Google Maps changes a route after the drama of young people lost on a ghost road
tekdeeps.com
tekdeeps.com
-- Manual on Uniform Traffic Control Devices, 2009 edition, section 2A.04, paragraph 01
Replace 'signs' with anything--no one's patience/attention will hold up in the face of a constant barrage of information.
We had to get a change made to Google maps to prevent people getting killed here, and it took national news, and personal contacts within the organisation.
Something is rotten in the state of Denmark.
I think this is the right approach. Why do free ground-proofing for a for-profit company? (Who is also so unresponsive). Better to submit fixes to a community project.
A friend of mine just bought a property in regional Australia, and Maps wants to send anyone going to going to his place via forestry roads and then across two private properties to get there, despite having main road access.
Once in another part of Oz, we were sent up a cliff face, in the middle of the night, and it took an hour and a half to correct it's mistake.
Every time I try and "fix" maps by reporting it, and very very very rarely is it fixed
Another property has two "crown roads" on it, but to all purposes, they're very dangerous 4wd trails, again thru private property. The owner has signs, but the belligerent and unaware swear by the Goog, insisting it's correct.
I prefer paper maps over the Google rubbish.
Hiking in Hong Kong, Google maps disagreed with the route I was attempting to take on Sir Cecil's Ride back around the back side of the rock quarry. Google maps took us across the front face of the quarry, past the police shooting range / Explosive Ordinance Disposal Depot, and up a sheer rock face next to the quarry with the assistance of a cobbled-together rope some kind soul left for others, then through a bamboo thicket that cut my wife's legs up. My wife no longer insists on following Google maps if it disagrees with the hiking route I'm planning.
there is a situation similar to this near me and it has caused damages and near fatalities as vehicles moving 50 mph roar down the driveway, round the bend and suddenly there are people, a well head a house and a seafright container up front and center.
It's not just Google. Other GPS maps will send you to my house on a road that hasn't really been passable for 30 or 40 years - if you make the mistake of not telling it to avoid dirt roads. But, if you tell it to stick to paved roads, it'll have you park about 2 km away.
I have no idea how to fix this, except to suggest that they update the maps that the GPS relies on. It's such a remote area that it's not a priority for them. Every year we get a few people who think that GPS can't be wrong. Maybe if one of 'em dies and someone sues they'll get around to updating it.
How does Google tell the difference between someone on an electric bike vs a car going on a dirt road doing 20-30km/h to not have rocks flying everywhere?
With all the sensors in the phone, it is not that hard. There's a lot of research in that field...
The best approach I could think of is they expand the time they're looking at to classify the vehicles before the outlier situations.
It's a privacy issue to use that kind of data, because the employees who build the map shouldn't be looking at one guys GPS traces. So instead the traces are cut into tiny chunks and aggregated and only shown where there are more than 50 people at the same spot. That effectively means there is no data available to make decisions on the most rarely used roads.
The same algo that does that could tell the field mapper to go look at roads that are barely used (or whatever criteria ends up being the best), it doesn't need to tell them how many people or who went on that road.
Sometimes I think that Google create their road maps solely using satellite pictures.
Also, even if they choose to not do this for maps even though they do it for lots of other things, merely aggregate traffic info on roads would be useful. 364 days of zero traffic and one day with 30 minutes of one car would be useful.
> Every year we get a few people who think that GPS can't be wrong.
A bit of a self-fulfilling prophecy.
Probably not. At least one person dies each year in the Mojave Desert because they were following an electronic map instead of a paper map. It's in the local news most times it happens.
In my area, it sees like some of the driveways have turned into roads because of overly helpful contributors.
Fixing it in OSM fixed it slowly in all drivers navigation systems and after a year no deliveries were lost anymore
They can look in some government maps, depending on the licenses.
[1]: https://help.openstreetmap.org/questions/615/can-google-maps...
But they do not make wide use of OSM data, because one of the terms of the license is attribution, of which there is none on any Google product I have seen.
I would guess they are big supporters of OSM as a hiring thing - by supporting an opensource project, you will get young engineers interested in mapping, some of which will then go work on Google Maps.
It also helps any antitrust cases if Google isn't the only game in town, which was the case for many years.
Cue Googler who will find this comment and magically fix this problem. The HN effect.
The right answer is for users not to be brainlessly following whatever their corpo device tells them to be the truth.
Love the whole "nothing can be measured, who can say!?" attitude on Google of all fucking things. They log everything and delete nothing. Yes, they absolutely do know how many people they've routed down that person's driveway, and a subpoena/discovery would clear it right up.
Technically you could make the argument that the map directions are incitement/solicitation to commit a crime. But I doubt you’d get a prosecutor to pick up that case.
Yes, you absolutely could lobby for someone to write a law requiring Google to respond to nuisance requests from their navigation services. That is basically what DMCA does, after all.
Hypothetical example law: OK, you have a good-faith presumption up until someone emails you a copy of a deed and proves it's their driveway, then you legally need to correct it within X days or penalties start to rack up.
If that has been done, then it simplifies the dispute as you can objectively verify whether the complaint matches the official data and Google should not be routing cars there, or if the maps are correct and the appropriate solution would not involve Google - e.g. if a road is physically impassable then complaints to whoever is responsible for maintaining that road; if there are barriers or "do not enter" signs on a legally public road, then complaints to the authorities to remove these barriers and punish whoever put them up, if a path is marked public but you think it shouldn't be then legal action to try and correct it (which may or may not succeed), etc.
