Now that we have both the Outer World and the Hollow World close to being locked in, it’s time to go back to the 3D model and explore the results of our labours.
But first, let’s take care of a couple of loose ends.
Companion Set Brun map
The Companion Set included a continental map of Brun before the Master Set presented the whole world. It’s significant for two reasons: it presents a more detailed coastline for the continent, and it adds a few small details that we can use.
I have adapted the coastline to my Outer World model by using three different placements: one for northwestern Brun, from top of the Arm of the Immortals all the way over to close to Frosthaven; one for the Arm of the Immortals itself; and finally one for Norwold that extends down to Kamminer Bay. This last one showed part of the Orol Island archipelago, with more detail than the Master Set, so I have decided to change my base reference for these islands from Skothar to Norwold, with the Companion Set map deciding the location for the islands:
The other details are in Ochalea, which showed a small island off its southwestern coast, and the two small islands between Dwair and Davania:
All of these have now been locked in, to be incorporated into future maps.
Hollow World Set Outer World map
Next up, let’s put the Hollow World Set world map into place. The fit is not perfect, but it’s more than enough to act as a nice guide to expand the world map’s locations:
Looking good.
And that’s us done! There are very few sources left to add to either this or the Hollow World map. That means that all we have left to do is decide which placement to use, create a 3D model for it, and take it for a test drive. Once that’s done to our satisfaction, the Lining Up Mystara project will officially be over! So close now…
Looking Ahead for a Moment
A word on secondary sources (i.e. fan-made maps): these will come into play in full force during my next project, when I come to deal with all of the stuff this project has generated, and expand the maps to cover the whole world. My ideas have changed over the years somewhat, and my current feeling is that it’s better to adapt (or even adopt) fan-made maps than for me to come up with yet another version of something. Let’s face it, whatever we do it will never be the “official” version, because there was, is and will likely never be an “official” version of these things.
Besides, who cares? This has never been a big issue in our community, and it has been fading more and more into the background, I think.
So my current plan is to make up terrain based on my above collage where there is nothing else to refer to, and to adapt fan-made maps where they exist — and simply adopt them wholesale if they fit into the above map nicely. There will undoubtedly be some exceptions, and I’ll likely avoid using some maps if I already have my own strong ideas about an area, but in general I’ll try to choose the most accepted sources and work from them.
One last thing about this: I hadn’t planned to do the world map in 72 miles per hex, but looking at the sheer size of this map, I’m now reconsidering this. Would you like to see me do a full world map for the Outer World at 72 miles per hex? If I do it, it won’t be an Equirectangular projection, but rather something much more appropriate for hexes — likely either an equal area projection of some sort (to keep the areas of each hex constant) or else a conformal projection (to keep the shapes — but not the relative scales — of terrain constant).
I can’t talk timescales yet. Once Lining Up Mystara is all wrapped up, I have a number of other projects waiting in the wings, including of course Bruce Heard’s upcoming Calidar book. So all I can say for now is that I hope to start the next project sometime later in the year.
Up Next: Placing the finished maps on the 3D model. Will Placement 5 do the job? You’ll have to wait and see!
In which I slice and dice the Hollow World to get the straightened out map to fit the hex maps…
More Size Issues
There’s one more issue here that needs to be tackled: the size of the Hollow World. Back in 2012, when we were talking about World Dimensions at The Piazza, both Sheldon Morris and I came to roughly the same size for the Hollow World, based on expanding the hex grid from Iciria to the rest of the map, and counting hexes. My count was 124 or 125 hexes, as I have mentioned in previous articles. But let’s look again: what was I counting?
The Hollow World map has latitudes marked just as the Outer World did. There’s something crucial to notice here: between 60ºN and 60ºS there are 23 hexes for each 30º of latitude. 60ºN to 90ºN and 60ºS to 90ºS only consist of 16 hexes for 30 degrees of latitude.
But if we take the hex map at face value, we can at least conclude that the north-south dimensions of the central meridian provide an accurate measurement of the extents of the Hollow World — at least, the extents of the part up until the polar openings.
This is all getting a bit complicated, but here’s my conclusion: the vertical height of this map is the accurate measurement, while the latitudes are a red herring that we can and should ignore.
So what needs to be done? Well, we need to unwarp the map — straighten it out so that north is always straight up, south straight down. But we don’t need to map it to the grid of squares an Equirectangular projection will give us. Look again at my Equirectangular map above. See how all of the squares are regular throughout the map? But as we have seen, the hex map dictates that the areas from 60-90º are in fact smaller than the rest of the map.
