12 replies to this topic

[woneil]

For a new writing project I have drawn the first of what will be quite a number of maps. This one is to illustrate a discussion of the choices facing the top American leadership at the critical juncture in July of 1941. They had to decide whether to continue to try to appease Japan, thus yielding to Japan further significant strategic advantages and risking alienating important allies, or risk a violent confrontation for which the country was not ready and which might distract from the struggle against Hitler.

I wanted to give readers a sense of the geographic relationships between the United States, Axis-occupied Europe, Japanese-occupied areas in Asia, and the Soviet Union, all in a map that must fit on an 8vo page (or an e-reader screen). Moreover, the map must be legible whether viewed on the screen of a color e-reader device (such as an iPad) or in a conventional book or a grayscale e-reader screen.

To me it seems that in a case of this sort the orthographic projection gives an immediate and direct visual sense of scale. Its distortions are comparatively unimportant because our brains automatically correct for spherical distortions. A case could also be made for an azimuthal equal-area projection, or possibly a stereographic. None of the cylindrical or pseudo-cylindrical projections appeal because the discontinuity dislocates the perceptions of global geographic relationships. I might make an argument, however, for a projection in which a complete image of the United States appeared at each end.

I experimented for quite a while using Versamap, which provides instantaneous re-draws as projection parameters are changed, before deciding on a projection center at 75N, 90E – Uncle Sam peeking over the pole. Then I constructed the map in Manifold, patching together polygons from several sources and doing some manual trimming. The results would not bear close scrutiny at large scale, but are acceptable at the very small scale actually used for the output.

All the coloring of the areas and boundaries was done in Manifold, with just a little bit of later patching in Photoshop. The shafts of the arrows were drawn as lines in Manifold, but the arrowheads were added later in Photoshop. I exported the map from Manifold as PDFs, with one representing the base map and blue areas, another for the arrow shafts, and a third for the Axis and Japanese occupied areas. These were imported into Photoshop and composited into a layer stack before doing some fairly minor editing and adding the labeling and legend.

I'm attaching both the actual color version and a representation of how it will look in grayscale.

Comments and suggestions welcomed.

Attached Files

•  World_Strat_7_1941.png   321.39K   163 downloads
•  World_Strat_7_1941_mono.png   290.86K   100 downloads

[DaveB]

Tricky subject to map. I think it's good you are trying a creative solution. Took me a bit of looking to kind of get in the mindset of looking at that view of the world, just because I'm not used to looking at it that way. That doesn't mean there's anything wrong with it.

Slight criticism - I find the blues/grays for the US and Allies to be a little too similar to distinguish easily. Hard for me to determine which is which when they're not side by side. Also, I wondered about the diagonal line fill for the Soviet Union. Obviously that shows some difference between that country and the solid blue ones.

I think the black and white/gray scale version actually works better in some ways. I picked up on the small Japanese bridgeheads on coastal China more readily with that, for example.

Overall it's a clean crisp legible design.

[Adam Wilbert]

I also like the grayscale version better than the color. On e-ink and other BW readers, do they simply present a desaturated version of the color image, or do they adjust the levels? I would be nervous about giving them a color image and assuming that all e-inks render grayscale the same way. You might be surprised that your orange and blue become the same shade on some random device for instance. Much safer to create the image in grayscale for everyone. In my opinion, it's a cleaner map, and the color doesn't really add anything to the experience.

I think a legend swatch for the Soviet Union's fill would help clarify alliances, or mixed alliances.

I like the projection of the map, but wonder if it might work better rotated 180º. Since most of the "action' takes place at the top, it might help with the north/south reading if that were at the bottom and the US was "peaking" over the top of the globe. That way the off-center pole would still be in the top half of the image as well. Its much easier for me to mentally flip North America over than to flip all of Europe, Asia, and Africa.

I would caution against using solid black for the German fill. Since your next darkest color appears as only 50% black, there is a huge jump up to 100k. Something in the 80k range would balance the image more. Save 100k for the text elements.

The "fi" ligature in Pacific should probably be removed since you've increased the tracking on the rest of the word.

Just minor suggestions. Overall, it's looking sharp.

-Adam

[Dennis McClendon]

I'm not so crazy about looking at World War II through a polar point of view, because I think it gives a Cold War missiles-over-the-Arctic mindset to an era when wars still had to be fought on the unfrozen surface (or at least on the ground every 1000 km). I'm a big fan of Harrison's unconventional points of view, but I think it might be distracting from the story of military strategy in the late 1930s to put at the center a region that was completely irrelevant—while giving short shrift to Southeast Asia, Indonesia, and North Africa, where there are important stories to be told.

