Arrakis.py map script discussion

[Ahriman]

Maybe. If we did it, it would probably be for a specific worker unit with a small national limit.

Even if we did it then though, how would the resource capture work? Building a harvester acted like a fort, providing your faction's influence on just that tile?

Partly though, this would put the entire maps' spice in play, which could be too much when its all being channeled through a single city for each faction.

I worry then that basically the entirety of a player's commerce income would end up coming from the spice corp, which would weaken the buildings in all the other cities, and the commerce producing improvements, and mess with the general commerce pacing.

I think its probably best to stick with just getting resources within borders, as current.

 

[davidlallen]

If we could make an AI to intelligently run some automated spice spice collection, similar to automated workers, would we want that? Wow, I think I've lost my fear of the AI code. Maybe we had better test my latest changes first.

Loop detection! I tried this in version 1.1.1 and there was negative feedback. Please read this post for my original proposal, this post for the release note with the feature, and this post for the retraction. Maybe you can think of a different approach, but it may be worthwhile to read that section of that thread before going too far.

 

[cephalo]

Loop detection! I tried this in version 1.1.1 and there was negative feedback. Please read this post for my original proposal, this post for the release note with the feature, and this post for the retraction. Maybe you can think of a different approach, but it may be worthwhile to read that section of that thread before going too far.

Thanks for pointing that out. I guess what I would like to see is the open dune areas to become a separate theatre with different rules of obtaining value and how you defend what you gain there.

I was thinking a harvestor would be like a carrier with ornis flying CAP or something. Battles in the dunes would be air based kinda like WWII Midway.

 

[Deliverator]

Wow, I think I've lost my fear of the AI code. Maybe we had better test my latest changes first.

I'm sure we have, and will have plenty of uses for your new AI expertise!

Do we have any other major AI issues or things it doesn't understand? If not, we're bound to run into some once we add religions and a few more custom units/mechanics.

 

[davidlallen]

Interesting AI related problems?

Fremen amphibious assaults.

Lousy initial city placement, and separate but related, lousy colony site placement. Hills are a lot more valuable than you think because of windtrap yield.

There are a half dozen unsolved implementation challenges in the religion design; see this post.

 

[koma13]

Do these minor Ergs have spice in them? I kinda wonder if we could at some point go to automated vehicle based spice collection, instead of culture borders spice collection. Kinda like Dune 2000. Or, maybe we could build culture spreading forts out on the dunes for spice collection purposes. Then I could make more Dunes!

I am not happy with the current spice mechanics too. No competition, deep desert feel like ocean. I tried multiple times to suggest changes, but all other team members of Dune Wars fanatically defend the current concept. My english language skills are too limited to convince them and I gave up. Good luck.

 

[koma13]

Lousy initial city placement, and separate but related, lousy colony site placement. Hills are a lot more valuable than you think because of windtrap yield.

The current CvPlayerAI::AI_foundValue reponsible for ai selecting city sides isn't working very well. It should reject all city sites without food bonuses and hills. We have to remove distance calculations that lower plot values. Coast land decisions doesn't make sense anymore (currently ai's 1st city in a new area have to be coastal).

 

[Deliverator]

Lousy initial city placement, and separate but related, lousy colony site placement.

I see the AI placing cities on bonus resources too often. This is probably due to poor general yield of the terrain. The trouble is if the found a city on Groundwater then that tile will only ever produce the default 3 water and not the higher potential yield it could have.

 

[davidlallen]

Not quite a big enough topic for its own thread, and related to terrain types.

In vanilla, you have jungle on top of many useful resources like banana. When you have enough tech, you can chop jungle and then build the improvement. You can actually order both at once, or individually; you can just order a plantation, and the worker will first chop and then build.

In DW 1.5.1, we have salt pan in the same location that other, more useful terrain would be. Salt is a terrain, not a feature like jungle.

So, how do we handle "chop salt", which is desalination? We want to remove the salt, and then build other improvements.

One way is to have a fake improvement, Desalinator, which never actually exists. But it can only be built on salt, and it gives a yield which is exactly the difference between salt and graben. Today it is +1H +1C. So the human or AI would build it; when the improvement is built, a couple of lines of python would swap the terrain type to graben. We would use a fake improvement with no graphics, which would be deleted, so no building type thing ever appears. After it is built, whatever improvements can be built on graben can then be built.

This is a little different than the automatic "chop + build" of vanilla jungle, but it may be close enough. The AI would build the desalinator, then discover that a bunch of new improvements. What do you think?

 

[davidlallen]

I am not happy with the current spice mechanics too. No competition, deep desert feel like ocean. I tried multiple times to suggest changes, but all other team members of Dune Wars fanatically defend the current concept. My english language skills are too limited to convince them and I gave up. Good luck.