Here’s what I propose: we can still use my reprojection to straighten out the map, but it needs to be cut up and rescaled vertically, so that the latitude lines match the pseudo-Robinson projection’s lines. This will result in an image like this:
Here’s a comparison of the two. Note how the latitude lines match up.
This image is no longer 2:1. It’s now 1925:1000 — in other words, a little too short in the east-west dimension.
This makes me remember that when we fitted the world map to Iciria’s hex map, we needed to stretch it slightly east-west. Since we’re reprojecting based on the graticule, the stretching will be discarded in the reprojected image. Does this mean that we need to stretch the map out a bit to get it the right east-west dimensions again?
Well, the original hex map’s east-west dimensions should be true at least along the equator, so we can use this as a check.
In simple terms, we want the reprojected hex map to match the original (pseudo-Robinson) hex map along both the equator and the central meridian.
Let’s see with an overlay:
It’s pretty obvious right away that they don’t match.
The latitudes will of course need to be rescaled manually as we did for the Equirectangular world map.
Note how the central column of hexes is quite regular now that the latitude lines have been matched up. Remember, we’re ignoring latitude and going with the hexes as an absolute measure of distance north-south.
But what of the longitudes? Shouldn’t they have been accurate along the equator?
Apparently not. Looking at other Robinson projections, they do seem to be taller than they are wide.
Conclusion: the Hollow World hex maps are probably all slightly wider areas than they appeared to be! This is because the Robinson projection is a compromise projection, which doesn’t present any single aspect (shape, area, distance, etc.) of a map completely accurately.
But wait: there’s more. It turns out that the pseudo-Robinson world map’s western (right) hemisphere is significantly wider on the map than the eastern (left) hemisphere. Check out the width of the grid squares on the left and right, and it’s quite obvious that the right is bigger. This all gets normalised when you reproject, but I propose that we match the longitude graticule in the same way we matched the latitude lines. Specifically, we should match the junctions at the equator, restoring the visual appearance of the map. It may be necessary to add some sea on the far side of the world in order to keep the central meridian where it is, but this is fine by me — I’ve always thought that the outlying continents are too close anyway.
Here’s what we end up with:
The resulting world is now 1967:1000. Closer to the 2:1 ratio than before, but not quite there. Let’s fix that by centring the central meridian and adding sea to the blank space:
This looks pretty good. It would be easy to add some islands to make sure the equatorial mountain range isn’t completely broken.
The ratio is now 2031:1000. So we’ve passed the important 2:1 ratio. More on this shortly.
Building an Equirectangular Hollow World
This may have been hard to follow, so I’ll sum up. We now have:
An Equirectangular projection world map, scaled to fit the visual dimensions of the pseudo-Robinson world map. The graticule grid is now totally wrong, but we can fix that in future maps.
An Equirectangular projection Iciria hex map, scaled to fit the north-south visual dimensions of the pseudo-Robinson Iciria hex map.
It’s time to put these together.
If you remember, this involved stretching of the world map last time. We’ll do the same this time, adapting the whole world map to the hex map’s dimensions. Here’s the resulting world:
It’s reached a ratio of 2234:1000. Add a graticule and fill in the north and south extremes with sea, and here’s what we get:
It’s messy due to the old graticule being burned into the map, so I made the new graticule thick red. You can see that it’s mostly the eastern (left) hemisphere that had longitude problems, which makes sense as it’s the one that came up short.
Hollow World Placement
The polar lips starting at 66º look fine. The Blacklore Valley was always going to end up close to the polar openings, but its exact location may need some tweaking so that it doesn’t go too far into the opening; the Hollow World Set polar opening map seems to place it not quite inside, after all. This should be possible by adjusting the amount of sea at the edges of the map.
But the big question for placement is the size of the Hollow World. All of our tweaks have regularised the Hollow World so that it now appears to have a steady 7.5 hexes per 10º. With 40 mile hexes, this works out as:
7.5 x 40 = 300 miles per 10º latitude. 300 / 10 x 360º = 10,800 miles. Unsurprisingly, this is the figure we originally came up with. The official figure was 11,908 miles. Doing the formula backwards, that would mean 8.27 hexes per 10º.
We can easily add a little more sea at the edges to accommodate this size. This has the added benefit of bringing the Blacklore Valley up to right after 66ºS, which should mean it will be at just the right spot, on the gradual slope into the opening but still facing the inner sun.
Here’s what this looks like:
I would definitely add islands to the equator.
It took a surprising amount of fiddling to get to this point, but I’m pretty happy with the result.