You've used the suggest a change feature within Maps? I've done that several times and I think it always worked for me.
Full disclosure I work at Google, not on Maps, and I always did the suggest a change with my personal account.
I've had a road that is mis-named near me. I've used the "suggest a change" feature multiple times, but have never gotten a response other than the automated "it'll be reviewed". Road is still mis-named.
Other changes i've suggested have gone through in hours.
OSM strongly discourages the mapping of trap streets — fake streets that only exist on your map for unscrupulous competition to copy so you know that they did so — but it's possible that some mappers do put these on the map. I don't think one of the big proprietary maps has ever been caught copying one of these though.
I added a placemark in OSM in the middle of Bali, Indonesia for an Internet cafe a few years ago, and it's currently also in Google Maps. (The cafe has since been closed for a few years.)
What's interesting is that's why I added it to OSM - to get picked up by Google Maps. :)
What's even more interesting is that when I told people in SV that I was doing that, they told me since it was a copyright violation, Google Maps wouldn't do that. :) :)
I ask the labelling of a school on the map, specifying the name and providing the home town site as proof (a page listing all schools with their adresses and picture), and two days later it was labelled as “<Name of the town> school” completely ignoring the provided info and making the spot misleading. Requests for correction a few months apart didn’t do anything at this point.
So I can vouch for the process sometimes working, and I have nothing to do with Google. Now, why, when and how it works is a complete mystery.
Even when you're dealing with their non-public products!
Every holiday, Google reminds me I need to check the holiday hours for our business on the Google My Business page. So I do.
Every time, despite being the only authorized user on this page and despite the changes I make being nothing strange (things like being closed on christmas, and not to mention their system asked me to check it)......it sits there saying "pending review" for a week or more.
I don't really know what to think about it when they don't trust the business page's actual manager to get the info correct.
I wonder if it would be easier to just change the street signs rather than convince Google to make the correction.
https://www.wsj.com/articles/czech-prime-minister-wants-to-k...
Even in the same area. A neighborhood name was added, but the street name change in that same neighborhood wasn't.
Obtuse af. Is this default assumption that the user is the one who has to "create" associations in the company's systems based on anything? The big tech cos have been social mapping since before 2012 and they undoubtedly register when two accounts use the same IP.
Creating a system to secretly map my work account and my personal account in order to prioritize my Maps suggestions seems like a huge amount of work for no benefit whatsoever.
E.g. a coffeeshop has a constant stream of accounts on its IP that appear only once or twice. So it might be considered "a public place". So no association is done over accounts arriving from that IP. Your home, OTOH, might have only two accounts on its IP for years, so the association here is stronger.
A cookie shared between two accounts is even a stronger indication, as a person used the same browser to log in with both.
The system might not be created "to prioritize Maps suggestions" only and has multiple benefits in other places. Most obvious of these is to prevent people banned from the service to come back with a different account.
(I don't work at G and never did, but worked at another place where we linked accounts.)
>A cookie shared between two accounts is even a stronger indication, as a person used the same browser to log in with both.
I never use the same browser for my work and personal accounts. But in theory that could run into the spousal problem as well for some people.
In short, the amount of data you are passing to Google, willingly or not, allows them to link these to a single person — you — with an extremely high certainty.
It also makes the simple connection that a user also has a google.com account without doing anything special.
Although I suspect they do have something home built as every single company I’ve worked for with a broad customer facing product has had some flag to note employee accounts. I built this sometimes and was usually to help understand if some random suggestion was from a bigwig worth looking up. You never want to be in a situation where “Ms So and So, EVP of some department you don’t know submitted a question a week ago and got no answer. She mentioned it to me at the EVP monthly hobo hunt, yadda yadda yadda.”
And saying "hobo hunt" shows you obviously aren't engaging in serious conversation.
I was trying to convey that Google just already tracks all linked email accounts so it seems natural that they would be able to trivially identify or prioritize inbound suggestions from employees.
I also wanted to say that even if they did build a custom system to sort employees, it’s not hard. It’s not a super secret tech, but something I’ve personally implemented in an afternoon 20 years ago, so I think it’s safe to assume that it’s easy to do. If they want to.
I added the “hobo hunt” joke as I think infusing humor in arbitrary engagements makes it easier to engage in serious conversation. And all my serious conversations involve humor. Of course, I seriously doubt that any company executives routinely engage in hunting homeless persons and used something extremely absurd and impossible.
You say you work for Google, so you must know that the company you work for amasses huge amounts of information on billions of people on the internet. It then uses that information to sell personalized ads to these people. Advertisers love that, and they pay the rates, and that's how Google makes money for your paycheck — and much more.
What makes you think they "require" any action on the part of these people to "create associations" between accounts? They are making billions by doing it because they can.
Do you have any evidence that Google spies on its own employees for the purpose of giving them privileged manual treatment when they don't ask for it?
Do you have any evidence that Google create associations between separate Google accounts to the extent of treating them as the same accountholder (asside from deleting duplicate spam accounts)?
For months after that, Google bombarded me with the links to the disease and cures and remedies and whatnot.
I remember even seeing that at my kid's gaming PC, and I don't remember ever logging in to my Google account there. We only shared the same IP in the household.
