The Atlantic Magazine Reveals How Google Builds its Maps – Not.
At last! We are close to delivering our final report to the US Census Bureau on their Master Address File and I now have time to devote to one of my favorite pastimes – writing blogs for Exploring Local. Hooray. What this means in consultant speak is that I am “on the beach” or between assignments, although truth be told, I am not looking very hard for anything to do for a while. I have my personal projects all laid out.
Over the last few weeks I have read a number of articles about maps and mapping that have renewed my interest in what’s going on in the profession. I guess that is something of a misstatement, since there a no longer enough people who actually are cartographers to make a profession, at least one that has any hopes of future growth. However, not to worry. Alexis C. Madrigal, a noted madrigal singer, oops I meant a the senior editor at The Atlantic, where he oversees the Technology Channel, recently wrote a marketing piece, oops I meant an “even-handed” review of the encyclopedia of all cartographic wisdom – Google Maps. His article “How Google Builds Its Maps – and What It Means for the Future of Everything” is a monumental tribute to revisionist history.
I suspect that The Atlantic magazine will soon be renamed “The Atlantic & Pacific”, since Googleland appears to be the heart of the cartographic universe. While reading the article I thought “Wow, he gets paid for writing this crap” and as I read further, I began to wonder “But who pays him?” The entire read resembled a poorly thought out advertorial.
I guess Apple’s entry into the mapping arms race has the big “Goog” upset and they decided to get ahead of the curve by bringing in a cub reporter, who knew little about mapping, to whom they could show why Apple doesn’t have a chance. Why Googs did not make these stunning revelations to a writer from a real tech magazine is an interesting question, but one to which we all know the answer. Madrigal’s article was enough to make me want to change my opinion of the problems that Apple must overcome to be a player in this venue. Well, let’s start down that road by focusing on the startling new truths that Google revealed to Mr. Madrigal about the world of mapping.
I know that it may be hard for some of you to realize that mapping was not discovered by Google. Last February I was examining a floor-mosaic–map in Madaba Jordan that dated from the 6th century AD that was designed to show pilgrims the layout of the Holy Lands. I can assure you that it did not have the Google’s copyright notice anywhere on it and I can, also, assure that it was not particularly old as maps go.
It may come as a further shock to you that Google did not invent map projections, including the one they use, nor did they invent tiling, symbolization, generalization, decluttering, zooming, panning, layers, levels, satellite imagery, map tiling, crowd-sourcing (both active and passive), their uniquely named “Atlas tool”, as well as most everything else associated with Google Maps. Even Google’s Street View had its origins in robotic sensing systems developed to enhance computer vision research, although Google, with the help of Sebastian Thrun, was smart enough to figure out how it could give them a competitive advantage in compiling street maps.
Where to start on the Madrigal article? How about the byline that reads “An exclusive look inside Ground Truth (no, Google did not invent this either), the secretive program to build the world’s best accurate map” Phew, I was glad to learn that Google was not attempting to build the world’s worst accurate map. I had hoped that the Googie was going to attempt to build the world’s most accurate map, but I guess that they just wanted the world to know that what they were building would be “bester” than whatever Apple could supply.
Did you notice the secret information the former NASA rocket scientist who was mentioned in the article told Madrigal? It was a howler, as a matter of fact it sounded like something I wrote in a blog when I was being snarky about Google. Anway, here is the exclusive/secret info from the horse’s mouth that was revealed exclusively the Madrigal of The Atlantic
“So you want to make a map,” Weiss-Malik tells me as we sit down in front of a massive monitor. “There are a couple of steps. You acquire data through partners. You do a bunch of engineering on that data to get it into the right format and conflate it with other sources of data, and then you do a bunch of operations, which is what this tool is about, to hand massage the data. And out the other end pops something that is higher quality than the sum of its parts.”
Wow was that informative. Before we go any further, I would like to note that Mr. Madrigal might have received a better education on Google Maps by reading this blog than visiting Google, but then that would be shameless self-promotion. So, will you tell him instead?
For some reason the opening figure in my copy of the article is a photo of two people playing ping pong. That had me stumped for a while, as I could not figure out what it had to do with project “Ground Truth”. Well, it still has me stumped, but I am working on it. I thought we might get a photo of a person at a workstation with a map on a monitor and details of the environment at the Google Maps office in Mountain View, CA. Apparently ping-pong is more interesting and newsworthy and the great Googs was not about to reveal any detailed information about how they compile maps to Mr. Madrigal.
I was more than mildly surprised that Madrigal seemed not to understand that there was more to Google Maps, or any map for the matter, than meets the eye. How did he think that routing occurs? Did he really believe the “Men in Black” idea that there were tiny aliens inside Google servers that supplied routes on demand as they were requested? Does he know that computerized routing has been around on a commercial basis since the early 1970s? Did he ever hear of Barry Glick, the founder of MapQuest, hawking online routing capabilities before Google was founded? Does he have any idea what NAVTEQ does with its highly instrumented vans and imaging systems? Has he ever looked at Bing Maps, or the hundreds of other services out there that provide competition to Google in the mapping sector? Put more simply, did the author of this article have the least little bit of inquisitiveness about what Google was telling him? My conclusion is a big, “Nope.”
