[Mod] crops [crops] - (Farming|Food|Cooking)

[sofar]

Could you make plants get destroyed if the player walks on them?

That would be something I'd enable in the difficult mode. Are you sure you want this for "walking on" and not "jumped on" ?

I do not know whether the proper methods/event exist to make this work consistently, it may not be that easy to implement.

[Hybrid Dog]

Are you sure you want this for "walking on" and not "jumped on" ?

l thought about both.

I do not know whether the proper methods/event exist to make this work consistently, it may not be that easy to implement.

you could use globalstep or abm and retrace the path of the player (maybe http://dev.minetest.net/minetest.find_path) if the round position is not the previous round position

[Sokomine]

Could you make plants get destroyed if the player walks on them?

The Mineguild.de server had a mod that caused wheat to be trampled down. I believe it was called paths. I didn't like the mod at all. The paths created didn't look good, so we ended up digging them up and placing dirt again. Trampled down wheat was also very annoying as it offers new ways of griefing on a multiplayer server (in this case, mostly unintentional).

- I'm thinking of adding a hydrometer tool that can be used to determine the watering level of a plant. Using the hydrometer on a plant would display the watering level in the charge bar indicator of the tool.

Please do so. Else it might get too complicated. There are already plants in farming_plus where the growth stages are very difficult to distinguish due to the textures looking so similar (and no, strawberries don't turn red before they're ripe!).

- Also possible in this scheme are: bug infestations, diseases. I guess I also need to make bug spray.

A few bugs running around amongst the plants (there's a bugs mod out there) or as an animated texture could indeed be entertaining. Please offer biological alternatives to bug spray as well! Players could build an insect hotel to attract enemies of the plant eaters.

[sofar]

- I'm thinking of adding a hydrometer tool that can be used to determine the watering level of a plant. Using the hydrometer on a plant would display the watering level in the charge bar indicator of the tool.

Please do so. Else it might get too complicated. There are already plants in farming_plus where the growth stages are very difficult to distinguish due to the textures looking so similar (and no, strawberries don't turn red before they're ripe!).

Mostly done already. Here's a screenshot of how it looks after you use it on a plant that's at 50% humidity level (healthy):

- Also possible in this scheme are: bug infestations, diseases. I guess I also need to make bug spray.

A few bugs running around amongst the plants (there's a bugs mod out there) or as an animated texture could indeed be entertaining. Please offer biological alternatives to bug spray as well! Players could build an insect hotel to attract enemies of the plant eaters.

I'd do bugs as particles - flying around the plants, or crawling around on the plant. The benefit of particles is that I do not need to modify the plant textures. This is how I'm doing soaked/withering right now and it looks OK already with little effort and code.

A soaked plant:



A plant that's withering:



I'm gonna take a pass on the plant trampling idea. Not only do I think it'll be more than awkward to implement, the chance for grieving is large, and it's way too much damage to a plant all at once.

The humidity code works really, really slow instead. If you don't water your plants for an hour, most likely you just get a tomato less in the end, and the plant took longer to grow. Even if someone soaks your plant with a watering can, he likely helped it's growth more than hurt.

I'll write some detailed text on the mechanics of watering once I have all the code figured out. It's somewhat impactful and I'm still finalizing the integration into all the crops. Only tomatoes are affected right now, and beans / corn will be more complex as well due to their 2-block high nature.

[Hybrid Dog]

you can support protection for trampling because you know the player

[sofar]

you can support protection for trampling because you know the player

It's literally dead last on the list of things that I would like to see enabled in this mod.

If you'd write a patch or a better "paths" mod, I'm all in for taking your contribution, but it's open source ... I'd rather work on stuff for my sedimentology mod, or even my warps mod first.

Besides, I'm nowhere near done enabling the humidity code for all the plants yet. Plus all the plant stuff needs balancing and tuning (e.g. potatoes are growing way too fast and easy).

[Hybrid Dog]

By default farming only works if a player is waiting there until the plants are grown, else they're just grown a bit every time someone visits the chunk and the plant didn't grow one moment before. l guess you can also earn more cacti if you place 4 next to each other than making the grow interval 4 times faster.

[prestidigitator]

By default farming only works if a player is waiting there until the plants are grown, else they're just grown a bit every time someone visits the chunk and the plant didn't grow one moment before. l guess you can also earn more cacti if you place 4 next to each other than making the grow interval 4 times faster.

Yeah. I find this rather annoying, myself. It would be nice if mods like this (and the default farming mod) stored the last ABM execution time (either wall time using os.time() or world time using minetest.get_gametime()) in the node metadata, and did a bulk (statistical or deterministic; don't care) calculation based on the number of intervals since that time.