That's pretty anecdotal, I know, but that's what there is.
Why do we allow these tech giants to treat us so disgracefully?
Because we've already been conditioned to accept this kind of behavior from our local governments.
My home address is wrong on Google Maps. It's correct on our national postal services' postcode system, but that only covers public post: most private couriers use Google for some reason. As do Uber/delivery drivers/taxis/etc. I live in an urban centre (Dublin, home of Google's EU HQ) so it's not some obscure rural farm.
I've used their suggest a change feature within Maps pretty much every month since I've lived here and even received automated responses telling me the suggestion has been "processed" or similar. No change to the map though.
I get the problems are both of the same origin, but the discussion has a much better chance of making progress and getting Google's attention which is hard to do when people's lives are at stake. Or when they have been lost directly due to their own oversights.
Seems to be the Google way of doing things: they can't do everything with algorithms!
It happened gradually. First I was directed to a well maintained gravel road, then to dirt track, which forked and forked and slowly faded to nothing.
I was driving a 4x4, had an almost full tank, a load of groceries and 12 gallons of water in back, and plenty of time, so I went with it for quite a while. There are lots of long dirt tracks in this area and I kept hoping that it was still a short cut. The shortest alternate route was around 90 minutes longer.
I think most sane drivers would have bailed at the first turn off of the pavement. I waited until the stream bed sand was getting deep before turning around. For some optimistic mobility impaired person it could have been a death trap. I've adjusted my expectations of Google Maps accordingly.
The worst I've had, which wasn't really that bad, was when I was taking a friend to pick up a canoe in rural NH and they were on a snow covered gravel road. Google routed me on a long drive on said road. Which was OK but still slow going. When we got there, they guy sort of laughed at the direction we had gone and said we should have come in the other way which only involved a mile or two of said road.
Once you get off the beaten track, you really want to tap into local knowledge if you can. Don't just trust Google or paper maps for that matter.
"wrong way" via Basin Creek & Umatilla Creek
https://www.google.com/maps/dir/44.2207582,-119.1331445/Rile...
"right way" via White Creek & Riley creek
https://www.google.com/maps/dir/44.2207582,-119.1331445/Rile...
With the tracks, the journey to our nearest village is 5km, and our second nearest is 7km away - without them, it’s 15km and 55km, respectively - as there are numerous bridges over the multiple north-south rivers here which don’t feature on google - probably quite sensibly, as many of them are in “is this is a road, or a landslip?” territory.
I mean, we’ve been here a year now, I know plenty of routes that look like they must just go off a cliff or something, but cut miles off journeys - but it’s largely been trial and error - more error than anything else when we still had a 2x4 - now that we’ve a 4x4, 30 degree mud-chutes qualify as roads.
There should be a “show me public routes which are theoretically navigable for a vehicle” option - Waze does this, but it’s hit and miss.
Had we continued using Google Maps on our phones, god knows how far into the interior we would've gone before being in real trouble.
I don't know if BC's forest service roads are thoroughly electronically mapped anywhere. Last I knew of, you had to buy a book.
Between maps + satellite imagery navigation was possible.
Well-maintained gravel road seems pretty reasonable in that area of the country. But I'd probably consult my paper map at that point which I would also always have for that area of the country (had I not done so already).
I'll mostly put my trust in Google etc. to find my hotel in Silicon Valley. Not so much to get me somewhere in the rural West.
Why? If it's an area unfamiliar to them then it's very reasonable to assume that Google might know it better than you.
I'm a consultant in a rural area in the PNW and Google maps is literally unusable to access 80% of my client's properties.
OpenStreetMap did the same thing and it's an ongoing hassle. It gets better all the time, but there's a lot of those features in sparsely populated areas.
Sure, you and I know that. But driving around with Google Maps in the city all year gives a false sense that it works great. Then when, once a year, you drive round the back of the mountain to visit Grandma you discover the hard way that Google Maps doesn't quite work so great.
No one at Google validates any of their results. They simply operate on a large data set within the constraints of the machine.
IMO this is the problem with the technology industry; the only correctness that matters is the services are online.
It’s similar to the stock market being completely decoupled from the real economy. Oh yes all those math operations were handled just so, but you drove the lemmings off a cliff.
I’ve adjusted my expectations to society accordingly; lemmings.
Taking a few minutes to zoom out and check the actual maps in terrain view would save most people on these journeys.
If you don't have data access then Google Maps routing won't help much either. It all comes down to offline/hard maps, situational awareness, and basic wayfinding to get through these areas safely.
EDIT: what are people disagreeing with here?
Streambed: https://www.google.com/maps/@34.991622,-108.0730561,94m/data...
Lightly used dirt roads: https://www.google.com/maps/@34.9927729,-108.076162,158m/dat...
As for the data issue, Google maps (by default?) caches the turn-by-turn directions when you start navigation. Spotty cell service isn't really a problem until you make a wrong turn.
Zoom out, assess the terrain, look for markings, check for buildings, and compare paths to the destination. Unless you're visiting the bottom of the crater there, you shouldn't ever be going off that main highway.
The map tiles should also cached (how else would it show you the turns?) but yes offline/paper maps are a must when going into unfamiliar areas.
That said I do think people should carry a $15 paper map booklet if they plan on venturing outside the city. I both have offline maps of the whole US and paper maps in my car, but I also do this a lot.
