Cornflakes_91 wrote:

ThymineC wrote:[stuff]

Albeit i still dont like the idea of having small cubic sectors i could live with this approach.
As it makes leasing space away pretty straightforward

(I would still like a full metaball approach more...)

Not just leasing, but feedback and possibly AI design too. We spoke about how Star Trek uses sectors as a coordinate system - we could use it similarly for LT here as well. If every sector has its own associated coordinates, it becomes very simple to implement a decent feedback system. Say that someone breaks a law somewhere within your territory - your computer could tell you something like "A pilot is unlawfully attacking other pilots within sector 88.129.91", which can also be highlighted on your map. With a pure metaball-approach, there is no convenient means of conveying feedback with this kind of precision - at best, you could say "A pilot is unlawfully attacking other pilots in the territory around station 5", which isn't terribly helpful if station 5's territory is large. On the other hand, the map could give just as helpful feedback either way.

Another thing is designing the AI to work with the system. The whole idea of zones was introduced in part because it simplifies the task of developing AI to properly understand the concept of territories and how to deal with the mechanics surrounding them, as Josh explained in a recent dev log. I think there's a fair likelihood that discretising space into sectors can further simplify the task of designing AI that can manage territory-related gameplay well.

One potential limitation that someone mentioned in IRC about a blocky grid-based approach is the difficulty of managing things within such a grid. However, this is something Josh will very likely have to address anyway based on his plans mentioned in a recent devlog:

JoshParnell wrote:Here's the truth: capability is created by allocating some portion of an object's total mass & value towards it! Yes. This makes so much sense. What does it mean? Well, a ship has some total volume. You can choose to devote that volume to cargo storage, or to a tech lab, or to a production facility, etc. Do not think of capability as just coming from an object, but think of it as coming from space itself.

Since spaceships are three-dimensional, it's very likely that Josh would need to develop a 3-D grid structure in which you could manage each individual volume of space to specify whether that volume constitutes part of the ship's cargohold, production lab or research lab. If this is something Josh needs to resolve, it's fair to assume that the same approach as he ends up taking for internal ship design can be applied to managing individual sectors within a 3-D blocky grid space.

Edit: Although now that I think about it, there are many easy (but less fun) text or slider-based ways to specify the allocation of different parts of the vessel to different things. I really do hope it's based on a blocky grid approach though, so that we get something like FTL except 3D and with the ability to control where everything is placed and in what sizes:

AFAIK Josh does not intend to use volume in any aspect of ship management. "Capability" is an abstraction he has invented that doesn't relate to volume.

Space also probably won't be divided into cubes. The zones are likely equivalent to spheres of influence, where they are defined by origin, radius and type.
The are not homogeonous throughout space. They are specifically for points of interest.

A point of interest is anything that is recognisably different to the rest of a system's space. This could be a space station, a planet, a jump gate, a wormhole, a patch of asteroids with a different density/type distribution, etc.

Josh has not clarified how systems are owned. I would assume an agent owns a system when they own some to-be-decided percentage of the points of interest in that system.

Spheres of influence isn't unreasonable. It's how I implement control in my space game prototype, in which zones of control can overlap.

But I'd think that makes it harder to satisfy what I've been told is Josh's requirement that zones of control don't overlap. If what you have are spheres defined by a radius, then you have to special-case every volume where they would have intersected.

You don't have to do that if space is composed of smaller chunks. Zones are composed of chunks, each of which is tagged with the ID of the owning faction and the strength of that faction's control. For the cost of tracking multiple chunks of space, you don't have to worry about overlaps.

I have no idea which direction Josh might go on this, but I'm looking forward to finding out.

Control could be generalized for all spherical zones, even if they intersect.

Give each zone a "control level" and "controller" So a zone or an intersection is controlled by "controller" at a value of <control level/sum total of intersecting spheres' control level>

It is apparent that an agent that controls an isolated sphere with a control level of 100 has 100/100 control.
Say another agent sets up a nearby station that has its own zone with a control level of 80, where the spheres intersect, Agent 1 has a control level of 100/180, Agent 2 has a control level of 80/180.