[Hybrid Dog]

The furnace already does this not.
What is http://dev.minetest.net/NodeTimerRef good for?

[sofar]

By default farming only works if a player is waiting there until the plants are grown, else they're just grown a bit every time someone visits the chunk and the plant didn't grow one moment before. l guess you can also earn more cacti if you place 4 next to each other than making the grow interval 4 times faster.

Yeah. I find this rather annoying, myself. It would be nice if mods like this (and the default farming mod) stored the last ABM execution time (either wall time using os.time() or world time using minetest.get_gametime()) in the node metadata, and did a bulk (statistical or deterministic; don't care) calculation based on the number of intervals since that time.

This needs to be solved on a global level, Doing this on a per-mod scale makes little sense. I personally do not think that this is really needed, and it has negative effects too, especially performance wise.

[prestidigitator]

This needs to be solved on a global level, Doing this on a per-mod scale makes little sense. I personally do not think that this is really needed, and it has negative effects too, especially performance wise.

Eh? Why? Load a single metadata value and check the time in each ABM call. That's not much, and actually it would make sense to decrease the rate of ABM execution if a more accurate time model were used. So it could easily be done with the same or possibly even less execution cost.

On the other hand, I CAN buy the argument that it should be done in the main farming mod if this one relies heavily on the API of that mod or something.

[sofar]

This needs to be solved on a global level, Doing this on a per-mod scale makes little sense. I personally do not think that this is really needed, and it has negative effects too, especially performance wise.

Eh? Why? Load a single metadata value and check the time in each ABM call. That's not much, and actually it would make sense to decrease the rate of ABM execution if a more accurate time model were used. So it could easily be done with the same or possibly even less execution cost.

On the other hand, I CAN buy the argument that it should be done in the main farming mod if this one relies heavily on the API of that mod or something.

You can't easily recreate the effect of what happens when you miss ABM's, since none of it is logically following strictly the idea that crops grow linearly. I mean, a furnace does exactly cook one piece of something per time unit. But plants grow randomly, they're affected by light, they're affected by nearby water, a player watering, and all these other factors.

None of that can be recreated afterwards. If you would come back 3 weeks after you planted to your farm, then all the crops from this mod would instantly wither and die, or at least end up in their least productive state (e.g. only 1-2 potatoes from a potato crop, normally 3-5) on normal difficulty settings.

Also, my argument was that this shouldn't be solved in the farming mod, it should be done at the global ABM level, otherwise you're just asking for the next farming mod to also solve this problem.

Of course, my sedimentology mod would be really, really interesting if this would be implemented - you most likely won't even recognize your landscape anymore, and millions of blocks would be moved or changed. Neat, but massively degrading the game play I think.

I still don't think that the solution is useful. For a furnace? Sure, it's lineair, it's entirely predictable. For a living thing, that has inherent randomness?

A good way to think of it is to ask yourself if you want chickens in the wild to lay eggs while you were not there. Are you sure you want to see thousands of eggs in the forest?

[sofar]

Three of the crops now require watering: Melons, tomatoes, potatoes.

In the default settings, the plants won't die, but the yield of the plant is reduced and you may end up getting almost no value for the plant (if you neglect it). The watering can is a pretty tough starting recipe, needing iron. The hydrometer isn't really needed, but with the hydrometer watering becomes an easy task, since you can see quite well if a plant is way too dry or too wet.

At night time the plants recover a bit of water. At day time they mostly lose water. Growing costs water! Each growing step costs 10 water. Growing a melon costs 25. The watering can adds 25 water per use to a plant. There's 20 charges in the watering can, so you don't need to refill that often. Of course, plants adjacent to water need less water.

Once I get beans and corn using the hydration system I'll cut a formal release I think.

[prestidigitator]

You can't easily recreate the effect of what happens when you miss ABM's, since none of it is logically following strictly the idea that crops grow linearly. I mean, a furnace does exactly cook one piece of something per time unit. But plants grow randomly, they're affected by light, they're affected by nearby water, a player watering, and all these other factors.

Not if there are other environmental factors, perhaps, but if everything is a function of the independent variable of time (which would make sense while a block isn't loaded), it's quite possible. For example, start with a simple Poisson distribution, and then factor in the day/night cycles if you like. Or simply run a number of loops equal to the number of ABM periods that would take place, modified by psuedo-random variables to account for the probability of the ABMs running.

