Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[PTTG??]

Looking at the tiles, you can see there are boundaries between adjacent tiles where the meshes meet. The problem is that the meshes touch but don't connect; thus there's no interpolation that leads to the nice smooth effect, so the sharp angles are visible in the change in shine. What would make this look a lot better is if adjacent tiles could be aware of each other- if this can be done.

[Jay]

you can see there are boundaries between adjacent tiles

I don't see what you're getting at.
If you're talking about the actual models, that's a moot point, as the interpolation has to be done by the program itself, otherwise you'll only have interpolation for certain tiles being next to each other, and it won't always happen like that.
Or you'd have to actually have models for every single possible matchup of tiles being next to each other, which would be tedious.

[Vactor]

I know what he's talking about, its due to the normals of the models aren't being calculated at the seam, which is something i've been kinda hung up on, one of the solutions i'm considering is to have the program extract the verticies from the model, and put them all together into one vertex buffer, then have it draw polygons between them all.  This would clean up the draw method quite a bit as it will no longer have to run a draw for each tile, rather it would calculate it all once, then draw once each frame, and not have to calculate it until a tile is lost or added.

[PTTG??]

Sounds good to me. The only problem I forsee is what to do if tiles should meet but are off by a few percentage points.

...Or you'd have to actually have models for every single possible matchup of tiles being next to each other, which would be tedious...

Tell me about it. It's the best system, though, because you can add new tile shapes easily enough.

[Neonivek]

Wow, of the three projects Ive seen started up in this board... This one has lasted the longest! (well... longest with updates)

Congradulations Vector!

[Jay]

Congradulations Vector!

Congratulations to Vactor indeed.

[Yanlin]

All vectors to Vactor!

[Boksi]

Anyone else think Vactor sounds like some sort of latin title?

[Yanlin]

Ask Google.

http://www.vactor.com/home_2.asp

http://en.wikipedia.org/wiki/David_Van_Vactor

[Boksi]

According to Wikipedia, being a Vactor means he's not real! A Virtual ACTOR! A fake! OH NO!

[Vactor]

I'm going to change my direction a bit.  I'm not satisfied with the current limitiations of onscreen tiles (~100 and you start seeing slowdowns on my machine) So I'll be looking towards having viewing levels.  In the inter-airland view your airland with be abstracted to a certain degree, hopefully still conveying the shape and design of your airland.  (this should be relatively simple with low poly tiles without transitions)  The current work will be the airland level view, with which you can view a particular airland, and anything docked to it.  I'm thinking of implementing a closer tile view, where a tile in the airland view can be designated with a zoning type, which affects what kind of things can be built in the tile view mode.

(one of the current factors limiting this is the way in which the current view needs to be updated in order to move in real time in 3 dimensions, with this change only the most zoomed out view will move in 3 dimensions in real time)

The player would be able to designate a tile as a village tile, then zoom to that particular tile, where they can site particular buildings.  This will maintain the idea of having the player make decisions as to the layout of tiles in relation to eachother, while allowing a greater density of improvements, (rather than say having all of the assets of an empire being built on individual tiles)  The contents of adjacent tiles can have an effect on what can be built and sustained on particular tiles. 

I'm thinking I'll try to move towards tracking all individuals, however this may prove unworkable.  I am also toying with the idea of resource flows, (rather than stockpiling materials, you would build the infrastructure that would allow you to transport a greater amount of material, allowing projects with ample worker supply to be finished more quickly) however I think I may prefer to keep stockpiles, i'll be considering this further.

[Yanlin]

Hmm. It's not too late to change the entire design and story is it? You could just make it a futuristic game where you are the government of an airland. You see it from a "computer interface" where it can be represented by icons and grids. No fancy graphics required. Example: MMO tycoon. http://forums.tigsource.com/index.php?topic=1735.0

[Sowelu]

My honest first thought is "Wow, those mountains are too fancy, give them less polys".

[PTTG??]

Well, looking at what you're thinking of, you're probably going to want to have procedural maps- save the grid data and let the game generate the transitions and tiles on the fly... that's not as difficult as it sounds, and you can generate the detail to any depth you want- so one could zoom in on a single tile and zoom out to the world fluidly.
About villages; how about this system; you place a "Town Center" which allows you to give general town orders, and then the village grows around it. You can then place special structures such as smithies and barracks which join the town. Generalized homes and shops are condensed into a tile statistic like moisture level.
Give me a little while and I'll make a mock-up of the Proc. terrain.

[Virex]

If you're going procedual anyway, a way to cut down on the lag is to reduce the amounts of pollys and the resolution of the textures used by the tiles further away from you.

Also, seeing a slowdown with 100 tiles on the screen is odd. By the looks of it, you're using aproximately 20 pollys per tile. That would mean you'd experience lag with 2000 polies on the screen. That seems like quite a low number to induce lag. Also, are you sure you've got backface culling active?