A friend of mine and I rented a AWD vehicle to be able to drive on the "F roads" in the internal part of Iceland, which is beautiful but very desolate. (I'd be surprised if the inspiration for Mordor didn't come from some places Iceland in fact.) I had downloaded the OSM maps on my tablet and was just manually plotting routes that looked like they could be interesting.
It was great fun, but it sort of turns out that there are F roads and there are F roads. Some of the roads are fairly flat, easy to pick out, and you see a car at least once every hour. Some of the roads... well, one road went steeply up, then without warning into what was basically a giant sand bowl half a mile in diameter. We were in the "bowl", with sand who knows how deep (deep enough that the car had trouble making forward progress) before we knew what was happening. In the middle of nowhere, with no cell signal and not having seen anyone for hours.
Thankfully, by turning around and just keeping the accelerator on, we were able to build up enough momentum to get back over the ridge by which we'd come in; but it was definitely a close shave.
I could imagine the same story as TFA happening with OSM maps.
If such tags are present and seem to be reliable, then OSM can be used to plan trips. Otherwise it is just too risky to rely on OSM. In particular in harsh climates and if there is no way to get help.
I.e., before using OSM, assess the quality of the mapping done by the local community. Anyone can draw lines of a map by looking at satelite images. It doesn't mean that such lines are passable.
https://www.strangeoutdoors.com/mysterious-stories-blog/2017...
https://www.google.com/maps/dir/Gold+Beach,+OR/Galice,+OR+97...
There's still some odd behavior around bicycle routes. For instance, Google does not seem to show bike routes down CA-1 in middle California. I'm not sure what's going on there--hundreds of people do that route every year.
Yeah, there's a reason roads in that area are so open.
I do most of my intercity navigation trough state and road signage, with the navigator on but only for the last stretch, like driving me to the address once we're close.
Around here, "navigator shortcut" had already become a derogatory term for their inane suggestions.
There's some that provide a truck mode, but I don't know what weights they have on their algos.
Via Michelin used to have a main roads mode, but they didn't give turn by turn back then only planning.
(https://images.trafficsupply.nl/imgfill/800/800/i-114795-69c...
I suspect it would take me a moment to process this into some action.
It started off a reasonable road and quickly became dangerous and then deadly. I rolled down the mountain side with the motorcycle and got very, very lucky that I was only mildly hurt (on top of the brutal sunburn I got on the ride). I ran into the local village (more like hobbled) and the first two people didn't believe that anyone could be stupid enough to do that (pantomiming because no one spoke English). I needed to find a teenage boy before anyone thought it plausible.
Happy to live to tell the tale, and I certainly won't be following maps blindly anymore.
[0] http://www.cultofquality.com/index.php/2018/02/on-glory-and-...
In September, I was in Hunza, northern Pakistan. The mountains are so tall, like 7k to 8k+ meters tall, I couldn’t even get a GPS lock most of the time. Even with a downloaded, offline map, it Google Maps was less than useless for navigation (still worked OK as a “paper” map).
I've forwarded it to a few people. It's also made me slightly sad about being stuck in the UK lockdown for at least the next few months. I love South East Asia.
We're doing well despite the pandemic; its been interesting to see how adoption of machine learning in consumer research has been accelerated during COVID because its no longer possible to gather lots of people together for tastings.
It was a surprise to get into such thick jungle so close to civilisation.
Though, I myself though I will never get into such a stupid situation, and probably those two thought of the same.
As somebody who spent childhood in Russian far east, I think wasn't even told not to fare into forests for it being so obviously dangerous as shown by lengthy necrologues in style "Ivan Ivanov 1980.1.1 went to forest on 2000.2.2, never returned, presumed dead"
Just a couple weeks ago, we located and brought back a middle-aged couple who went out on an afternoon e-bike ride in the backcountry. They found a nice 20 mile loop of singletrack and forest service road on the AllTrails app and headed out with minimal clothing or supplies. The trail was beautifully maintained, taking switchbacks down from the ridge into the valley and to a campsite, but beyond there, it was in disrepair, overgrown, washouts, treefall, etc. But they kept pushing on - the loop they had planned was on the map after all. But no, the second half probably hadn't been maintained much less traveled on in 5+10 years and soon it completely petered out for them and they were left, deep in a valley, at night, with the temperature dropping, in shorts, with no water or food (he was a diabetic and they both had heart conditions).
My team dropped in from the far side which was closer to their location but even with multiple maps and satellite imagery could not locate the trail for any significant stretch which led to several miles of late-night bushwhacking down steep hillsides to reach them. Once checked out, provided with warm clothing, food, and water we all hiked out the way they came in which took the rest of the night and we reemerged at the trailhead at dawn. They were very lucky that due to the orientation of the canyon, they had cellphone coverage.
Happens fairly often.
Another one, last summer, was a guy who set out on a 3 day, 40 mile loop, got sick the very first day and instead of turning back, kept pushing on and was found severely dehydrated and still ill 36 hours after he was supposed to have completed his hike. He spent several days in the hospital...
It has lots of great stuff about lost person psychology, search strategy, persistence, local knowledge, and extremely remote areas.
Highly recommended.
[1] https://www.otherhand.org/home-page/search-and-rescue/the-hu...