[sofar]

Not if there are other environmental factors, perhaps, but if everything is a function of the independent variable of time (which would make sense while a block isn't loaded), it's quite possible. For example, start with a simple Poisson distribution, and then factor in the day/night cycles if you like. Or simply run a number of loops equal to the number of ABM periods that would take place, modified by psuedo-random variables to account for the probability of the ABMs running.

I'm sorry, you lost me entirely. Do you mind if I focus on the list of items to work on that I posted earlier?

Don't get me wrong, feature requests are good, but dragging on the discussion without contributing code only demotivates.

[sofar]

All 5 plants now support the hydration model

I made some per-plant tweaks to the scaling:

- corn likes dry soil, soaking damage chances are high
- beans are finicky, doesn't like wet or dry soil
- melons soak up a LOT of water when they grow a melon.
- tomatoes soak up plenty of water when they are in the harvesting phase
- potatoes are probably the easiest crop, don't soak too fast, don't wither fast.

Further notes:

- At all growing steps the plants consume more water
- growing slows down if a plant is soaked or dry
- if a plant becomes damaged, the crop yield shrinks... don't let your crops sit unattended!
- plants near water have it easier
- at night plants don't evaporate and therefore can replenish water themselves a little bit. So watch out that those crops standing near the water edge don't soak, which could happen even without watering.

[twoelk]

argh
this mod is way to beautiful.
it's consuming my spare time......wait got to water those melons.....

I am allready such a slow builder .... why are those stupid tomatoes not happy? .....

now I have to decide either farmer or builder ... aren't my beans a beauty? ....

hmm, got to dive into technic and pipeworks for some advancedwateringmachinetrickery.
wonder wether I can combine this with Hydroponics or even Aeroponics

... are my potatoes too close to water? ...

[sofar]

argh
this mod is way to beautiful.
it's consuming my spare time......wait got to water those melons.....

I am allready such a slow builder .... why are those stupid tomatoes not happy? .....

now I have to decide either farmer or builder ... aren't my beans a beauty? ....

hmm, got to dive into technic and pipeworks for some advancedwateringmachinetrickery.
wonder wether I can combine this with Hydroponics or even Aeroponics

... are my potatoes too close to water? ...

hahahaha, excellent post! :^)

It should be compatible without issues with Hydroponics, but there will not be any interaction with plants and mechanics from Aeroponics(farming_plus), the code doesn't handle unknown plants.

[twoelk]

oops forgot that mauvebic/peacock lost his code.
I was fiddeling with HybridDog's version as the old original version I had ran into problems and wasn't the last anyways.

[sofar]

oops forgot that mauvebic/peacock lost his code.
I was fiddeling with HybridDog's version as the old original version I had ran into problems and wasn't the last anyways.

The growlight will likely work on the plants in "crops", but the hydration system won't work for HybridDog's version. However, the code is mostly setup so that HybridDog could integrate support for the hydration system with a few changes:

- create a plant.properties table
- register the plants with crops (a table insert)
- calling crops.plant() instead of minetest.set_node()
- calling minetest.swap_node instead of set_node to retain meta values used by the hydration system
- handling crop damage at drops calculation time

some of this could be an easier API, but it's not a lot of code. Obviously the API can be improved with little effort, which I'll likely do myself.

[prestidigitator]

I'm sorry, you lost me entirely...discussion without contributing code only demotivates.

No problem. So, a Poisson distribution is useful for modeling how many times a random event with a regular frequency occurs in a given period. For example, if an ABM had a period of 10 seconds and a chance if one in 3, then in 5 minutes it would occur an average of 10 times. A random variable with a Poisson distribution having a parameter (lambda) of 10 would could give you how many times it actually occurred. In fact, we could use the same info above to simply calculate that the average growth period (in the right conditions) is 30 seconds. Whenever you update a plant, you can simply calculate the (possibly fractional) number of periods since the last update, and use that for lambda. Then the growth is basically independent of when the ABM actually runs.

Now of course you probably want to factor in things like light and water. So if you can estimate those things over time, you can use it to determine the amount of "good" growth time that has occurred. For example, you could use minetest.get_node_light() to help determine this, based on the current time of day, the length of the day/night cycle, and the amount of game time that's elapsed since the last update. Then obviously you don't care about more actual growth periods than the maximum number of growth stages of the plant, so this serves as an upper bound (max) on the output of the Poisson distribution.

Your phone or window isn't wide enough to display the code box. If it's a phone, try rotating it to landscape mode.

