You didn't exactly write a brisk read, yourself, you know
People report massive FPS spikes because of changes such as stone temperature with magma or constant water flow. I can't help but cringe when I think of my computer trying to chew another enormous set of variables for every square of soil (which is A LOT of squares). The problem is not so much that it will eat up memory. Rather, it will consume too much processor time. The game would constantly have to check a whole set of attributes on every soil tile on the map. That translates into a big slowdown.
A fairly common concern.
Note that temperature slowdown occurs because
every tile is calculated for heat. If you have a 4x4 embark (each embark tile being 48x48 regular tiles) with 200 z-levels, then you have 4 * 48 * 4 * 48 * 200 = 7,372,800 tiles in an embark. Cycled through every temperature check.
Now then, the soil system I am proposing makes soil a "floor covering", and in that same area, would be 147,496 tiles to check (and I further proposed conglomerating embark tiles worth of wild soil together at that, which would significantly reduce this number down to about 16 plus however many plots the player makes, which is not altogether dissimilar to what you just said in "tracked by layer", anyway). These are performed on growdur cycles, which occur every 100 normal frames, further reducing the number of calculations taking place by two orders of magnitude.
Many of these calculations are not terribly complex, at that, and this is thanks in large part to keeping much of the system as an abstracted system outside the sight of the player. You can fudge around nutrient values between soil tiles near one another if the player never can measure the exact amount of phosphorous that last tuft of grass took from that one tile.
The changes you mentioned are awesome mainly because I really like the idea of limiting resources. However, I think that could really be accomplished without dramatically overhauling the soil system. Just make seeds more scarce. If each plant only had a % chance of producing a seed when processed, that would end the whole "infinite agriculture" problem. The % chance of getting a seed could be influenced by dwarven skill, whether the plot was fertilized, or what-have-you. This would be the way to re-implement some of the ideas you mentioned.
Overall, I don't know if your suggestion solves the original problem of the game being too easy. Even if crop yields are reduced greatly, it would still be very easy to thrive. First of all, you could just move your farms every few years to more fertile land. Second, with the huge availability of meats and fish, you really don't need to eat plants at all. Of course, you still need booze, but if you use all your ag resources to grow booze crops, you'd never run short of the stuff.
There is more than just the problem of the game being "too easy", it's also a problem of the game being "too shallow". DF has complexity, but not very much depth. I like to compare it to the creatures in 40d - they had plenty of stats on them, but really, all that mattered in a creature's stats was their size. Right now, there's very little plant diversity because there's only one or two metrics you can really even judge a plant on - the rate at which you get food, and the value of whatever other use you might be able to get out of that plant. Nobody wants hide root because it's just a silver barb or a dimple cup, but worse.
We can make the game much more deep and involved than it currently is, but it requires that we actually add into the game the metrics by which we can actually differentiate one plant from another.
Just changing seeds, or just changing how many dwarves work the fields doesn't address this problem. Plants are already easy enough to care for that if you give them only one concern - getting enough seeds to keep planting them - then you're just changing the decision metric to still be only about judging how best to solve one simple math problem - what gives me the best probability of getting seeds?
I'm going to post
this link to Extra Credits' monologue on choice again, although you probably didn't see it the first time, anyway, so it's all good. The key to setting up a challenging or difficult choice is that you need to make the player want to further at least two mutually exclusive goals at the same time. If all your choices are based upon just one metric (having enough food, based entirely upon having enough seeds), then you can't make a player stop and choose, it's just a matter of solving the math problem of "what course of action will get me the most seeds, and hence, secure my food supply?" There's nothing much in that decision tree that gives a player some sort of conflicting desire to do anything else, except maybe some players' desire to spend as little time thinking about a farming system as is absolutely mandatory.
In order to set up conflict in the choices, I have involved several ways in which the player's innate desire (get as much valuable stuff as possible) are in conflict with other desires the player has. (The fear of an environmental backlash from overexploitation of the land, long-term degradation of the land versus short-term boosts in productivity, the threat of pests vs. the potential to grow greater numbers of more valuable crops, the choice between different forms of biological resources if trees are put on farms to compete with food crops, as are very tedious to grow but powerful alchemical ingredients when I get around to my next suggestion on alchemy, and of course the simple amount of player time and dwarf labor it takes to design and manage systems of ever-increasing complexity.)
Do you not believe that these are sufficient conflicting desires? I am always open to more suggestions on how to set up a conflict of player choice, but at the same time, you need to make sure it actually does this in a way that sets up a conflict of choice, and is not just a way to try to make the game "harder" without really adding to the depth of choice. If all you have is one calculation to make on what is the best way to get enough seeds to sustain your farm, then once you have made that calculation, and make it every time, then the game is just as "easy" as it was before you made it "harder" - the problem is solved, and there is nothing else to think about.
Here are some suggestions of my own:
1) Slash n Burn
This is somewhat similar to the way that we were discussing making Plump Helmets grow (by cutting down logs, and letting plump helmets grow on the logs in cool, moist caverns). The specific idea that burning the tree on the spot as a means of producing potash, however, is a bit new. I'll try to include that in my fertilizers section when I get to writing it.
2) Rebalanced meat industry
I have talked extensively about this in several other threads, including the previous thread, and believe that the major solution to this is to let animals produce stupidly large amounts of meat, but to make dwarves eat more (and crops produce more food).
If cows need to eat grass, and produce less meat than turning that field into a corn field would have produced, then you have essentially eliminated the problem of livestock producing more meat. To make animals not worthless, those cows could eat grass from more marginal land, and their manure could help in the reclamation of soil fertility. Herd size would need to be maintained at a level balanced to the fields you are capable of providing for their feeding.
If elephants need to eat a stack of 50 units of plant matter at a time, 250 in a year, then suddenly their 200 units of food when you slaughter them seems much less enticing.
Work continues on the Interface section. This is taking far too long, but some of these decisions are much harder than I would really think they should be, in terms of deciding how to actually give players the buttons to actually push to make the things happen.
I expect this to be my most controversial section, since scheduling is still a bit on the obtuse side, even considering how long I have spent trying to find a way to reduce it down.