He worked for CNET, and it was a big story back in 2006.
https://elpais.com/internacional/2020-12-23/perdidos-en-la-t...
Here is a video of a drive along it between the cities mentioned in the article:
https://www.youtube.com/watch?v=J46vglp6g98
Insane cold in this area - I'd be curious what type of preparation they had for this, and why not turn back if things were impassable?
Google in the USA maps lots of roads that are 4x4 only (but totally fine on 4x4).
Is this really a ghost road with no use in 20 years? I could pretty easily find some towns that would seem to be connected with it. Also plenty of adventurers and motorcyclists (??? how cold would that be?) seem to take it?
Did they by ANY chance turn down a farm road or some other side track?
A brief check of sat views doesn't really match the "abandoned for 20 years" claim being made here.
https://siberiantimes.com/other/others/news/tragedy-in-yakut...
Edit: It was pointed out in the comments below that Google maps does go on the correct route now. It is the same general road, but there was a section called the "Old Summer Road" that had fallen in to disrepair but was shorter.
Here is a link that illustrates the difference: https://cdn.newsapi.com.au/image/v1/481952e93a2fed32c94a09bd...
Google maps now follows the red dotted line in the image if you enter the route.
Judging by a couple of photos I could find on Google Maps it looks like in the summer the road is barely accessible using vehicles like the famous UAZ-452, I'm talking about this photo [2] showing a UAZ-452 crossing a river. In the winter I saw that these two people did it with a heavily equipped Toyota off-roader [3].
[1] https://en.wikipedia.org/wiki/Toyota_Chaser#6th_Generation_(...
[2] https://www.google.com/maps/@63.351051,140.967979,3a,75y,90t...
[3] https://www.google.com/maps/@63.258317,143.201208,3a,75y,90t...
If you dress right, have a good thermal sleeping bag, have calories in your system, you can do pretty well. A fair number of sleeping bags have survival temps as low as negative -40. I've done OK once or twice at surprising levels of cold (sleeping in the back of car / SUV on a ski trip other winter trek is not end of world, even without cars heater).
Not totally clear still why this is a story about google.
In summer there are sections of mud pit, swamps, river crossings because the road is unmaintained. Anything with at least three wheels and a heated cab is fine in winter.
If the vehicle was actually unsuitible they would have gotten stuck within walking distance of town instead of putting a branch through the radiator trying to make an N-point turn in the bush after going the wrong way (probably down some little side fork).
https://www.openstreetmap.org/directions?engine=graphhopper_...
Even a fairly straightforward off-paved road route like to the Racetrack in Death Valley, sees the local Jeep rental/tour company rescuing standard passenger car drivers all the time because they get tire punctures with no or crappy rental car jacks, etc. (Or they change tires and the doughnut spare lasts for 5 minutes.)
I've fiddled with Google Maps in the Death Valley area and my experience was that it wouldn't take you on stupid routes just because they are shorter. But if you give it a destination that can only be reached by a challenging road, it will route you on it.
The sharp rocks of Death Valley and other super arid parts of Nevada was not something I had really thought about until I saw a picture of the car from one of my friends in Nevada. In addition to the normal full-size spare tire, there were two other tires stuck in the back of the car. He worked in a mine and apparently the road up there was full of these sharp rocks so there were days (not often) that he had to change 2 tires during his commute.
I imagine that it is particular type of rock prone to fracturing with sharp edges. But I also wonder if the sharper rocks more common because of the super arid conditions and are not worn down by the weather.
Several people have done the road and other remote dis-used highways on motorcycle (in summer/autumn when it's more difficult but the weather is less brutal). The AVrider forum has several threads with lots of pictures.
Traveling these kinds of roads is common in this part of Russia. I would bet money that they had driven the "correct" route before and then when Google said the old road was faster they knew exactly what they were getting into and figured they'd give it a try because it might save time. Their mistake was not being prepared.
http://sherrijosbecauseicanworldtour.blogspot.com/2010/08/ol... (There's a part 2 + 3)
And the other thing - I really wish they waited just a year or two with this adventure, the tech they used was literally on the verge of maturing but 90% of their problems were due to driving prototype vehicles. Literally just a year after filming, there are both electric cars and motorcycles that don't run effectively alpha firmware.
If they had delayed too long they may have run into covid times and that has ruined an awful lot of things.
From what I can find, Kolyma Highway is still used and not disused since the 70s, as the article says.
AUSTRALIA - definitely not a "road" more like an unused hiking / herding path with nothing but lazy kangaroos in sight.
CHILE - down an unpaved half-lane road making its way down the side of a cliff, marked "mucho peligroso"
ITALY - completely off road over giant potholes and big rocks. Middle of the night.
ABU DHABI - across a clearly pedestrian plaza.
In all cases it worked out - as it happened to be traversable in my tiny rental car and indeed the way to get where I am going. But that's in retrospect, could have easily been "death by GPS" types of scenarios.
A few things I recommend when driving in places like this: - refuel religiously. Don't let your tank get below half, as theoretically that should give you enough to backtrack to where you got gas last time. Not always possible of course. - try to ask yourself - if I had to back out / turn around, could I? And be very leary of locking yourself out of that option (although easier said than done.) - be mentally prepared to sleep in the car. At some point it is safer to do that and figure things out in the morning, even of there's a chance that your amazing tuscan villa airbnb is just another mile down the washed out road (cuz it may also not be) - think of it as a fun adventure and don't freak out.
