws for the plants that have been provided so far, I don't think I've seen any combinations that end up with a net 0 resources, all are net negative resources.  Which means you won't be able to produce a completely self-sustaining farm, or even if you can, or make one that comes close, it will probably be a very low producing farm, having to till plants back into the soil for resources and such.  That means, for basically every farm, you're eventually going to need an outside source of nutrients, in the form of some sort of fertilizer or biomass or such.

And since you already need an outside source to increase those resources, it's not a huge deal if your planting isn't very optimal in terms of soil conditions, it just means you'll have to produce and use a bit more of one fertilizer or another.  In fact, if you have a large supply of materials you can use for some fertilizer, you may not want to try to optimize for minimal soil depletion.  You could just use a couple types of crops that produce a large amount of food, but are very resource intensive, and produce a large amount of fertilizer to offset that.

Another thing that could help there, is simply provide a running total for what all of the plants in your farming rotation will do, or normalized to per year, if you can have multi-year rotations or something like that.  So you can look at it at a glance and see what gets depleted, what gets added, and by how much.  And then if you want it to be different, you can just tweak it a bit, add a different plant in, change this plant to that one, and see what works out.  Maybe you have an abundant supply of one resource, and so you just want to tweak it so that the whole rotation uses a lot of that resource, but not much of the others, etc.  You can just pick a plant off of the list that does what you want(plants should show this information in-game as well as to what they would do), and note the change in the total of what will happen.

As for 'challenges', easy and challenging are mutually exclusive, you can't have challenges to keep your dwarves from starving to death, if it's always easy to keep your dwarves from starving to death.

Oh... when will people learn and stop doing this logical leap? Easy controls =/= easy game with no challanges. Difficult controls =/= fun game.

But you haven't been talking about controls, you've been talking about AI and automation.  AI and automation aren't control mechanisms, they're lack-of-control mechanisms.   Sometimes they can be quite useful lack-of-control mechanisms, like when you want someone to focus on broad overarching plans, and not minute details, but if they end up taking something over in it's entirety that's not easy controls, that's non-existent or irrelevant controls.  And if the game plays for you, and does it well enough for you to survive, then that's an easy game.

I think the main question to ask, the first question you should ask, for any system of automation or other farming systems is this: How does it reach the failure state?  What state of events will result in a fortress starving to death, how will it be noticed, and what can the player do about it?  I'd be opposed to any system of automation, or farming in general, that can't reach this failure state, or that requires extraordinary circumstances to do so.

Edit: An addition on what I think a good level of automation, or at least a maximum level of automation for the player, for this system might be:  It doesn't choose crops for you, it does display what each crop does and what effect your total rotation has on nutrients, and what the current state of the field is.  If a plant is scheduled to be planted, and the nutrients won't be suitable for it to grow, then if possible the planting dwarves will first go and grab fertilizer or materials of the necessary type and amount from your stockpile and apply it to the field, to make it suitable, if it's something that can be fixed in that way. (with an optional toggle per farm, for conditions where you don't want them to restore the resources and just want them to stop instead when it gets too depleted to plant more)  (Probably a good idea when choosing a set of plants, if it displayed how much resources you would need for the first planting or something, in addition to what it does each rotation, so you don't accidentally try to grow some plant in wildly unsuitable conditions, and end up having your dwarves use up all your resources in attempting to do so).  If the dwarf doesn't have the necessary materials or ability to make the farm suitable for some condition, then they'll say so (produce a warning message).

[NW_Kohaku]

I really like the full nutrient system described here (NPK+Ph+Biomass+Water) as it creates a lot of divergent gameplay. The trick, as has been mentioned before, is communicating this to the player, and making it both meaningful yet not all-consuming, game wise. This isn't a farming sim after all, but farming is important.

