It has been a number of centuries since medieval cartographers filled the uncharted regions of their maps with mythological creatures. When the known world ended, the early artists were left with only their imaginations, and so turned to what they feared. Hic sunt dracones, the old ink reads. Here are dragons.
Today, the dragons have departed; the basilisks have slithered from their nests, and the fears of modern cartographers are unlikely to include fabled beasts. But they might, instead, include concerns over our increasing faith in proprietary mapping data. Google has assembled the largest collection of geospatial resources ever seen outside government. They have mapped the streets of every nation, and in over a quarter of those countries their Street View cameras, bolted like sailing masts atop sedans, have stitched together mosaics of the roadside. At a technology seminar, a Google engineer declared that “anything that you see in the real world needs to be in our database”; a recent New York Times Magazine article recounts a Street View project leader speaking of rendering the world in “about one pixel to the inch.”
It’s unclear exactly how much money Google has spent on Street View, but the amount is considerable. We also know there have been questions about whether the investment will eventually turn a profit. But one might rightly view the undertaking as a calculated wager, a gambit, with all losses to be recouped once the prize is reached.
The Street View image of Google’s headquarters in Mountain View, California is an accurate picture in that it is a likeness of what was seen by that particular camera, at that particular moment, in that particular location, from that particular angle. It was, apparently, a sunny day (Google cars are forbidden from taking pictures while it is raining). One can tell by the foliage that it was spring or summer, and the long west-facing shadows would indicate the early morning. On the roadside, some people are brandishing cell-phone cameras; others are waving. One Google employee, ID badge dangling from his jeans, extends a paper sign: “Hi Mom!”
This image is true, or was true, once.
“Google takes a picture of your house in the morning when there’s no one in the front yard,” says Rachel Franklin, Associate Director of Spatial Structures in the Social Sciences, and a geographer by trade, “But you have a house at night, a house in the winter, a house in the summer.”
In Street View’s rendering of the world, the sun does not set. The employees do not return to their offices. The summer does not turn to autumn, and the leaves do not fall from the trees. A location is particularized down to a single image, which, until the next Street View car rolls around, becomes indelible. Jean Bauer, Digital Humanities Librarian at Brown University, says “I know people who knew that the Google scanner was coming to their street and tried to mow the lawn.”
That has the cadence of a joke. But for the business owners, landlords, and realtors whose products will be seen time and again through Street View, what other choice will they have?
Concern with the permanence of map displays is not only a modern problem. For centuries, cartographers have struggled with faithfully portraying what we know about the world because any printed map of the earth must be “projected”; that is, the latitudes and longitudes of the sphere must be systematically transformed so as to be seen on a twodimensional surface. But just as the world cannot simply be unfurled, a flat surface cannot accurately represent the features of a sphere. It is impossible to preserve both the area and shape of landmasses; accuracy in one must come at the expense of the other, and these distortions can be seen in the overblown sizes of Greenland or Antarctica on some maps, or in misrepresented distances and directions.
For much of cartographic history, the dominant display was the Mercator projection, derived in 1569 by Gerardus Mercator, popular because it preserved rhumb lines and loxodromes, which sailors use to chart their courses. During the nineteenth century, popular use of the Mercator began to decline because the projection wildly distorts areas, exaggerating the size of Europe and diminishing Africa. Again, the controversy was over the static and singular nature of the image: if too many elementary school students looked only at their classroom’s copy of the Mercator, they would believe certain continents to be much bigger than they are.
Our interactions with maps today have become less global, more local. There has been only muted talk about which projection Google uses (yes, the Mercator), but as our attention localizes, so do our cartographical problems. We know that any representation of reality is prone to error, but the errors have moved from being the best we can do with imperfect tools, to alterations of the rendered world, either by error or design.
“Google’s worldview becomes the geographic truth. The stuff that they choose is the stuff that is real,” Franklin says, “That means we have to trust Google not to lie. Something always gets generalized, something gets lost, and Google gets to choose what is lost.”
The concern—editing the world—is not hyperbole. Google has covered over areas of the White House and other federal buildings by request of the United States government. Security is the noblest of causes, but it is a lie, Google’s lie, that suddenly becomes the truth. And that lie will be seen by everyone. Unlike the standard data in Google Maps, which is available for use by any developer, the Street View data remains guarded, and we are uncertain of the full extent of the world-edits. Bauer also thinks the images would be exceedingly valuable if released: “You could do vector analysis on architecture images. Color palette analysis on a city.” But, she says, “You have to get the right information.”
The right information—that is the grail, and the hope. It is also, we know, a fiction. One grows leery that as private interests map the world with ever-increasing degrees of precision, we begin to trust the dubious results with ever-increasing degrees of faith.
Thankfully, Google is not the only player in the game. Two years ago, Google began to charge high-bandwidth users for taking their map data. Writing in the New York Times Magazine, Adam Fisher relates that “the change prompted an exodus,” with Foursquare, Wikipedia, Craigslist, Apple, and others ceasing to use Google’s data and instead turning to OpenStreetMap, an open-source alternative run mostly by unpaid volunteers. When Apple released their own Maps app several months later, users described receiving directions to drive through buildings, and across bodies of water and airport runways. Roads, highways, landmarks, and entire towns were reported as missing from the database. Apple issued an apology, and even suggested using a competitor’s product as they repaired the issues. Perhaps they should have added dragons.
KYLE GIDDON B’15