First, the highway of bones I think refers to the Stalin-era political prisoners that died building the road.
Second, Kalyma colloquially refers to an incredibly far away place. Like in English you may say "it's on the moon" to mean that something is ridiculously far, in Russian you say "to Kalyma"
Fundamentally, you don't end up anywhere near that region randomly. It's not like they were going for milk and died on the way to the store. It's hard to really understand what they were thinking - but there's no way they didn't understand that the whole area was ... challenging.
The other commenters is right that the name comes from the prisoners who died building it. Allegedly they were buried in the road. I say "allegedly" because everyone who's ever built a road knows that the road base is key to longevity and that adding things that will decompose to it is the opposite of what you want to do.
Oh, that's just how roads are in South-Eastern Europe.
The Wikipedia page also says that the road is called the Kolyma Route "since it is the only road in the area and therefore needs no special name to distinguish it from other roads"
[0] https://en.wikipedia.org/wiki/R504_Kolyma_Highway
Edit: Okay, it was pointed out in the comments below that Google maps does go on the correct route now. It is the same general road, but there was a section called the "Old Summer Road" that had fallen in to disrepair but was shorter.
When the road was upgraded, the route was changed to bypass the section from Kyubeme to Kadykchan via Tomtor, and instead pass from Kyubeme to Kadykchan via a more northern route through the town of Ust-Nera. The old 420 km section via Tomtor was largely unmaintained; the 200 km section between Tomtor and Kadykchan was completely abandoned.[6] This section is known as the Old Summer Road, and has fallen into disrepair, with washed-out bridges and sections of road reclaimed by streams in summer. During winter, frozen rivers may assist river crossings. Old Summer Road remains one of the great challenges for adventuring motorcyclists and 4WDers.
Google locates the park address as 40000 Paseo Padre Pkwy, Fremont 94538. The pin is dropped in a fine place, at the building by the Lake Elizabeth boat ramp.
I mostly know where I'm going anyway, and as I come up Washington Blvd I notice it's asking me to turn early, so I follow just for kicks. Inexplicably, gmaps directs me to an area about a mile to the south of my destination, at the end of the Railroad Ave trail (https://goo.gl/maps/G5xV7MDemtbtmNYRA). It then appears to suggest that I park and walk the trail for a mile to get to my destination. In the middle of downtown Fremont! When using driving directions!
The weirdest part is that gmaps clearly knows that it's taken me to the wrong place. There's a dotted line instructing me to walk a mile along a footpath. But then, why not just take me to the actual destination address in my car, which has a huge parking lot? I can't wrap my head around what's going wrong in this scenario.
This wasn't a fluke. I reliably get told that this is the best way to the park. If gmaps can be this wildly wrong in the middle of Fremont CA then I can only imagine how bad it gets in rural areas.
In the future, if you encounter a similar situation, consider reporting it (https://support.google.com/maps/answer/6194894). Any user feedback is greatly appreciated!
Actually I would also have been happy to report it to Google a few years back but I have to say our relationship seems to have turned adversarial in the interim and I'm fairly sure it wasn't my doing.
"Tap Suggest an edit or Report a problem."
(Note: I don't work for Google, and haven't for many years. I mostly noticed this change because it finally fixed a longstanding directions issue that I first reported ABOUT TEN YEARS AGO, involving directions to the Pittsburgh International Airport.)
It is completely conceivable for an app to say: "we don't know, please use a different app" -- it's just pure greed for user engagement that prevents owners from doing this.
Similar life-and-death situations have happened in the past with Google Maps users in Middle East.
The epidemiological concerns haven't gone away and there isn't much interest in building across the gap.
That's fascinating!
Wikipedia shows that the disease is nearly eradicated in South America:
https://en.wikipedia.org/wiki/Foot-and-mouth_disease
I suppose there still wouldn't be an economic incentive to build a road when there are two large bodies of water on either side?
So they are neither categorically opposed to warnings for any of the reasons you mention, nor are there technical difficulties. I guess the case of roads that are dangerous just happens to be extremely rare.
Another time in I encountered another tourist in the middle of nowhere in Nevada, totally lost. Again, blindely followed Google Maps.
In Death Valley I crossed tourists in a rented SUV who wanted to cross over the Lippincott Pass to get to the Racetrack Playa, again following Google Maps. I talked them into turning around, there was no way they were going to make it. Also, they were not prepared at all (no water, food, etc.)
In Ecuador, Google Maps lead me down a mountain road which turned into a mountain bike track. I made it out in one piece, it made for a good story[2] but it is pretty damn dangerous.
If you are in a remote area, always double check google maps suggested, eg. with an OSM based map.
We were cruising through the desert in mid-August, well stocked up with water as everyone told you to always pack twice as much as you think you'll need, and mildly pissed-off by the gas price of the only gas station we've seen that day. We barely met another car on the whole journey through. It was late afternoon, we were about 30 minutes away from getting out of the desert and we saw a man waving at us at the edge of the road. We pull up and we see he is desperate, his car got stuck in the sand a few meters away. Together we managed to dig a bit under his wheels, gathered some flat stones and shoved them under the wheels to improve traction. After a couple of tries he was able to get the car out of the sand and back on the road.
