« Sisters Are Doin' It For Themselves | Main | Coming out of the woodwork post friday open post »

June 02, 2011

Comments

I don't have any more info, but my Yuan is on 20something programmer, or maybe a small group of 20someting programmers, making the change without really thinking about it. The whole stupidity/malice thing.

Though maybe they didn't want to be perceived as taking sides in a "political minefield"?

Though maybe they didn't want to be perceived as taking sides in a "political minefield"?

I have no tolerance for that. And how, seriously, could it be a political minefield to recognize the actual legal truth of the existence of these sovereign nations on their maps?

In fact, I would point out that reservations are more independent than states are: Google would have had more legal justification to drop the state boundaries from the map of the US. Tribes are domestic dependent nations, not counties or municipalities or commonwealths, and they should at least get that recognition.

I'm with ugh on this. Welcome to the decentralized world. The sources that Google relies on for these sorts of maps and information are less likely to recognize reservations than they are state lines. Assuming that it was a hi level decision and Eric Schmidt said 'let's erase these, who needs them' is really missing the mark.

One way to look at this is imagine that Google made a mashup algorithm that would take all the maps in the world and weigh the lines according to how many maps they appear on. As they add more maps, the lines that are on all maps, such as state boundaries, would get darker and the lines that are on a smaller proportion of maps would grow fainter. Then insert in the algorithm some cut off point where lines are no longer drawn. That formula would have the precise same effect as what you describe, yet not be the result of any human decision to erase the lines. So when you say 'why did you erase those lines?' the person might look baffled and say 'I didn't erase any lines'

While there is probably not a map crunching algorithm like that, the basic process a lot of what mashing up is doing is basically that. This is not to say that it shouldn't be pointed out or that it is somehow harmless, it can be quite pernicious in many ways. But attributing it to some specific human decision making process will lead to a lot of passion expended that might be better used in other ways to deal with the problem.

Geronimo meets Kafka.

Tibet has China and American Indian reservations have Google to define their non-existence.

Are any Indian Reservations on the Death Plain Summer Vacation Tour? I mean, after, the The "Statute" of Liberty (which she called the Lady in a tweet, thank you, Gary Busey. Please let her win the Presidency and hand her the red phone and the nuclear button; it's time to cancel this fatuous reality show called America) and pizza with the sh8thead underneath the toupee that Custer gave up once before.

The United States of Stupidity/Malice.

We're a joke. And we want the rest of the world to adopt our mirthless, malign laugh track and call it, well, what should they call it?.

The sources that Google relies on for these sorts of maps and information are less likely to recognize reservations than they are state lines.

Cite please?

I ask because the US government makes GIS data describing the contours of Native American reservations available to the public here.

Assuming that it was a hi level decision and Eric Schmidt said 'let's erase these, who needs them' is really missing the mark.

Did someone assume that? Who?

While there is probably not a map crunching algorithm like that, the basic process a lot of what mashing up is doing is basically that.

Google Maps is not a mashup. Please, I beg of you, if you've never written a computer program or worked extensively with GIS data, just refrain from speculating on how Google Maps works. Otherwise, you'll miss minor details like the fact that they have an extensive test suite that verifies that maps don't change in unexpected ways when they update the software which means that any change that caused the reservations to disappear probably had to be explicitly approved by a staff engineer and signed off on by at least one another engineer.

imagine that Google made a mashup algorithm

and

there is probably not a map crunching algorithm like that

Expending all this passion and energy just seems so wanton.

imagine that Google made a mashup algorithm

Why should we? Imaging that will only confuse people since that's not how the actual software works. At all. Why do you want to make people more ignorant?

Expending all this passion and energy just seems so wanton.

Wow. You have no idea why that statement was a joke made at hsh's expense, do you?


Hey, if you don't want to have a serious conversation, that's fine. I asked you for a cite, I provided a relevant link, and I tried to explain some basic software engineering principles of which you seem to be ignorant, but if you'd prefer to randomly snipe rather than have a real discussion, that's cool. I can't keep you from trolling your own blog, although I don't see why you'd troll in Dr Science's posts per se.

Turb, I said imagine. I can't make you imagine something, but that seems to be a deficiency in your mental makeup rather than anything I can do anything about.

