New Map Design Thread

[jwilly]

The ratio between distance and height was surely kept.

For several design reasons, elevations are 1:1, and horizontal distances are compressed 2:1.

[AHWULF]

Coarse is 400 meters between xy points. The altitude is in meters. I was thinking that instead of just contour lines, which are harder to calculate, I could try a gridded approach, like sampling at a 20meter size or something, this could then be color coded as an overlay at a zoomed in map, so you can easily see the peaks and slopes. Just an experiment. Contours are not easy to read and because of the data may wind up being unusable with way too many curves in a small area.

3D visualization is not the goal. Right now collecting the data is the first step though.

I would love to build a map completely from data. I know where all the buildings are. The only missing data is rivers, roads, rr and coast/town boundaries. That information is encoded in textures so it becomes a problem in analyzing pixels. If I could automate that I would have the mesh and vector data for roads etc. With all that one could build a completely vector game map. But that's pretty distant.

[AHWULF]

Here is random supercell (likely not populated) drawn with contours with Mathematica. Even here it's pretty confusing.

[saintj2]

For several design reasons' date=' elevations are 1:1, and horizontal distances are compressed 2:1.[/quote']

I don't think so and for the reason YOU stated: if it was so, in game hills would be huge and very steep. We are talking about heightmaps ofc.

[jwilly]

I don't think so

As noted, the DEM data was only used as a starting point.

[jwilly]

The only missing data is rivers' date=' roads, rr and coast/town boundaries.[/quote']

I apologize if this information is way primitive relative to the level you're working on:

Those five geometry types, I think, are all parts of what I call the "tile designs"...the 400 m square "tiles", four of which make up an octet. (Forest blocks are another such geometry type.)

My understanding is that some of Doc's terrain creation work has consisted of coaxing the world creation tool to insert a tile from the tile-set into the world grid at the right location, with the right polar orientation, replacing a base tile that was there.

If you look at Five's maps, you note that the tileset isn't all that large. I don't recall the number as I type, but there are only something like 65 to 70 tile designs. And, if you superimpose a 400 m grid on Five's whole theater map-set, you can extract from that map the whole tileset...albeit at quite low resolution. But knowing which tile is where, I'd think it'd be possible to capture the tile content in 2-D via overhead "photography", and use this higher res information to extract the geometry information that you want.

I have the whole tileset (I think--based only on the above extraction process, with no check-input from CRS). Here's a map-piece of the north central Belgium area, showing a simple map-mod (but it'd still be a huge amount of Doc-work to implement, I'm sure) that I proposed a few years ago to add the Dyle River to the game. Historically, the entire Allied defensive intent was called the "Dyle Plan", and the extra-mobile British were to defend the "Gembloux Gap" between the Dyle and Meuse Rivers, so the game is a bit odd for the original design having omitted the Dyle entirely.

If you want, Ahwulf, I can send you a customized Five-map with the 400 m superimposed grid (provided to me by Five...full credit to him, all that I did was extract the low-res tileset for my own use), from which you can instantly see which tile is where.

Edited July 15, 2014 by jwilly

[AHWULF]

I now have the entire game world heightmap in a file. Now I have to stitch it together to make the mesh. It's a big file.

[SNIPER62]

I thought you were doing a map to put on a web page. Is the 3d heightmap just to make contours ? And do we really need contours ?

IMHO having a "simple" map like lagus in 1 month is better than having a super-map in 6 months.

i've got to agree on this one...

i'm assuming there is some sort of priority list that would go something like:

wiretap2

webmaps

contour maps

...right? it seems like there are other things that are more important that money was put up for rather than getting crude contour maps. get wiretap2 going then other players can access game world heights and can mess around with getting a contour map working while the webmap gets finished.

side projects dont help get wiretap2 and a webmap done

[AHWULF]

It's not up to me, currently my tunnel to the live databases is broken whilst other broken things are being fixed that are higher in priority like billing and csr. While I wait I am working with what I have which is the game data wherein resides the height maps.

Plus I now have the mesh so nothing more to do on that until later.

Edited July 15, 2014 by AHWULF

[saintj2]

It's not up to me, currently my tunnel to the live databases is broken whilst other broken things are being fixed that are higher in priority like billing and csr. While I wait I am working with what I have which is the game data wherein resides the height maps.

Plus I now have the mesh so nothing more to do on that until later.