I was, also, stunned to read Manic (okay, that’s supposed to be Manik) Gupta’s comment that the information in offline real word in which we live is not entirely online. When did this happen? Why is it allowed? Wow, this is beginning to sound like a science fiction thriller where there is no distinction between offline and online. Maybe Google really is an earth changing company, in more ways than we realize. Hopefully Tom Cruise will play the part of Gupta in the movie version of this thriller.
Gupta’s follow-up quote was even better – “Increasingly as we go about our lives, we are trying to bridge that gap between what we see in the real world and [the online world] and Maps really plays that part.” Hmmm. I had always thought that maps were a representation of the real world, and not the original thing. Based on the Madrigal article, it appears that he thinks that maps can and should serve as the real world. I guess Mr. Madrigal may not understand the real nature of project “Ground Truth”, or the use-warnings Google puts on their map and routes. I don’t know about you, but I have heard that trusting ground truth is usually a better strategy than ignoring it and trusting its representation, whether that representation is online, printed, in a box on your dash, in your phone, hiding in the ether while encoded in a radio wave, or packed as a new innovation labeled “Ground Truth” created by Google (and thousands of people before them).
I smiled when I read that Madrigal was stunned to learn that humans were involved in Google’s map making process. Yes. Humans are apparently needed to remedy problems that software cannot seem to solve. Imagine, data compiled from various sources that does not quite fit. Is that possible? Hmmm. Did Google invent that too? And is using crowd-sourcing to mine knowledge another Google innovation? No, I don’t think so. Is there no end to Madrigal’s naiveté? Well, the answer to that also appears to be “no.”
I hope you noticed his comment that, “This is an operation that promotes perfectionism.” I, also, liked this one “The sheer amount of human effort that goes into Google Maps is just mind boggling.” Followed by this, “I came away convinced that the Geographic Data that Google has assembled is not likely to be matched by any other company.” Well, guys, apparently it’s time to give up on mapping. Google, according to Madrigal, appears to have thought every thought about mapping that could ever be thought. Well, maybe not.
I hope you noticed the section of the article with this comment, “The most telling moment for me came when we looked at couple of the several thousand user reports of problems with Google Maps that come in every day.” A couple of thousand error reports every day? Is that like saying Google Maps only has 347 million known errors left to remedy? Seems that just like most of us Google will not be the first to achieve perfection in mapping. If you read my series of about Google’s map correction program, you know more about this than Mr. Madrigal, so you should consider applying for his position at The Atlantic.
I wonder why there appeared to be only one workstation for Mr. Madrigal to observe in Mountain View? According to Madrigal’s article hundreds of operators are required to map a country and many of them are in Google’s Bangalore office. Hmmm, So much for local knowledge. In part this remoteness of operators from the geography they are editing is why there are so many errors on Google Maps. In addition, maybe all those Google-driven innovations in mapping don’t quite help when the sources that Google harvests contain incorrect information to begin with. Adding erroneous information the edit queue of an operator who must use other surrogate information to validate it can be a recipe for disaster, as Google has proven time and time again.
I do applaud Mr. Madrigal for realizing that Street View is an important distinction in the marketplace for mapping services. Whether Google is actually using Street View for all of the processes mentioned in the article is unclear to me. Didn’t it sound like something a VP would say when Brian McClendon indicated that, “…… We’re able to identify and make a semantic understanding of all the pixels we’ve acquired.”? Wow, that’s great, but do you think they really do this? Someone should send McClendon some articles to read on image processing, as well as some older texts on information theory – seems as if they are doing a lot of work they do not need to do. And how about the number of businesses, addresses and logos they have collected with Street View. If only they could create a comprehensive address database with this stuff, but they can’t because of footprint limitations related to the deployment of their Street View assets. However, whether Street View provides a sustainable competitive advantage is something that Apple, Microsoft, NAVTEQ and TomTom will have to decide. It may a competitive advantage today, but I can assure you that whether or not it is sustainable, will not depend on Google’s wants.
So to Apple and all the apparently second level mapping companies – Don’t give up the map race quite yet. The fact the Google thinks you can’t catch them may be the best news you have had this year.
Finally, shame on Google for participating in a public relations effort masquerading as a report on technological innovation. While I have great respect for what Google has achieved with Google Maps, the interview behind the Madrigal article was not designed to reveal any details on Google’s technological innovations in mapping. Instead, it was an interview strategically planned to denigrate Apple’s mapping capabilities by implying that it could not compete with the great Googie. Revealing old news to someone who did not have a background in mapping, GIS or navigation is pandering and something I had not expected from Google. Just what is it about Apple’s mapping program that has them so scared? Hmmm. Something to think about.