I can't understand why you would choose starting a fight with me over reading what I wrote, but I can't keep you from doing that.

LJ, I have no trouble imagining, but I do find that imagining absurd incorrect things when trying to understand a problem can be counterproductive. If your kid is stuck on a math problem, telling him to imagine that 2+2=22 is not very helpful.

Secondly, I'm not fighting you. I might be chuckling at you just a little, but not fighting.

Third, I'd like to bring the discussion back on point. So, do you have a cite for the point I asked about? Can you tell me who here assumed that dropping the reservations was a high level decision made by Schmidt? And can you explain why you believe that Google has no test coverage sufficient to catch large areas of their maps disappearing overnight?

Wow. You have no idea why that statement was a joke made at hsh's expense, do you?

Well, the light bulb just came on for me. I didn't even catch it the first time, and they were my words.

*Ohhhh ... now I get it. (smacks forehead)*

On the sovereignty of Native American tribes, I can cite the story of the Iroquois lacrosse team's attempt to participate in the international tournament held in Britain in 2010. http://www.9wsyr.com/news/local/story/Update-Iroquois-Nationals-Lacrosse-Team-returns/rrAG8HtNx0i4HWLs7-KWKw.cspx. The Iroquois tribe had issued its own passports, but the British felt they were insufficiently secure, although the US State Department backed them up. The passports have served for travel between the US and Canada,
Incidentally, the Iroquois invented lacrosse, a fact that did not embarrass anyone sufficiently to successfully resolve the problem.

Turb, I think that LJ was not so much suggesting that Google had done a mash-up which changed the map, so much as he was atempting to point out that there are possible scenarios (of which this was only one) where something can change without an explicit decision by a human being to change that particular detail.

Did it happen in this case? I suspect not. But that is not to say that the possibility doesn't exist. And, as I read LJ, what he was saying is that we should not get outraged at a human decision before we are sure that it actually was one. (And LJ can now correct all the ways that I am imputing motives to him without any evidence. ;-)

Turb, I think that LJ was not so much suggesting that Google had done a mash-up which changed the map, so much as he was atempting to point out that there are possible scenarios (of which this was only one) where something can change without an explicit decision by a human being to change that particular detail.

Well, anything at all is possible...perhaps an angel waved a magic wand and whisked away the reservations. I hope you can see how writing long paragraphs about said angel's motivations doesn't really help us understand anything.

But in the realm of the plausible, LJ's imaginings are just absurd and wrong. Google Maps is not a mashup. Period. The federal government publishes a huge amount of GIS map data, including reservation contours. Changes to Google Maps software and the underlying data are carefully reviewed. There are automated test suites in place to ensure that the maps which users see can't change unless someone (really, multiple someones) at Google approves the changes. Which means that either more than one Google employee knowingly approved removing the reservations or more than one Google employee screwed up and failed to notice that an unrelated change removed the reservations.

And, as I read LJ, what he was saying is that we should not get outraged at a human decision before we are sure that it actually was one. (And LJ can now correct all the ways that I am imputing motives to him without any evidence. ;-)

wj,
Sorry to come out of the woodwork on this, but correctomundo! I'm pleased that my message got across to you, in that it was completely blocked out by turb's heady combination of belligerence and reading fail.

In the post that Dr S links to, Indian reservations are described as 'strange tan blotches'. If they aren't identified (and 'strange tan blotches' seems to fit that description, no?), it is relatively easy to imagine one engineer, or even a whole passel of engineers, to not really notice when they are gone. Comparing to state borders does no good. There are any number of states who I could hold up a silhouette of and most Americans (of which I believe Google would draw those engineers from, though I'm not going to find a cite for that either) would be able to identify. I doubt there is not one (and I'm sorry, I don't have a cite on this, but I trust you can imagine and I hope you can help others to) who could identify indian reservations based on their outline. Careful review doesn't make a bit of difference if you don't have that base information there.

I doubt there is not one...who could identify indian reservations based on their outline. Careful review doesn't make a bit of difference if you don't have that base information there.

Tiptoeing near where angels should fear to tread: lj, the scenario you posit, as an excuse (of sorts) for why someone wouldn't notice that the Indian reservations were gone, isn't very plausible. Much more plausible is that a careful review would specifically compare before and after, so that it would be impossible not to know that "before" identified reservations, and "after" didn't.

