Cartography Over Time (in different eras)

[Gabriel S]

Hi everyone!

I'm working on building a playable demo of an improved Civilization (video game) style map generator and viewer, and I wanted some feedback from folks who know a little bit about maps and video games.

First off, is our current planetary coordinate system (longitude by latitude by elevation) an ideal one, or is it a system that was created a long time ago and was just embedded enough that we never switched from it to the better systems? If there are better planetary coordinate systems, where could I read up on them?

I would like my Civ-type game to change the look of the world-view as the technology changes, starting off with the kinds of maps that ancient peoples used, and getting map upgrades as new technologies and methods become available. What are your favorite eras of cartography, and / or what map styles would you like to see in that style of video game?

Do any 4x game fans have annoyances with the current Civ-style maps and / or requests for improvements?

[Azélor]

I would like my Civ-type game to change the look of the world-view as the technology changes, starting off with the kinds of maps that ancient peoples used, and getting map upgrades as new technologies and methods become available. What are your favorite eras of cartography, and / or what map styles would you like to see in that style of video game?

Do you want to change the whole map or just the mini map or just change the textures?

[Gabriel S]

Do you want to change the whole map or just the mini map or just change the textures?

I'm talking about building a map generator and manager from scratch.

Generate a (spherical) planet.
Units can traverse the planet in any direction. No edge of the map.
The size of the planet, the ocean, and the distribution of continents could theoretically be used to simulate weather systems:

• areas of frequent typhoons and hurricanes
• doldrums in the oceans (where unmotorized ships can get "stuck")
• coastal area temperatures (and terrain) can be affected by cold and warm currents
• oceans can have more realistic resource distribution based on water temperature and currents
• add your own wishlist of things that can / should be calculated based on macro level terrain (oceans, continents, mountains)

Subdivide the planet into traversable terrain with a coordinate system.
Older games used square tiles with XY coordinates.
Newer 4X games (Civ style) use hex tiles with warped XYZ coordinates.
WarCraft / StarCraft / Command & Conquer style games switched from big square tiles to itty-bitty pixel-sized square tiles to simulate an untiled map. They also gave up turn-based play, however.

What kind of map coordinate systems have real (non-video-game) cartographers come up with to handle placement of objects on the surface of an entire planet? Is longitude / latitude / elevation an ideal system, or just a "good enough" system? (I'm thinking of the lines of longitude near the poles being so close together.)

Tile-based maps are simply easier to implement, but there's no reason a turn-based, overhead strategy game still needs to be using a map system based on limited computing power. What kind of mapping system would delight a cartographer to find in use inside a strategy video game?

[Azélor]

You could make the map as if it was large dice like this : http://en.wikipedia.org/wiki/Pentagonal_hexecontahedron
but probably with more sides

[Hai-Etlik]

For most global data sets when storing data for general purposes, we use spherical/topocentric/lat-lon coordinate systems. Sometimes geocentric/cartesian coordinates are used for certain specialized uses. Each projection is also a coordinate system (or getting really specific, each projection, when combined with a datum, a unit of measure, and an axis ordering) These may be used for specific regions or for special purposes. Various equal area, gnomonic, and conformal projections are used for specialized forms of storage/analysis.

A common coordinate system is UTM/UPS which is based on a combination of secant transverse mercator zones in north-south strips with the poles covered by polar stereographic zones. The coordinates then combine the zone you are in along with a position within that zone in metres. This can be combined with various datums, with WGS84 being the most common.

A datum is a model of the planet. Real cartography is generally based on a spheroidal datum (flattened sphere) used to approximate the geoid (an equipotential surface of the earth's gravitational field) The most common spheroid for modern geography is called GRS80. This has an equatorial radius of 6,378,137 m and a polar radius of 6,356,752.314,140,347 m. Different datums then move the centre or sometimes rotate it very slightly, and select a central meridian. WGS84 tries to position it so it matches the geoid over the whole surface while using greenwitch as the central meridian. NAD83 tries to match the geoid over North America. The vertical and horizontal components of a datum are often handled separately. GRS80 based datums generally have an error of 2 metres relative to one another. Other datums, particularly spherical ones can produce significant errors in comparison.

The European Petroleum Survey Group maintains a database of tens of thousands of coordinate systems which are frequently identified by their "EPSG Code". Anyone involved in GIS has memorized several of the important ones like EPSG:4326 (WGS84) and EPGS:3857 (Web Spherical Mercator). UTM involves a separate code for each zone, for each datum. EPSG:3157 is UTM Zone 10 North in a Canadian specific update of the NAD83 Datum.

The datums also receive periodic updates to account for improvements in the measurement of earth's gravity and for continental drift. New Zealand has special coordinate systems that take a time parameter to account for the comparatively rapid rotation of the Zealandia microcontinent.

If you want to do distance or geodetic bearing calculations globally, you have to use spheroidal mathematics. There's no coordinate system which will make it work as if euclidean that works outside of a restricted extent, and even then it's only approximate.

[Gabriel S]

A datum is a model of the planet. Real cartography is generally based on a spheroidal datum (flattened sphere) used to approximate the geoid (an equipotential surface of the earth's gravitational field) The most common spheroid for modern geography is called GRS80. This has an equatorial radius of 6,378,137 m and a polar radius of 6,356,752.314,140,347 m. Different datums then move the centre or sometimes rotate it very slightly, and select a central meridian. WGS84 tries to position it so it matches the geoid over the whole surface while using greenwitch as the central meridian. NAD83 tries to match the geoid over North America. The vertical and horizontal components of a datum are often handled separately. GRS80 based datums generally have an error of 2 metres relative to one another. Other datums, particularly spherical ones can produce significant errors in comparison.

The European Petroleum Survey Group maintains a database of tens of thousands of coordinate systems which are frequently identified by their "EPSG Code". Anyone involved in GIS has memorized several of the important ones like EPSG:4326 (WGS84) and EPGS:3857 (Web Spherical Mercator). UTM involves a separate code for each zone, for each datum. EPSG:3157 is UTM Zone 10 North in a Canadian specific update of the NAD83 Datum.

Thank you! This is a perfect place for me to start!

New Zealand has special coordinate systems that take a time parameter to account for the comparatively rapid rotation of the Zealandia microcontinent.

That is fascinating! I am off to bury myself in reading, thank you!

[johnvanvliet]

if you are coding a generator
you will need to decide on how it is done

a spherical subdivide in 3 dementinal space
( diamonds)
a OLD example of this is "Fractalplanet"
https://github.com/ldo/fracplanet

or
taking a 2d array out of a multy denominational dataset
( libnoise)
http://libnoise.sourceforge.net/index.html
single threaded and SLOW


also have a look at
http://www.crystalspace3d.org/main/Main_Page

there is also the OLD program "planet" by:Torben Æ. Mogensen
https://github.com/makhidkarun/planet

replacement for "libnoise"
https://code.google.com/p/accidental-noise-library/


found it
a article by some Stanford students
http://www-cs-students.stanford.edu/...ap-generation/
-- the code--
https://github.com/amitp/mapgen2

and that is just a start