Well, if you cant do anything else and if you have the heightmap (png files ?) you could just dump those on a 3d js library and get instant 3d terrain on a browser ... You'll still need layers or objects to make it useful (rivers, roads, towns, trees, ...) but it would be a start .

Well this is kind of a nerd thing - there was no 3D maps on WWII

[tomppeli2k]

is it for this site map or live game map?

btw... it should be look like ww2. could be cool to have marks like from a pencil or something what marks attacks and defends etc not too modern looking please :/ but still should be quite simple to read it

[AHWULF]

Site map, the in game map is in the client and is ... a massively ugly thing. I do want to make multiple kinds of maps eventually for different uses.

My VPN access is getting closer to functioning.

[AHWULF]

@jwilly

I have all the client data including the raw tile data. But if I want to create vector roads, rivers, etc, it's still a OCR type problem. There is no vector data for these things, it's a combo of textures (tiles) and prebuilt triangle data (think rivers with sides), none of which is all that easy to analyze.

The alternative is to build a time machine and go back and make them build it all from a database of vectors :-0

[jwilly]

Bitmap tilesets may be challenging to work into a vector map system, but they do have the advantage of being simple to extend for amateur map-concept development.

I think this three-water-depth Channel project took something like 25 additional tile-types, plus of course some postulated new tile parameters to affect object behaviors.

(If you happen to know the Channel, this probably is way off now...it was done from a mid-1940 map based on then-current charts. It obviously represents a mongrel set of depths because tides can't be modeled...I figured the more important characteristic was the channelization caused by the shallows. I figured it was good enough for my purpose.)

Edited July 25, 2014 by jwilly

[AHWULF]

Never talk to DOC about changing the map data this much :-)

[jwilly]

Oh, I'm well aware that changes/developments of these kinds even if CRS otherwise liked them would be practically impossible with the existing toolset.

[DOC]

Why not talk to me ? Is it because the past which I cannot undo and which I'd rather forget even happened, cannot be changed ? Because if it could, I'd change EVERYTHING.

[jwilly]

In regard to my Channel-map post above that was the reason for Ahwulf's comment, I think you've probably seen that work before...it was up in another venue, no longer available, at one point.

It however has been the case for a number of years, and appears to be even more so now, that there's no room on CRS's agenda for projects involving big map changes and maybe corresponding changes to server and client functionality.

So, stuff like the Channel map is just re-posted here to perhaps inspire other customers about the potential future.

Edited July 26, 2014 by jwilly

[genxs]

Hiya Ahwulf. Very nice to hear of progress an mapdata.

Looking forward to some kind of (expected) roadmap for development.

I have been a big user of the wiretap (so lots to update). Hoping that we can use raw feeds again due to use of nonbrowser scripts that wont process JS.

GL on the work there.

S!

Genxs

[saintj2]

Hiya Ahwulf. Very nice to hear of progress an mapdata.

Looking forward to some kind of (expected) roadmap for development.

I have been a big user of the wiretap (so lots to update). Hoping that we can use raw feeds again due to use of nonbrowser scripts that wont process JS.

GL on the work there.

S!

Genxs

I think he is using JSON (as if it was any better than XLM - pffft), but you can probably find JSON parsers for any language, almost as easy as XML parsers.

[SNIPER62]

I think he is using JSON (as if it was any better than XLM - pffft)' date=' but you can probably find JSON parsers for any language, almost as easy as XML parsers.[/quote']

yup

[jwilly]

There's not a lot of posting in this thread from Ahwulf, for news of whose work most of the thread participants are here, so in the meantime here's another one of those ideas-for-the-future maps:

Lower Scheldt Estuary, Modeled With Three Water Depths

(intended to support water-depth-dependent naval gameplay and ordnance)

(based on early-1940s charts and maps, simplified for no tides, practical tilesets and gameplay needs)

Edited August 2, 2014 by jwilly

[SNIPER62]

any update one this or wiretap2?

[AHWULF]

Sorry I haven't been on lately, I had my second and final surgery of the summer the end of July. Today I finally get the damn stitches and bandages off and can sit in front of a computer like a normal person again.

Assuming nothing else goes wrong internally I hope to have the connection setup to get the data on a regular basis from the live database to my map/data server and then I can start making the map actually update.

tldr - ouch lesser, code more

[jwilly]

Second = not good, final = definitely good.