I do agree that dunderheadedness is at least as likely an explanation as ill intent. And, I'd be very curious as to what the ill intent was, if indeed this was done for some deliberate/political reason.

In the post that Dr S links to, Indian reservations are described as 'strange tan blotches'. If they aren't identified (and 'strange tan blotches' seems to fit that description, no?),

Unidentified on the maps that the public sees does NOT mean that Google engineers don't know what it is. They almost certainly have more detailed viewers that show them metadata describing where map elements came from. People who run public websites often have much more powerful tools than the sites' users do.

it is relatively easy to imagine one engineer, or even a whole passel of engineers, to not really notice when they are gone.

This might be true if we assume that Google does not do any quality assurance testing at all. However, I know for a fact that this assumption is wrong.

Let me tell you how Google Maps does not do QA. They do not change code and data and then have some random people eyeball randomly selected maps to see if they notice anything gone missing. This is very obvious to anyone who has ever written a computer program. Therefore, your whole supposition here is just wrong.

Instead, they almost certainly run automated tests that compare the old and new maps (or some intermediate data that gets turned into maps) and highlights differences. Updates are not allowed unless enough people sign off on the differences being acceptable. If an engineer has no idea what about a splotch that his test set is telling him disappeared, he's not going to approve the code change until he finds out. He can't "not notice" the splotch; the whole point of automated testing is that the machine notices everything and highlights differences for you.

Having new code generate results that don't match the old code is usually a bug. Even if you think the tan splotches aren't important, you still need to find out what they are because until you do, you can't be sure the new code isn't dropping other important features.

'strange tan blotches'

If the previous iteration failed to identify it, would marking it as a difference somehow help them figure out?

'The machine notices everything'

Yep, that's why these things never ever happen. Why bother imagining when a machine can do it for you?

lj, I am working on the specs for a rewrite of some very old computer programs that lie at the heart of the business of the company I work for. Every quarter, one of the programs takes, as input, several hundred numbers that are carefully reviewed and set by humans, one by one. Then those numbers are crunched with a lot of other data to produce 6000 (potentially 40,000, but never mind that part) numbers that are then used by other programs to produce potentially hundreds of thousands of other numbers.

There is no way on earth that humans are going to review these numbers one by one. Yet we still have to know if some of our numbers are going wildly (or even not so wildly) out of the range of what doctors call "within normal limits."

Machines do that checking for us. We write programs to tell us things like "give us a list of the numbers that represent a change of more than X% compared to last time." Then we can look at just that subset to try to figure out what's going on.

Sure, machines, programs, and programmers make mistakes. But that doesn't invalidate Turb's point about how google would be reviewing changes, nor does it mean that we could get some of these things accomplished without machines. We just wouldn't do them.

When I worked in a college admissions office a gazillion years ago, databases were just starting to replace index card files for keeping track of applicants. Someone pointed out that it wasn't remotely that the "databases" (very crude in the late 70's, I can assure you) meant less work for the human beings that ran the admissions operation. But they did allow the human beings to do a wild variety of things that could not have been done using the old system of having two big boxes of index cards, one sorted by last name and one sorted by zip code.

So what does "why bother imagining when a machine can do it for you" have to do with google checking new output against old? Nothing remotely like google/maps would exist without machines. He||, this blog wouldn't exist without machines. ;)

If the previous iteration failed to identify it, would marking it as a difference somehow help them figure out?

Yes. The data had to come from somewhere. There's a good chance it came from the US Census Tiger data set, specifically from the American Indian Reservation layer. They might not have complete labels that allow them to add names for each reservation, but they almost certainly have internal data that tells them 'this splotch represents a reservation and it comes from Tiger 2010 data layer IndianReservation'.

Yep, that's why these things never ever happen. Why bother imagining when a machine can do it for you?

I am not claiming that errors can't happen. Rather, I am claiming that it is likely that the disappearance of the reservations was noticed in testing and specifically approved. I can imagine many reasons why Google staff might knowingly approve this change.

Hmm, I evaluate quite a bit of software these days, and the scheme Turb lays out looks like the kind of thing a mid level manager may well be convinced is true.