I think the solution revolves around managers, auto-rotation, and dwarven free will.  Basically, what I mean is that you can put down a farm, and have your farmer dwarves worry about all the details so you don't have to.  On the plus side, you don't have to worry about a thing!  On the downside, your farmer dwarves may make decisions you don't want-- and suddenly you aren't producing as much food as you need.  This is FUN(tm), and in my opinion, the GOOD kind of fun.

...

Which is when you find out that your starting farmer is a greedy entrepreneur who throws caution to the wind. He's depleting your soil with cash crops that aren't even feeding all the new immigrants coming in. That's when you take matters into your own hand, and set a Farm Guild leader to oversee things for you. With a skilled Guild Master, it would show you the soil levels (as far as the guild master knows them) and preferably give you a nice graph of how things are predicted to go based on what plants you put in. In this interface, you would schedule the crop rotation yourself, using the PREDICTED values as your guide. The manager would even give you sorting suggestions based on the above criteria. Min-maxers rejoice!

You see, I wouldn't want this sort of autonomy in my farming - I'd much rather micromanage my farm than find out too late that my farmers have been screwing me.

I certainly like automation - that's why I've purposefully argued for a way to automate essentially everything but responding to a pest-related crisis - but only automation you actually have a reasonable expectation of control over. 

One of the things I was assuming in the watering system was that all the farm tiles that would be watered by that single sluice control would need the same amount of water - if you start putting different crops in the same farm that need different water levels, but all watering of a single farm is handled by a single sluice, and you can't give more water to one area than another without manual bucket-brigading, you've just destroyed the value of that water piping system you worked so hard to create.

There are several soil properties being tracked (6 if I count correctly). All of these continually change (depending on what you grow and what's happening to the soil) and affect the crops. Let me repeat it: you have six separate, continually changing variables for something as critical as farming. This sounds like trouble.

No, it sounds like the start of a system that is complex enough to be worthwhile.

It means the player will regularly have to check their farms to see if something important has changed...

No. 

These aren't random variables, they are variables you are in total control of.  The number of stones in your stockpile is a variable, but how often do you really have to check to see if you are running out of your layer stones for your stonecarvers?  Food is a variable, but how often do you really have to worry about running out of food for your cooks to prepare meals with in a fort with an established farm?

The ENTIRE POINT of these variables is that you will be able to predict how they change, and are able to plan ahead and make a manually programmed automated system that accounts for all these variables.

(And, of course, what Soralin says about fertilizers is exactly true - it should be easy to set up a system whereby the dwarves are scheduled to apply fertilizer to the soil up to the point where you need it.  Farmer skill may play a part in making sure they don't apply too little or too much, but generally, you can just have farming schedules that send out fertilizer at appropriate amounts.)

Translation: we have added a menial uninteresting task to the game that needs to be done regularly and can't be skipped. Awesome.

No.

Please, please, please read the things I have written.  You are apparently willingly ignoring everything I have spent the better part of a month arguing for just to be able to fight some sort of strawman.

OK BUT... the player can set the system once and then let it work indefinitely by itself, right? So they can, for example, spend 30 minutes browsing through the nutrient numbers, game wiki, farming tables, crop properties, etc. and after a lot of difficult calculations come up with a clever system that will keep all the six variables on desired levels without them drifting away in any direction... but only if they hadn't made any mistake in any of the million calculations.

Or you can do exactly what I set it up to do: manually play with the system early on to find out the system that works properly, and then, when a stable system is found through manual trial and error, simply set it to be repeated.  It's not that hard, and it doesn't take wiki crawling.  That's why I've been arguing for exactly that, in spite of what everyone seems to want to assume I'm arguing for, just because they can make me into something to argue against.

This leads my back to my farming wishes:

• I designate a field
• I turn on farming labours
• I tell the dwarves: "grow this crop"! (or better yet: "grow anything!")
  
• It works!

Anything else is just a bad design. 

No.

I want to worry over my farms.  There's no point to all this if it takes even less effort to farm than we already do.  A system where the player doesn't matter is just bad design.