Huh? Please re-suggest any improvements or alternative systems you have in mind.

 

[Deliverator]

Sounds like a good way to do it. The build button can just say "Desalinate Pan" or something like that.

We should do as Ahriman says and take the +1 hammer off Sink (formerly Hills) so that Sink/Salt is yieldless and Graben is +1H, +1C. Simple YieldInfos change.

 

[davidlallen]

Could you be in charge of collecting the improvement/bonus/terrain changes for 1.5.2? I want to work on redoing the terraforming implementation, so that I can add more onto it. It is a little too complex to make more changes in without a rewrite.

 

[Ahriman]

I am not happy with the current spice mechanics too. No competition, deep desert feel like ocean. I tried multiple times to suggest changes, but all other team members of Dune Wars fanatically defend the current concept

I don't think that human players would appreciate having to try to defend huge swathes of deep desert; I don't think it would be fun for human players even if the AI could be rewritten to contest areas without culture or cities, and if somehow resources could be harvested even outside cultural borders.

The competition for spice is cultural competition; the stronger your culture, the more territory you control (and the more spice you harvest), and when your culture hits enemy culture, the strength determines where the border is (ie how much of the spice each one of you gets).

The biggest problem with spice is not that you can't harvest spice outside your cultural borders.

The biggest problem with spice is that the only mechanic we have for actually using it is the house spice corporation, which concentrates all your spice commerce in one city.

We should investigate trying to graft corporation resource-consumption mechanics onto a building. If we could construct buildings that could, say, give +0.25 gold per spice resource you controlled, then spice would be helping to boost empire-wide economy, rather than just in one city (where it gets massively mulitplied by buildings).

This is a little different than the automatic "chop + build" of vanilla jungle, but it may be close enough. The AI would build the desalinator, then discover that a bunch of new improvements. What do you think?

Well, it was my idea, so unsurprisingly I support it

The current CvPlayerAI::AI_foundValue reponsible for ai selecting city sides isn't working very well. It should reject all city sites without food bonuses and hills. We have to remove distance calculations that lower plot values. Coast land decisions doesn't make sense anymore (currently ai's 1st city in a new area have to be coastal).

Agreed, better AI city selection needs to be a priority; it needs to put high value on water sources and bonuses (and no value on spice), and very little value on anything else.

 

[davidlallen]

The biggest problem with spice is that the only mechanic we have for actually using it is the house spice corporation, which concentrates all your spice commerce in one city.

I don't use corporations very often in vanilla. What happens when you spread a corporation to multiple cities? Does each city get its own copy of the full resource bonus? Or does each city get its own slice? I don't recall why we eliminated the corporation executives, which prevents the corporation from spreading to other cities.

 

[Ahriman]

Does each city get its own copy of the full resource bonus?

Yes.

I don't recall why we eliminated the corporation executives, which prevents the corporation from spreading to other cities.

This is why you eliminated corporation executives.

Weakening the corp and making it spreadable isn't the right solution.
The corp upkeep costs are also complex (and I don't fully understand them) and are a function of the number of cities the corp is present in and the number of resources being consumed.

Basically, we want a mechanic so that most of the income you get from spice depends on the number of spice resources you have, but you want to be able to have empire size matter a little too, and have the economy bonus spread around a little.

The main spice corp mechanic works well, but I'd also really like to see buildings or something that also gave some kind of benefit depending on how many spice resources you had.

 

[cephalo]

Ok, v 0.52 is up. I removed badlands and I opened things up a bit in the center and made the map just a tiny bit larger to compensate.

 

[Deliverator]

Could you be in charge of collecting the improvement/bonus/terrain changes for 1.5.2? I want to work on redoing the terraforming implementation, so that I can add more onto it. It is a little too complex to make more changes in without a rewrite.

Sure, can do. There are only a couple of things which sound like more than an XML fix. I'll post up the amended XML sometime in the next couple of days.

I still think we should use the Rivers graphic in the terraforming somehow. In Children of Dune, people are always walking along beside the qanat, or overlooking the qanat. So qanats are really like canals - open, flowing bodies of water. Try placing the river graphic down in World Builder - with some tweaks to the river graphics it could really capture the water flowing in the desert feel.

 

[Ahriman]

I would think that the little channels you get when you build an aqueduct (the art bringing the water from a hill or whatever) would probably be better than using rivers.

 

[Deliverator]

I would think that the little channels you get when you build an aqueduct (the art bringing the water from a hill or whatever) would probably be better than using rivers.

But that is much harder to implement - there is some crazy custom code controlling the layout of the Aqueduct.

 

[davidlallen]

wacky/possibly stupid idea: could we reskin *roads* and have the graphic automatically appear inside fresh water areas?