This has not been my experience as a user. An awful lot of junk sneaks out there. I would buy just about any other scenario than what Turb describes here. The products are just plain not reliable and professional.

Hmm, I evaluate quite a bit of software these days, and the scheme Turb lays out looks like the kind of thing a mid level manager may well be convinced is true.

I am not a mid-level manager. I write software for a living. I spend a lot of time thinking about, discussing, and writing tests. I write tests for every single software change that I make. And every single change, including the tests I write, has to be approved by at least one other staff engineer before it goes into the production network. Since my employer is responsible for a fairly large fraction of internet traffic (think: double digit percentages), our quality control efforts are important.

I do not work at Google, but a number of my friends do. Their office is just down the street from mine. Every few weeks, I visit there and have lunch with them and we chat. My impression based on these conversations is that Google takes testing pretty seriously and devotes a lot of resources to it. Are they perfect? Absolutely not. But the level of technical excellence is quite high.

My code runs on many thousands of computers around the world, 24/7. The same is true for Google engineers. We all probably employ a degree of testing that is not common for random off-the-shelf applications that typical computer users are familiar with. This is not because we are better engineers but because we are working in a very different environment. The average desktop PC has a vast array of possible configurations and failure modes and lots of testing time for desktop applications has to be expended testing against that variety. I and my Google friends don't have to worry about that: we have one platform that is, frankly, better designed than the crap that regular computer users get. It is certainly better managed. So we can spend our time writing better tests.

...looks like the kind of thing a mid level manager may well be convinced is true.... The products are just plain not reliable and professional.

How fun to have something so deliciously patronizing and superior combined with vagueness to the point of irrefutability.

"The products"? Which products? Airline reservation software? The software my company produces to help other companies manage their international payrolls? Payroll software in general? The programs that manage electrical utilities, mobile phones, TV, elevators, online searches? The internet itself?

Yeah, there's crappy software, and it's easy to focus on that. I do it myself every time I complain about the dumbass search engine Barnes and Noble employees have to use in-store. The squeaky wheel and all.

But the "amount" of software that works and works invisibly, reliably, and well is staggering. And it wouldn't work that way without -- as I see I'm cross-posting with Turbulence -- reliable testing mechanisms and protocols of the kind Turb is describing.

OK, let's assume that that Dr. S, Turb, et al.are correct: that there was a conscience decision to remove the indian reservations from google maps. Would anyone care to suggest a reason why that decision would be taken?

I can see how an initial decision might be made to omit something like that. Deciding to remove something which was once there, however, would seem (at least to my innocent mind) to require a bit more reationale. I just am having trouble figuring out what it might be.

Side note to anyone tempted to offer "white men all hate Indians" or some variation on that theme: If you are going to ascribe malice, you need to have more evidence than someone who is merely suggesting incompetence. Not to say that malice is an impossible motive; just that it requires more support.

I can imagine many reasons why Google staff might knowingly approve this change.

It's interesting how your imagination works when trying to support your views, but is totally absent in trying to read others. I've noticed that a lot, but it is always good to have a reminder.

Janie, apologies if being called 'ignorant' and being accused of 'trolling my own blog' bleeds over to anything I say to you. At the top of my comment, I said 'welcome to the decentralized world'. As you distribute these functions to a wider range of people, you depend more on their base of knowledge and if that base doesn't cover things like 'there are these sovereign entities that were created because of treaties with various Native American tribes', they are going to get lost.

This New Yorker article describes scrapping of card catalogues in place of online databases. Despite the fact that librarians should have been most attuned to what was being lost, whole card catalogues, and any annotations and notes on them are gone.

If you read the linked posts, the writer notes that Google maps and the whole Google enterprise has a specific visual aesthetic. I suppose at some point, 'this is the way I want it to work' becomes 'I don't want that information in there', but when one is talking about maps, you have so much data to choose from, the question becomes not what do you exclude, but what you choose to show. And, just like picking a basketball team, choosing one item immediately means that you can't take another item to fill that place. Saying that the data must exist somewhere misses the whole point. Asking for a cite on whether state borders are better known than reservation borders misses the entire planet of points related to this.