Any questions?
One thing you may have realised is that even though I adjusted the Equirectangular world map vertically and horizontally to match the graticule, I only adjusted the hex map to match latitudes — not longitudes. As I noted, the result of this is that the area of the hex map will be slightly wider than it appeared on the original map. So why did I do this?
Because the projection we’re dealing with artificially expands distances east-west as you move away from the equator. To put it simply, the three HWR maps cover less area than they appear to, because they’re all away from the equator. Allowing this slight widening should counter that a little throughout the continent, allowing the shapes we know to remain closer to how we know them.
Also, I want to place as much of the Hollow World land as possible within the 66ºN to 66ºS span. This means we need more land from east to west, so this slight stretching seems to make a lot of sense here, too.
Next up: Tying up some loose ends before we go back to the 3D model and testing placements.
Last time we looked at the Outer World’s maps from this set. This time, let’s delve into the seemingly far less problematic Hollow World maps.
Hollow World Set Hollow World map
I’ll skip to the good part — above is the image with graticule lines added, ready for georeferencing. 703 control points later…
This map was a revelation when I first posted it back in 2009.
However, I now believe there are some problems with this interpretation.
To understand this, we need to delve into the Hollow World’s hex maps — something we have yet to tackle on this project.
Assembling the Hollow World
The main bulk of the Hollow World’s hex crawl was presented in two poster maps in the Hollow World Set: Northern and Southern Iciria, at 40 miles per hex. These fit together quite unambiguously, although there’s a small area of conflict around the Traldar Kingdoms, where generally the Southern Iciria map looks better. This can be confirmed with maps of Traldar in the Poor Wizard’s Almanac series. I took the coastlines from this third source, as they seem the most interesting, as well as being in general a compromise between the northern and southern maps.
The result of this is my updated Iciria map, which I released back in 2005. It’s just as valid eleven years later, so I’ll start with that.
Next, there are four 8 mile per hex maps to place on this map: Azca, Milenia, and two Nithian maps. These also place quite easily. Since there are five 8 mile hexes to each single 40 mile hex (in height, at least), there are multiple possible locations, but I’m happy I’ve found the best fits for all of these. Here’s the result:
There is one more official hex map to place: Shahjapur from HWA3. But before we can do that, we have to try to reconcile the hex maps with the world map.
Reconciling the Hollow World map
The 8 mile per hex maps are solidly based on the Iciria 40 mile per hex map. But what about Iciria’s map? How well does it combine with the world map? Unlike the Outer World, these maps were all produced at the same time, for the same single set. So we can expect them to line up a bit more easily.
In general, they do. But there are some caveats. First, the Iciria map is clearly based on the poster map, which means that unlike the Outer World hex maps, all of the Hollow World hex maps are based on the pseudo-Robinson projection. This in itself isn’t really a problem, except that it’s important to realise that we’re dealing with different projections with the outer and inner world hex maps. They are very different beasts. The most obvious example of this is that north and south are almost never straight up or down on the Hollow World maps.
Second, for some reason the Iciria map doesn’t fit over the world map as printed. Instead, the world map has to be stretched slightly horizontally in order to get a good fit. This is rather puzzling, and I’m not at all sure of the ramifications.
Take a look:
The coastlines from the hex map align nicely in the vertical dimension, but horizontally the land has clearly been stretched for the hex maps.
The policy of this project is to give precedence to hex maps, so let’s stretch the world map to fit this. It’s not a large stretch: the horizontal is scaled 103.5% compared to the vertical.
Another problem is that the fit of the map is rather wonky, with different placements needed to get the fit right in different sectors of the map. Luckily, most of Iciria is shown on the hex map, so my proposal here is simply to use the world map to add the missing extrusions to the hex map, using multiple placements of the map as necessary. There will be no effect on the other landmasses of the Hollow World.
However, it’s obviously necessary to choose a single placement of the central meridian and the equator over the hex map. I already chose this years ago, and there doesn’t seem to be any need to change it now.
Finally, let’s place Shahjapur. It’s not a wonderful fit, as it changed the shape of its island a fair bit, but that doesn’t really matter.
This is about as perfect as it’s going to get in terms of lining up the Hollow World maps together.
Now we have a conundrum… What to do about the projection? Time for more control points and georegistering… The following is a bit of a rough version — I may well redo this for more accuracy later.
Look closely and you can see the 8 mile per hex maps. Here they are in close-up:
It’s interesting to see these in this projection, but remember that these are still not the “actual shapes” of the land — for that, we would need a conformal projection.