Like one or two other users, I'm interested in using QGIS to map an imaginary flat world. A previous answer suggested simply refraining from defining a CRS for such a world, but I cannot seem to make that work well with QGIS 1.7 on Mac OS X. Pretty much every time I interact with a layer, the software warns me that no CRS is defined, and if I ever inadvertently open the CRS dialog for any reason, QGIS appears to apply the default CRS to my map. That happens quite often when I'm fumbling around looking for a way to find a setting that will establish the appropriate scale for distances on the map.

The fictional world is a flat plane of unknown (and possibly infinite) extent. The map is centered on a location that is locally taken to be the center of the world. Although the world is of unknown extent, the mapped area is about ten thousand kilometers north-to-south, and the same east-to-west.

I can map the world adequately using tools like Photoshop and Illustrator, but obviously they aren't GIS systems. I'd like to use QGIS for the myriad additional data-handling features it would provide, but I'm pretty well stumped by my inability so far to find or create a CRS --or to get QGIS to accept working without one.

I'm close to concluding that I'll have to write my own quasi-GIS system for this purpose that obeys the (admittedly idiosyncratic) rules of the fictional world.

  • This question is not substantially different from the others that you mentioned. Do you have a specific question? Jun 19 '12 at 1:46
  • My question may not be substantially different, but the other answers did not provide a solution I can use. As I've already said, I tried refraining from specifying a CRS with the result QGIS persistently and repeatedly presented requests for such a definition. My specific question is: how can I create a project that maps a world that is a flat plane ten thousand kilometers on a side with a datum in the center of the square, and avoid having QGIS 1.7 (on Mac OS X) dun me for a CRS definition every time I touch a layer? Jun 19 '12 at 3:20
  • A CRS is a important concept in GIS and hence QGIS. You are going to have a hard time getting QGIS to pretend like they don't exist.
    – Nathan W
    Jun 19 '12 at 3:40
  • That's fine and reasonable, but doesn't answer my question. It's also somewhat at odds with the advice given by a previous answer to a similar question (this one: gis.stackexchange.com/questions/20158/…) which suggested refraining from defining a CRS. Is it possible to write a set of parameters that tell QGIS that the world is a flat plane segment ten thousand kilometers on a side, with a datum in the center of the segment? If so, what do those parameters look like? Jun 19 '12 at 3:49
  • Ok if you go to Options > CRS > Use Project CRS you can then just create a new layer without a CRS and it will use the project CRS without asking which is nothing.
    – Nathan W
    Jun 19 '12 at 3:58

Simply tell the GIS that your coordinates are for an equidistant azimuthal projection, polar aspect (either pole will do). Employ a spherical datum (such as an authalic sphere. Just make sure never to ask the GIS to do "spherically correct" calculations or to "unproject" the coordinates, so that it sticks to Euclidean distance, area, and angle calculations.

One advantage of this projection is that the usual latitude-longitude graticule maps to a polar coordinate graticule. (Otherwise, you might just as well use any extensive projection, such as the world Mercator.)

GISes have a proclivity to clip this projection to a hemisphere. Thus, if you map parts of a world that are greater than 10,000 Km from the origin, you will likely need a workaround. One way would be to change your unit of measurement. E.g., interpret one meter (as reported by the GIS) as a "disc kilometer" and adjust all your coordinates accordingly: that will allow the GIS to handle distances out to 20,000,000 km, which will be enough room for a while :-).

  • Do they have degrees of latitude and longitude on the strange planet Flatworld? <grin>
    – nhopton
    Jun 19 '12 at 18:18
  • @nhopton That's actually a very interesting question. I would say yes: use polar coordinates. This is equivalent to unprojecting the equidistant azimuthal projection anyway (which is why this projection is so natural for a disc world). Beings living there, as Terry Pratchett aficionados well know, would automatically think in terms of "rimward" and "hubward" and would orient themselves according to an azimuth around the rim (much like Larry Niven's Ringworld people use spinward and antispinward).
    – whuber
    Jun 19 '12 at 18:25
  • Great; just what I was looking for, thanks. Thanks for the suggestions regarding measurement. Jun 20 '12 at 8:43
  • 1
    It should work very well; this world already uses idiosyncratic distance measures, and an idiosyncratic polar coordinate system. Jun 20 '12 at 8:52
  • @whuber, see gutenberg.org/files/201/201-h/201-h.htm :)
    – nhopton
    Aug 10 '12 at 11:01

If your imaginary world is going to be a simple, flat rectangle hanging in space then it really dosn't matter which CRS you use. Projections (CRSs) only become important when you want to map your data onto a real, roughly spherical world. As your 'world' is imaginary and flat this won't be a problem.

So any CRS will do, they are all simple, flat, xy grids just like a piece of flat graph paper. I'd go for something like World Mercator, EPSG:3395.

Good luck. Nick.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.