Glest Engine with arbitrary placement

[will]

It is my ambition to make a Glest-compatible engine that uses arbitrary maps rather than tile-based ones, e.g. play on a virtual planet or such.

I am busy coding prototypes, e.g. https://forum.megaglest.org/index.php?topic=6354.0 , yet I have not yet thought of a good project name!

Suggestions please.

[Gabbe]

colournoisy engine, because it had random colour maps and random height noise Wait, does this mean another fork? 3 engines massive..

[John.d.h]

I have not yet thought of a good project name!

Suggestions please.

I suggest picking something meaningless and arbitrary so you don't limit yourself.  For example, the Blender Open Movie projects are all named for fruit (orange, peach, durian) before they decide on a story.

[will]

Its currently at https://github.com/williame/GlestTools/tree/master/icoplay but then name will be needed eventually when it migrates to a C++ project

I've had some fun with doing ray intersection for the meshes, but a bit disappointed there isn't a nice R-tree in python that I've found

[Omega]

Hmm... name suggestions could get a bit spammy... I agree with John, pick any random, unrelated noun. A lesser known term might be more fun.

For example:
http://watchout4snakes.com/CreativityTools/RandomWord/RandomWordPlus.aspx

I think it's probably best we don't have people just spouting off random names in one word posts though...

Anyway, I wish you the best of luck. This would be a really cool thing to see in Glest.

Somewhat relatedly, my map wish list holds the ability to make a tileset independant map that contains its own models and information on the "walkability" of the cells. Allows infinite customization of maps, and stuff our maps can't currently do, such as true cliffs, city ruins, and various other organized things that can't be made from randomly rotated, single cell, limited number objects. Of course, that's super ambitious...

At any rate, I don't think this should be a separate ENGINE. What I'd like to see is a testing version for something that can be implemented as a GAE option in the future (not to take over maps, but to allow another type of map to be played, enhancing gameplay).

[will]

Yeah, arbitrary maps as in maps with their own g3d scenery embedded in them - say a map with Mt Rushmore or an active volcano or something - is perhaps more playable than the globe thing.

But battling on micro-planets - they have to be small enough such that your units can walk round the other side to fight the enemy in just a few minutes, and they have to be dominated by land rather than water - is a great way to validate the technical aspects of the engine itself.

And with that in mind, here's what I've got for auto-generated game worlds right now -



I'll leave the world creation at this point for now - doing good land generation is a distraction all in itself - and think about how to implement building placement and then unit pathfinding instead - in that order.

[Gabbe]

I think that the worlds must be bigger than that globe, either that or very small units & camera but then it would have sad quality.

[Mark]

So, would the maps be about the same size?  Because maybe bigger maps would be cool.

[will]

1) the globe maps would look like normal glest wheb zoomed in and playing, just that the world is a true sphere

2) I can't see why the flat maps people make to play with the engine can't be arbitrary size, limited only by the player's gpu

[will]

http://gamedev.stackexchange.com/questions/7161/what-library-for-octrees-or-kd-trees if anyone knows of anything

seeing that I plan to jump to C++ after I get pathfinding and building placement - building placement involves terraforming, will be fun - any insights to C++ libraries will be cool too

[will]

I got distracted with generating worlds again:

[titi]

Why do you think tile based worlds are not so good?

[Omega]

Why do you think tile based worlds are not so good?

They aren't, but that's no reason we can't offer both tile based and arbitrary. Each has its pros and cons.

Arbitrary
Pros:

• Can create the map as it should be (full control over looks).
• Unheeded by limitations in the map editor (ie: 10 objects, which are usually pre-defined).

Cons:

• Much, much harder and slower to make.
• More difficult to impliment.
• Larger filesize.
• Incompatible with tilesets.

Tiled
Pros:

• Easy to learn to use and create.
• Supports tilesets for more versatile look.
• Much smaller filesize (binary list of what each cell contains versus many 3D models).

Cons:

• Less support for innovation, such as true cliffs, city ruins, etc.
• Less control over the objects available.

Of course, I'm defining the arbitrary maps as maps we can model and tiled maps as our current, default maps. Now that I think of it, arbitrary has multiple meanings. I refer to the "Any and all possible" meaning, not the "random" meaning. Of course, should I be confusing myself, disregard me.

And yes, this is all very ambitious. I don't expect to really see arbitrary maps in the forseeable future, though I would certainly love to make some if they were ever possible... And as a final note, I don't think they'd replace tiled maps. The two work fine side by side, different options.

[will]

That was spot on.

Additionally, if I have a gripe about tile-based maps its that units align to them.  I rather want nice reeling groups of units moving fluidly in formation.

Ambitious?  Certainly?  Fun and challenging too though

[-Archmage-]

Why is this in the Off-Topic section? I'll move this to Maps, Tilesets and Scenarios for now, because there is simply no reason for this to be in the Off-Topic section.