A few years back we travelled with friends to a mountain resort for the new year's eve. Instead of choosing a longer, well maintained primary road to our destination, it chose a slightly shorter secondary road through a high mountain pass, where it was snowing like hell and obviously no ploughs were around.
I was a passenger and wasn't involved in route selection, but the lesson is that it might be good to take time to discuss this before. Also, zoom out before driving and see high level plan instead of blindly following first-person view of the road. Or, you know, old school low tech solution: call the locals and ask for best road.
In one such instance, I didn't realize what was happening until I had spent 45 minutes inching past the chairs of patrons in tiny cafes not a foot from my side-view mirror to eventually emerge from an alley onto the plaza of a cathedral in the city center much to the confused looks of the passers by.
Even in well-populated, first-world cities it seems it's wise to take Google Maps' advice with some healthy skepticism.
FWIW, a few years ago Google Maps also gave me wildly inappropriate directions on the way to Yosemite (i.e., in Google's ~back-yard). I called it off when it suggested I start off up a dirt road that switchbacked its way up a cliff face in my mother's minivan.
If I want to find the nearest ice cream store open at 9pm use google maps, it's a great location database. If you want a map use OSM.
Many tourists drive with their normal rental cars into very remote areas.
So at the end I use OSMand+ (Openstreetmap) always and Google Maps only for searching.
We arrived at a roundabout, and saw two signs, both pointing to our destination -- we could keep straight, or turn right. The old GPS told us to take a right, which would put us on a course to reach the hotel in an hour and a half. Google Maps, on the other hand, recommended going straight, in a more direct route that would get us there in 40 minutes. We'd had great experience with Google Maps in Sicily so far, and opted to try the route it suggested.
Driving past the roundabout, the road became a bit rougher, but not outside of the norm for the area. We wound around the mountains in the dark, and Maps continued to assure us we were getting nearer to our destination. The first sign of trouble came when we took another right turn, some time later, into a smaller road. We carefully drove around the first pothole. Then came another, and another. The road kept steadily getting rougher, but the car was ok, and we were driving carefully. There were absolutely no lights anywhere to be seen.
After a few minutes on this road, we started following it as it curved left -- and came face to face with reeds taller than the car. Our road vanished into them, turning into an old, rough path that had clearly not been used in years. I got out of the car to look around -- and could see absolutely nothing not illuminated by the headlights. No house lights, no street lights, no sounds except crickets. No cell service. Maps said we were only 15 minutes away from the hotel.
I got back in the car up to turn around, and the car wheels into mud and started spinning. It was at this point that my SO started freaking out. I don't blame her. We managed to, very slowly, turn the car around, drive up over the potholes, and make our way back to that roundabout. This time, we went the other way. Eventually, we reached our hotel, which was pitch dark with no outdoor lighting, had peacocks running around, pictures of Jesus nailed to the room walls, and a deserted wedding banquet hall for a restaurant, but that's a story for another day.
I still wonder where we ended up -- what that road looks like in the light of day.
When there is a serious snow storm or blizzard Google Maps will inevitably tell me to use an old (but still used) county road, rather than the interstate. I assume that google thinks this path has less traffic (true) and is faster (very much not true):
1. Due to weather conditions it is very much NOT faster, but I suspect due to the lack of traffic it seems faster than the slow interstate. Plowing through this county road is quite slow in reality, but the recommendation always remains...
2. This is NOT a road you want to be on during a blizzard or heavy snow event. It is not plowed as often as the interstate, and there is little to no shoulder so if you slide around even a little there's a super high chance you end up in a DEEP ditch.
Even worse sometimes it redirects people through local neighborhoods that are plowed even LESS often, presumably because it assumes the lower amount of traffic = faster, again that's wrong on every level.
In reality plodding along on the heavily plowed interstate is both faster, AND safer, but Google Maps seems to be unable to recognize this during these kinds of conditions.
The road network in Sweden is comprised of 36 million+ distinct coordinates that form 3 million+ roads which are joined on 2.4 million+ nodes.
In the first half of 2020, 16 thousand+ class (funktionell vägklass) 9 roads were added to the network, consisting of 2,769 km of road. Class 9 roads are the highest classification of roads in the Swedish road network and is loosely defined as "where vehicle traffic normally doesn't occur but is totally possible". Class 9 roads compromise 38.2% of all road links in the Swedish road-network for vehicles.
This represents the lions share (64.22% number of road links, 75.33% km of road link) of the changes in the road network during the first half of 2020.
This data originates from the NVDB service of the Swedish traffic agency and can be viewed here on a map: https://nvdb2012.trafikverket.se/SeTransportnatverket
A radiator is not necessary for a car to work, specially in a freezing place.
A radiator is needed to refrigerate the engine. In a very cold place the engine is refrigerated just by natural convention. Probably you will have to stop and open the car hood from time to time.
Without globes touching metal could be dangerous, but it is strange. Traveling to very cold places,what you notice is that globes are not accessories, they are vital elements.
Seems like at the very least they could run the engine as long as possible until its about to overheat, then turn it off and it will very quickly return to operable temperatures in the very cold conditions. Certainly better than dying outside in Siberia!
Might still be fine, but... By the time you know it ain't, could easily be too late.
Maybe this plan could work if you drove 2 minutes at a time and kept the hood up. But some of the components in the engine heat up pretty fast.