Code: Select all

local function poisson(lambda, max)
   lambda, max = tonumber(lambda), tonumber(max);
   if not lambda or not max or lambda <= 0 or max < 1 then return 0; end;
   max = floor(max);

   local pdf = math.exp(-lambda);
   local cdf = pdf;
   local u   = math.random();

   if u < cdf then return 0; end;
   for i = 1, max - 1 do
      pdf = pdf * lambda / i;
      cdf = cdf + pdf;
      if u < cdf then return i; end;
   end;

   return max;
end;

local function potentialGrowth(time, avgPeriod, stagesLeft)
   local lambda = time / avgPeriod;
   local actualStages = poisson(lambda, stagesLeft);
   if actualStages > 0 then realGrowth(actualStages); end;
end;


By calculating a "pseudo-time" instead of an actual time, you could model periods when the growth is slower or faster than usual, and stuff like that.

[sofar]

I'm sorry, you lost me entirely...discussion without contributing code only demotivates.

No problem. So, a Poisson distribution .................

I'm well aware of what a Poisson distribution is, and where you were going with it. This is also not the code contribution I had in mind.

Sorry, I consider the discussion on this topic closed. If you desire, open a github issue request and continue the conversation on this topic there (which I'll gladly do here). We can also talk actual code there, and I'd be happy to open a development branch.

[sofar]

Here are some screenshots of the particles when plants are soaking or withering:

Plants soaking:


Plants withering:

[sofar]

I'm doing some functionality testing on the current git tree. There were lots of little bugs and missed corner cases. Once this is done I'll do an official tagged release.

Testers welcome and appreciated.

[Hybrid Dog]

usually you can't see water as blue circles and withering doesn't make smoke

[sofar]

usually you can't see water as blue circles and withering doesn't make smoke

Sorry, what is your point? Are you suggesting a change? If so, what change do you suggest?

[twoelk]

usually you can't see water as blue circles and withering doesn't make smoke

why not just accept it as signal of what's happening like some manga or comic code

[Hybrid Dog]

usually you can't see water as blue circles and withering doesn't make smoke

Sorry, what is your point? Are you suggesting a change? If so, what change do you suggest?

l would like that the soil node which is too wet becomes a node with a mud texture. Particles sadly can't use transparency as far as l know.
And l think that a withered plant needs to be a different node that it's easier to see the difference to fresh plants. The particles may be insects which are flying around the withered plant.

usually you can't see water as blue circles and withering doesn't make smoke

why not just accept it as signal of what's happening like some manga or comic code

ln my opinion interpreting these signals may not always be easy because the withering plant still looks fresh.

[sofar]

l would like that the soil node which is too wet becomes a node with a mud texture. Particles sadly can't use transparency as far as l know.
And l think that a withered plant needs to be a different node that it's easier to see the difference to fresh plants. The particles may be insects which are flying around the withered plant.

Bug infestations are not the same as withering or over-hydration. As a matter of fact, I'm planning to add bug infestations as another source of plant damage.

Another damage source would be frostbite/freezing damage.

Particles are a simple way to show a plant is under stress. Having different textures for a plant under stress would cause lots of node updates, and actually introduces a lot of other problems.

e.g night time comes, plants recover a bit from drought, now I have to update all the plants back to normal textures -> lots of node changes, which are complex.

Also, just because a plant is low on water and showing signs of dehydration, it doesn't mean the plant is actually withering (taking damage). The visual clue that a plant is determining a bit of stress through particles is easy to understand. It's also clear that the plant itself isn't dead yet (on the contrary).

There is a dead plant texture for each plant, and that's what the plant turns into if you don't water your plants at all (and have the crops difficulty setting set to "difficult"). That texture means "the plant is dead" and you get nothing back from the plants (no seeds, no potatoes, etc.)

So, for various reasons, I understand that you don't like how this mechanic works because it's not realistic enough, but I'm not convinced that changing it will improve gameplay (not realism) and improve code.

why not just accept it as signal of what's happening like some manga or comic code

ln my opinion interpreting these signals may not always be easy because the withering plant still looks fresh.

You make a valid point. But, as I said above, if you see the withering particles, it doesn't mean the plant is actually taking damage. It's often just an indicator that damage may happen if the player doesn't act.

[prestidigitator]

I'm well aware of what a Poisson distribution is, and where you were going with it. This is also not the code contribution I had in mind.

Sorry, I consider the discussion on this topic closed. If you desire, open a github issue request and continue the conversation on this topic there (which I'll gladly do here). We can also talk actual code there, and I'd be happy to open a development branch.

Reviewing the mod code, it appears a bit too hard-coded to benefit much from changes like this anyway. I've submitted a pull request to the "Farming Redo" mod instead. Have a look at that for an example of how this can be done if you like, but otherwise I'm good with your, "discussion is closed."