Also, Google has always made the value of simplicity take precedence over everything else. If the goal is to get the map to load faster, you take out as much information as you can while leaving what most people would need to recognize it was a map. I have no idea how much not drawing in reservations saves, but if your aesthetic is to carve out as much as possible, any lectures about how little computing power it requires to add that to maps is really meaningless.

Whether my comments are ignorance or trolling, I leave it to you to imagine. Of course, if it is attributable to a specific person or group of people, I'm sure that Turb can use his network of Google friends to register a complaint, as Dr Science did and perhaps it will get a better response, since these are his friends and all that and I'm sure that if he uses the same manners he uses around here, he's sure to get satisfaction.

Google Maps 'loses' major Florida city


It's home to 90,000 people, one of the nation's biggest malls, a National Hockey League team and, the city's mayor boasts, the first IKEA store in the state.

But, for at least a month this summer, Google's computers "lost" Sunrise, Florida.

People who searched Google Maps for the city were directed instead to Sarasota, Florida -- a place that, while an alphabetical cousin of Sunrise, is actually 200 miles away. No Sunrise business or addresses or phone numbers showed up. Even city hall and other public entities were strangely absent, according to reports on news sites, blogs and Google help forums.

Google says it has since fixed the "technical error," and if you searched Google Maps for Sunrise on Wednesday, you would find a healthy grid of streets, malls and parks on the far outskirts of Fort Lauderdale, where it's supposed to be.

But this week's fix didn't come until after public outcry.

Sunrise's mayor, Mike Ryan, said this is the third time Google has dropped his city off the digital map, which he says is unacceptable.

"I don't have any problem with the idea that mistakes happen," he said. "The algorithms they have to apply to understand what my search is are undoubtedly complicated. What disturbed us is that this wasn't the first time it happened."

When he heard of the most recent drop, he was in disbelief.

"I said 'holy cow,'" he said. "It felt like a bizarre novel -- that all of a sudden we disappeared. We woke up one morning and we didn't exist in the ether world."


Sunrise isn't the first city that Google Maps has misplaced. The blog Search Engine Land documented five cities -- La Jolla, California; Rogers, Minnesota; Wickliffe, Ohio; Woodstock, Virginia, and Imperial Beach, California -- that Google Maps has also lost and found.

The Mountain View, California, company did not respond to CNN's questions about those previous outages, nor did it provide details about the Sunrise case.

The company did say it had fixed the problem and that it strives to deliver "the richest, most up-to-date maps possible."

"We've built our map from a combination of authoritative sources, ranging from the U.S. Census Bureau to commercial data providers, and have used satellite, aerial and Street View imagery to help complete the map," the company said in an e-mail to CNN. "Overall, this provides a very comprehensive map of the U.S., but we recognize that there may be occasional inaccuracies that could arise from any of those sources."

It's interesting how your imagination works when trying to support your views, but is totally absent in trying to read others. I've noticed that a lot, but it is always good to have a reminder.

Seriously, I have no idea what you're talking about here. I assume you feel clever constantly dropping snide comments about my lack of imagination, but I just find it bizarre. You invoked imagination to describe a process that does not describe reality and that is likely to confuse people. I pointed out that your imaginative vision was incorrect and not helpful. When you're trying to understand something weird, as Dr Science is trying to do in this post, imagination can be helpful. But imagining absurd things that are not true is not usually useful. Don't you agree?

Janie, apologies if being called 'ignorant'

Wait, do you actually deny being ignorant? I mean, have you ever written computer programs or used GIS data? I don't think anyone should be ashamed of being ignorant. There are many many topics on which I am totally ignorant. But we should recognize our own ignorance and keep quiet on topics about which we know nothing.

you have so much data to choose from

Look at the maps that Dr Science included. I don't see "so much data to choose from"; I see mostly empty space. No one looking at this map is going to suffer from information overload, no matter how many reservation boundaries you add to it.

choosing one item immediately means that you can't take another item to fill that place.

That's not true. Look at the yahoo/bing/old-google maps Dr Science included in her post: you can clearly see towns and other features inside the reservations. Just like drawing state and national boundaries on the map doesn't prevent you from displaying any other data on the map, showing reservation boundaries is no impediment either. Remember, we're talking about zoomable maps where users can control how much information is displayed, not paper maps.