And so on for more intersecting spheres. Deciding how much percentage of control level an agent needs before they own a zone is a balancing issue. IMO, it should be something like 50%. If all control levels are less than 50%, the zone would be 'contested' or 'neutral'

Note that these are my ideas on zone control, not Josh's implementation ^_^

ThymineC wrote:What I'm more interested in discussing is how space is partitioned with regards to ownership and territorial management. Cornflakes is adamant that we use a metaball-approach, and many others support him in this. I'm quite adamant that we partition space into discrete equally-sized cubic volumes, each of which are called "sectors". Quite like this, but at a much smaller scale. Ownership would be generated by certain kinds of structures, military presence, etc. and spread (in an attenuating fashion) to adjacent sectors, producing what is effectively a voxel-like metaball territorial region.

Seems overkill to me since most of space is a useless and empty void.
There is no point owning it so there is no point recognising it as a zone.

Metaballs look pretty but are they even useful here?

IMO, a zone can be a simple sphere.
These spheres can overlap and have completely different meanings. There can be more than one type of ownership.
Maybe the owner of the entire sector "owns" a huge sphere that covers the whole sector. Within these boundaries you are subject to their laws... until you fulfill the requirements for "conquering the sector".
Then this sphere is controlled by you.

A smaller sphere inside that could be patrolled by a private security company, protecting a space station.
You are still within the governing larger sphere but the local security types have the force of law... provided it doesn't contradict the sector-wide law.
They can order you to stop and be searched for contraband, for instance.

How much independence such a local sphere of control has... is a matter of bargaining. =)
The space station "Gazz' Super Nova" may offers the services of a Brothel & Grill in exchange for rum being not illegal within it's zone of control.
Naturally this may end up being a hideout of smugglers but that's someone else's problem.

Intersecting and partial control of this or that just muddies things up. A hierarchy of zones is easier to understand and display.

Different types of control may overlap each other just fine.
Like a mining company "owning" the extraction rights in an asteroid fields but not having a police force itself.
It is supported by the sector's police with illegal mining being defined by mining within this sphere.


I am wondering if ships could have their own zone of control and what that would imply.
A mobile mining ship could carry it's own bubble of exclusive mining rights... for a price?
If a zone can be dynamic, it can be mobile...

thymine:
Why do metaballs rule out an orthogonal coordinate system? O.o

i assume zones spherical and continous because of joshs notion that they are intended for AI usage.

The AI does not need "cube 132.548.32 with 436 points of czerka corporation ownership"

It does need "area around czerka corporatio station 2"

With your cubes approach you'd have to organise the cubes in zones sccording to some sorting mechanism, that the AI knows that sector 132.548.32 is in fact in the area around czerka corporation station 2.
Which is a free meal with metaballs and "nested zones" as i already explained

Gazz:
The metaballs are for merging zones which are next to each other, so for example when 2 stations get next to each other the zones merge and get one bigger zone.
No special case for overlapping zones, no special case for concurrent factions being in the same zone. Just re-evaluate relative "ownership" strenght of the zones' inhabitants.

Gazz: I far prefer your heirarchial zones than my intersecting ones. They make far more sense

Thanks, Thymine, for writing it all out. A typically thorough job.

There are one or two areas in which the concept in my head differed from what Thymine laid out. I was thinking about establishing a system "ownership" mechanic. This means that once some metric about the system has been reached by a player/faction, the system is designated as theirs (this would be a "controlled" system; other system states would be "open" and "contested" which I think are fairly self-explanatory).

The key thing I mooted was that prior to system ownership being established, it's all non-overlapping spheres of interest. These spheres can grow - most obviously if you build trade lanes - and if two spheres owned by the same faction come close enough they will touch, as per Cornflakes' "metaball" concept. Two spheres owned by different factions will produce something that looks a little like Endless Space, but in 3D. This is nice because it can make contested systems interesting pretty quickly, and also gives a solid reason to build more infrastructure.

After system ownership is established, then the owning faction at that point becomes a defacto government, and can start parcelling out space for commercial or military purposes.This would allow other factions to build infrastructure in already owned space, and allow for the mining licenses that Thymine illustrated.

Now there really isn't a nailed-on reason why you need a grid system to do this, but it just seems reasonable. Tessellating shapes make more sense when you're establishing some sort of regulatory system (as shown by Thymine's link to the UK offshore sector). Really though, I just think it would look good and be a visible way to differentiate between "open/contested" systems and one that is "controlled".

There was some discussion about how to subdivide "metaballs" into sectors. I didn't see that as necessary - my view is that once the system is "owned" then a 3D grid is drawn around all areas of interest in the system and the owning faction can do what they want with that space.