So I would probably have chosen a projection that splits North America down the middle in order to show more conventionally the coastal threats from east and west, as well as showing the Soviet Union vulnerable at both ends.

[woneil]

Thanks for the feedback.

Regarding how e-ink devices render color graphics, I find that it's best before going to a final version to actually display it on my Kindle. Simply desaturating gives a rough approximation.

I want to try some other projections but have found that Manifold simply does not do cylindrical or pseudo-cylindrical projections with any limit other than -180 to 180! So I am trying it in QGIS, but finding problems there as well.

[woneil]

I made up a quick and rough catalogue of projections centered on 90E, for my benefit and that of others I've been talking to. I've attached the first page (of three). The others won't fit in the CartoTalk system but anyone who wants to see them is welcome to contact me at w.d.oneil@pobox.com. One page has pseudo-cylindrical projections including Wagner IV, Mollweide, and Hammer. The other has two Lambert conics, a Bonne, and an azimuthal equal area. All the maps are sized to fit the limits of an e-reader screen (4.75" by 3.5").

Attached Files

•  Projections_1.pdf   672.64K   17 downloads

[Dennis McClendon]

Any of those would work, though I'd probably make the split at 110ºW so North America is evenly split. Obviously Mercator and Gall have their drawbacks—and their cultural baggage. I don't want to put undue emphasis on the choice of projection; there are lots of good possibilities and other aspects of your project will ultimately be more important. I was just taken aback by the one that gave so much emphasis to polar regions in this context.

[Strebe]

I found Dennis’s comments to be spot-on. I suggest an equidistant conic with standard parallels around 10° and 45°, splitting North America, as Dennis proposed. The standard parallels suit the projection to the theater of war, more or less, whether in Europe or the Pacific. I could refine that estimate if you were need it.

Something like this. I’ve cut off above 75°N and 30°S.
 War.jpg   98.25K   31 downloads

It is unfortunate your tools hinder you from making the right map.

Regards,
— daan Strebe

[woneil]

The little catalogue of projections I attached to my previous messages can be used to visualize the problem with a projection of this sort. Publishing restrictions dictate only a certain amount of space, with an aspect ratio of 1.35:1. Because it spreads the areas out over a much higher aspect ratio, a conic results in very small visual elements on the printed or e-book page.

My current thought is that I probably will need to use a sequence of maps to orient the reader and then narrow in on the elements of specific interest. For initial orientation showing much of the world I'm pretty sure that a cylindrical projection of one sort or another is the best choice on grounds both of visual familiarity and efficient use of the rectangular print block. Other projections will come into play as focus narrows later in the sequence.

E-books are restrictive regarding format, but do have the advantage that there is very little cost involved in using more maps. I have to consider print forms as well, however, which will present a challenge in planning.

[woneil]

No one will be surprised, I'm sure, to learn that my first concern is data. With the help of CartoTalk contributors I've found a good series of shapefiles for Europe at intervals over World War II, and and a low-resolution world-coverage shapefile for 1938. There are shoreline files from today and I think that for the scales I need I can largely say that shorelines are the same as they were then. Many boundaries or at least sections also have not changed in the interval.

For the rest I think I'll need to rely on tracing from paper maps. I've ordered an atlas of Asia from 1939 that I have hopes for, and there are some maps in official war histories that may also be useful.

From all this I'm hoping I can piece together a reasonably accurate sequence of files from 1939 through 1945. In any event it should be possible to make files good enough that few if any will be able to spot whatever errors there may be.

I think I will wind up mostly relying on Manifold. It has a lot of frustrating and often rather silly limitations, but that seems to be universal and I think I know now how to work around them. For instance if I simply omite Antarctica, or at least the portion within a few degrees of the pole, I can avoid many of the projection problems. I cannot see that I really need to show anything south of about -50, so that should be no problem for my own purposes, inelegant as it may be.

[Dennis McClendon]

If filling a squarish proportion is a concern, here's another possibility: three "views from space" focused on the areas of most interest:



Or use two global views in the center of the layout, with insets of Europe and East Asia called out to the corners.

[jpbellavance]

One suggestion would be to draw your arrows differently. Within ArcMap it is impossible, as far as I know. But if you have Adobe Illustrator available you could draw curved arrows that may better display the directions. Sort of like in those old WWII news reels when this curved red arrow would come over the earth and strike a point on the globe representing direction and target of the attack.

You can use Adobe Illustrator for free for 30 days. Just enough time to finish your project.

.02

Jesse

[François Goulet]

I agree with Adam on that one. I was going to way to rotate the map 180° or so when I was his comment. Since WWII was mostly an European theater, it will realign our east-west relationships.