Beyond that, note that most reservations are in very sparsely populated areas. You won't find any reservations in downtown San Francisco.

Asking for a cite on whether state borders are better known than reservation borders misses the entire planet of points related to this.

LJ, I never asked for a cite about that. Go back and reread my comment please. I asked for a cite about your claim that sources of data that Google uses would be less likely to notice reservation boundaries disappearing. As I explained, Google is probably getting its data from the US government GIS data, and the people who curate that data would absolutely notice reservation data disappearing. Publishing this data is their job, and if they accidentally failed to include reservation data, they'd get an earful.

I asked for a cite because I (foolishly) assumed that you might know something about the GIS data source in question and thought you might be able to teach me something useful. You, um, sure showed me.

If the goal is to get the map to load faster, you take out as much information as you can while leaving what most people would need to recognize it was a map. I have no idea how much not drawing in reservations saves, but if your aesthetic is to carve out as much as possible, any lectures about how little computing power it requires to add that to maps is really meaningless.

Map loading speed is not the same thing as simplicity.

Beyond that, this doesn't make any sense. Look at the Google Map for any major city; it is way more crowded and less simple than the map Dr Science highlights above. As I explained above, most reservations are in sparsely populated areas where the reservation boundaries aren't even competing with other features; they're competing with completely empty space. Finally, I can assure you that the bandwidth and computation time needed to draw reservation boundaries are completely insignificant.

Partly responding to wj, and just to be sure it's clear that I don't think I have enough information, or even enough imagination, to know why the reservations would have been removed on purpose, I'll repeat myself: I'd be very curious as to what the ill intent was, if indeed this was done for some deliberate/political reason.

I missed at talk by Eli Pariser the other night (because of bad weather) about his book The Filter Bubble. It's not quite the same phenomenon, and I'm just making a guess based on not even so much as a snippet of a whiff of a memory of something I read in passing while surfing the web, but I would be curious as to whether Google tweaks its maps in part based on the number of searches for specific entities or types of entities. (This is not meant to respond to jrudkis's comment; that one surely comes under the "mistakes do happen" heading, which no one has disputed.)

This isn't to say that even if popularity of searches matters in Google's decision-making, and then even if the reservations are/were one of the least searched kinds of entities, it justifies not showing them any more (for any number of reasons). But I'd be curious -- Turb? -- if this is part of the decision-making process.

Or in other words, did some "rational" decision get made in isolation from other considerations that we would have preferred to see taken into account?

JanieM,
That video has been bouncing around the interwebs a bit. While Eli Pariser has a point, the example that he uses about Egypt is problematic. This is not what I originally read, but it has some of the points I've seen.

I'm not sure how popularity figures into Google Maps. What it presents you with is a map, If it presented you with something that was somehow derived from the popularity of searchs, you'd have a huge number of points around major cities and places where there would be nothing. I dare say the number of people looking for a particular haute cuisine restaurant would be more than the number of people looking for a reservation.

Furthermore, as the post notes, reservations were already unnamed in previous iterations of Google maps. No matter how technically proficient the google staff is, the choice is between a splotchy tan blot and a blank space.

It is not a question of 'ignorance', it is a question of how one looks at the information as well as the sources that one combines.

Would anyone care to suggest a reason why that decision would be taken?

My guess? If they're changing data providers in a hurry and the new provider either doesn't include reservation contours or does but requires extra integration effort. Such changes are often driven by contractual or licensing changes. For example, data is often licensed for a particular set of uses and if Google is planning on rolling out a new service (or if legal counsel reviewed a contract and decided that the services they're currently using the data for are not quite covered by the existing contract), then switching to a new provider would be required in very short order. I've been involved in cases like that where the legal team told us 'switch providers now or we're in huge trouble' and the transition was messy because we were under the gun: some stuff was knowingly broken because we didn't have time to get it right and some stuff was just unknowingly broken because we were in a hurry.

Another explanation would be that they're refactoring their existing code and data and that temporarily dropping support for reservations makes their rollout much easier. It is not unusual to refactor systems in two passes where you replace the complex existing code with a simpler (more performent) version that doesn't quite do everything the old code did, deploy it to production to shake out the bugs, and then start adding back new features. Alienating existing customers by dropping a feature is bad, but this approach can get you to a better working system than alternatives in some cases, so sometimes companies do that.