Finally there was some good chat about how a system that was owned may revert back to a contested state. The most obvious scenario would be military action, since a "controlled" system is analogous to a country in some ways. However I also wondered if there we should establish some level of "influence" that the faction needs to maintain to retain ownership of the system? I'm using "influence" here as the sum total of their infrastructure and ships a system. If their influence was X when the system became theirs, then I believe that some large fraction of X should be needed to retain the system. This means that factions going broke fighting a war in a far-off system may lose control of some other system unless they're careful.

Phew. It was a productive discussion on IRC last night!

Always interesting ideas in here !

Do you think Josh has considered using SCALAR FIIEEEEEEEEEEEEELDD to represent ownership?

Mcsven:
I dont see the reason to throw away metaballs once the system is under ownership.

I see the cube-allocation more of the regulation of "you are allowed to work in this cube and your metaball space has to stay in this cube(s)"
not the complete substitution of metaball space, but limitation.

The problem I see with dividing space into a grid is that it defeats the purpose of having zones around points of interest.

If space is going to be divided into a grid, all of space must be divided into a grid. Otherwise you just have isolated cubes around points of interest instead of spheres. (Spheres are easier to implement)

I also think dividing space into a 3D grid would look a bit messy. Too many lines (personal opinion)
I also think that if space was divided into a 3D grid, in a heavily contested system you'd see a whole bunch of differently coloured semi transparent cubes stacked around each other. IMO, Isolated spheres would look better

Cornflakes_91 wrote:The metaballs are for merging zones which are next to each other, so for example when 2 stations get next to each other the zones merge and get one bigger zone.
No special case for overlapping zones, no special case for concurrent factions being in the same zone. Just re-evaluate relative "ownership" strenght of the zones' inhabitants.

Not needed, either, because 2 spheres do the same thing. =)

IMO, overlapping of spheres should only be allowed if they have different meanings.

Conflicting spheres (property of Blue vs property of Red) should not. You are simply in the sphere of whoever is closer or in case of different "strenghts", in the sphere that has the most influence on this point.
It's the principle of metaballs - only simpler. =)
If you want to get rid of the sphere of influence that is encroaching on your turf you have to get rid of the space station or whatever that is generating it.
(see GalCiv cultural warfare =)

Gazz wrote:

Cornflakes_91 wrote:The metaballs are for merging zones which are next to each other, so for example when 2 stations get next to each other the zones merge and get one bigger zone.
No special case for overlapping zones, no special case for concurrent factions being in the same zone. Just re-evaluate relative "ownership" strenght of the zones' inhabitants.

Not needed, either, because 2 spheres do the same thing. =)

IMO, overlapping of spheres should only be allowed if they have different meanings.

Conflicting spheres (property of Blue vs property of Red) should not. You are simply in the sphere of whoever is closer or in case of different "strenghts", in the sphere that has the most influence on this point.
It's the principle of metaballs - only simpler. =)
If you want to get rid of the sphere of influence that is encroaching on your turf you have to get rid of the space station or whatever that is generating it.
(see GalCiv cultural warfare =)

in your concept you'd get a very mess of overlapping spheres pretty fast if a couple of people decide to build in the general area.

You could also never enforce your influence if all the other has to do is to build an inch outwards to get out of your jurisdiction. Which would devaluate anything resembling police very fast.

Im also try to keep in line with the primary reason for zones i read out of the devlog: AI
With metaballs close zones merge into one, that can be handled as a single object.
without it you just have a conglomerate of influences superimposed on each other.

(I imagine this right now as a bunch of legos: metaballs stick them together to something coherent, superimposed spheres just get a bigger pile of legos, not something greater)

ThymineC wrote:

Cornflakes wrote:Once space has content a zone gets created, the space around it is still no zone.
So when you build a spacestation gets built a zone with your spacestation at the center gets created (and maybe named by some means).

The issue with this is that it can be hard to determine where one zone ends and another zone begins, if say I were to build a space station close to you.

For thinly populated systems, I think it would be plausible to have a lot of "no man's land". Analog to the traditional three-mile territorial zone at sea, you would have some "bubbles" of ownership with lot of international waters, ^h^h^h^h^h^space in between.

When two stations are closer than twice the normal radius of a zone, there would be some compromise, as in this picture:
http://graphics.ucsd.edu/~iman/SoapBubb ... _3_RGB.png

Yup. Dissimilar metaballs would just flatten when they met. No grids needed.