Also, since Google's product was worse than competitors with regards to reservations before this change, I can easily see them thinking 'eh, we're awful at reservations, so there's no big deal in dropping support for them while in the course of improving our reservations handling or making other improvements...after all, people who live in those areas are probably already using yahoo/bing'.

And of course it might just be a screwup. Those happen.

Those are all just guesses though. Like Janie suggested, I'd imagine that the fact that relatively few people are affected by this change probably helped project managers decide to make it if it was indeed a choice. Reservations are sparsely populated and rarely have great cellular data service so they're not going to be generating tons of usage for Google's mobile maps product. Regardless of how the reservations got dropped, they're probably not going to be added back unless a project manager decides this is high priority enough to get a developer to drop what they're doing and fix this. That question boils down to priorities and I'd bet that Google's priorities are affected by the relative search popularity of the affected areas.

Reservations are sparsely populated and rarely have great cellular data service so they're not going to be generating tons of usage for Google's mobile maps product.

All of this is true, and I find your speculations about how this came about to be credible. I doubt there was some explicit ill will on the part of Google toward native Americans.

All of that said, there also aren't a lot of people in Nunavut, or Greenland, or Alaska, or Mongolia, but all of those places show up.

There are not, by design, a lot of folks living in the National Park system, but those show up.

All of which, I think, supports Doc Science's point. The presence of Indian reservations - sovereign territories within the US - is of sufficiently little importance that ignoring them doesn't even require ill intent. Mere inconvenience is a sufficient basis for letting them fall between the cracks.

Apologies in advance for the long-ish response.

Short a response from Google, I might be able to provide some context: Until very recently, I worked at the larger of the two providers of mapping and navigation data. (Google was a customer until 2008.) In a nutshell, mapping is an insanely complex world, with a zillion kinds of data and even more data providers ... almost none of which/whom are standardized or consistent in quality. And that's before you get to the process of actually making a map. With increased complexity comes increased risk, thus the likelihood something like this -- 1,000 steps down the supply-chain and production line -- will happen.

While it's certainly possible someone goofed a few lines of code that altered the display of certain data, there are many other possible explanations such as what Turbulence noted. A sudden change in data provider, raw data, data ingestion/processing, quality controlling, map rendering, or any/all the above could cause a disruption in what we see and how we use a map. Yes, it's just that nuts. (I have way too many surreal memories of things like Sunrise happening.)

Oh, and let's remember that this is the company that continually updates its products on a seemingly daily basis and almost always behind the scenes, so the chance we'll see something different tomorrow is practically guaranteed.

So I doubt there's knowing malfeasance at play; there are just way too many variables to control for.

Mookie, I'm confused. Turbulence said this

any change that caused the reservations to disappear probably had to be explicitly approved by a staff engineer and signed off on by at least one another engineer.
and
I can imagine many reasons why Google staff might knowingly approve this change.

So are you saying that there wasn't a sign off?

Along the same line: who does Turbulence suppose "signed off" on the omission of Sunrise, and the other cities mentioned in that link?

I genuinely don't get it. If they can lose entire cities of (white?) Americans without evil intent, why can't they lose Indian reservations? What's the difference?

Along the same line: who does Turbulence suppose "signed off" on the omission of Sunrise, and the other cities mentioned in that link?

A staff engineer. Why exactly do you think that Google dropped Sunrise without having any knowledge of what they're doing? If I had to guess, I'd say their data provider screwed up and dropped the city in a new data release and they were forced to choose dropping Sunrise while bringing in lots of new updates that people were waiting for or keeping Sunrise and foregoing the updates. From a business perspective, dropping Sunrise may very well make sense.

If they can lose entire cities of (white?) Americans without evil intent, why can't they lose Indian reservations? What's the difference?

I don't think I wrote anything about evil intent. I certainly don't think Google is full of people who hate Native Americans and wish to expunge them from maps. But I do think there's a good chance that someone at Google made a decision for all the reasons I explained above. Or it might have been a screwup. But Google has an enormous test suite and an entire staff of people specifically to prevent screwups like this from happening without someone making an affirmative decision.

Turb: Thanks for the clarification.

The comments to this entry are closed.

Blog powered by Typepad