The truck had a functioning radiator and appropriate coolant levels. The only thing wrong was the thermostat. Even in that configuration, the temperature gauge climbed into the "danger zone" if I neglected the brake-slamming trick.
https://en.wikipedia.org/wiki/R504_Kolyma_Highway
> Locally, the road is known as Trassa (Russian: Трасса – "The Route"), or Kolymskaya trassa (Russian: Колымская трасса – "The Kolyma Route"), since it is the only road in the area and therefore needs no special name to distinguish it from other roads.
Update: This image makes it clearer: they were told to go on a disused stretch of road, which is where they got stuck.
https://cdn.newsapi.com.au/image/v1/481952e93a2fed32c94a09bd...
https://andrew-max.github.io/adventure/2017/03/24/huamachuco...
Google maps would reroute me around perfectly good highways because some goat trail somewhere was technically a shorter path although the reality was that they were 10x longer because of how difficult the terrain would be and if you were to brake down on any of those roads, you would immediately be in a survival situation, it was incredibly remote.
It finally happened to me although luckily I was able to find an isolated farm but it's not hard to imagine if things went a little bit differently it could have turned into a life or death situation
Wow.
Considering both cities are some of the coldest in the world and the journey is 2000km, this really seems to be the result of poor planning. Part of the road contains driving over a frozen river, perhaps map apps should alert on crossings like this?
What's more interesting is that the vehicles that travel this road seem to not be in super great condition and for some reason I did not see any instance of chains.
Truth is, after 500-600m what remains of the road, more of a path, is totally overgrown and after ~900, ~1300 and ~1700 there are fences (I went for a hike down that path with a friend after repeatedly being guided down that path by google maps).
I only use OpenStreetMap for planning my hikes and bicylce routes now.
Google Maps gives me the new, longer route from Yakutsk to Magadan (34 hr, 1161 mi) but I can recover the old route by adding a stop in Tomtor (40 hr, 1077 mi).
This has been an issue with Google Maps for some time: being more optimistic than e.g. Apple Maps just so you use the service, as suggested by this article: https://arturgrabow.ski/2018/02/19/navigation-apps/
I will use Google Maps to plan routes, but I will also use common sense:
• I will look at the route on the map to become familiar with it
• I will use street view to see what each turn looks like (street view is also useful to see if the road is which needs a 4x4 with high clearance and pure rubber tires which do not use air to successfully drive)
• I will memorize the map (making simplifications as needed)
• I will then use the route
If one just lazily puts in a destination and depends on their phone to do all of the work for them, they can and will encounter problems. One time, I was with a buddy who does that, and the route given to us was horrible, so I just looked at him and said “Ignore what your phone says. I know how to get there.” It ended up he accidentally left “do not use freeways” enabled.
Mostly, I don't sweat it so long as I'm doing something straightforward although I find routing tries to overoptimize.
Great, the map app shaved a couple of minutes for some drivers in a hurry by putting hundreds of kids' lives in danger.
Mapping apps ought to be subject to some sensible regulation to avoid cases like this and the one in OP.
I've just tried to suggest a change, but as it happens often with Google documentation - the description of what to click matches nothing (there's no menu item labelled with the phrase from documentation - I use Polish version).
Edit: found it, but docs are misleading.
Then again, I got a land cruiser 3 years ago (with a 24 gallon secondary gas tank) so if a gps says to try something these days I usually do!
My understanding is that most U.S. ski areas are on leased federal land formerly leased to logging companies, and the resorts go into the clear-cut areas, remove stumps, put in chair lifts, etc. The cat tracks are the remnants of the logging roads used to get the equipment in and trees out.
It recommends alternative routes to "save" you a couple of minutes. Dirt tracks, unmantained roads and sometimes danger neighborhoods (favelas).
They should fix the algo to prefer the main roads and streets. At least give a configurable threshold. It seems quite doable with all GPS data/street images they have.
Google has many faults. But the story is not totally one sided.
So yeah, when the road does not seem right follow your gut feelings.
I get the problems are both of the same origin, but the discussion has a much better chance of making progress and getting Google's attention which is hard to do when people's lives are at stake. Or when they have been lost directly due to their own oversights.
To get to the in-laws place for Xmas, I used to drive my family on a good highway, but it went through a 50-mile-wide frozen swamp with little traffic on it. Year after year. So risky ... suppose I hit a deer. Now I wonder why in the hell I did that.
All maps are going to have limits - computer maps just moreso - overtrust is always unwarranted.
Generally, the best way to find bike routes in Japan (or most countries) is to ask Google for car directions and avoid highways and toll roads. But this time, it drove me straight into the radioactive danger zones of the old Fukushima plant. You were only allowed on those roads if you went by car, as a radiation shield. So I got stuck in a dead end, in a literal ghost town, trying to hitchhike. Man, I still remember the frustration and anger. Weird adventure.
For road trips or hiking in sparsely populated areas, though, it’s treacherous to rely on automated routes. They make a path seem more reliable and well-known than it is.
A fairly straightforward and generalizable way of implementing a fix to issues like this could be by adding probability estimates and corresponding warnings to suggested routes. A long hike/ride through not-so-well-trodden path with scant data accumulated? Mark it, visibly.
I do see now that the Mount Butler HF radio station is no longer marked "Gas Station" (hint: barbed wire and guard dog warning signs aren't common at gas stations) on Google maps, so at least some things have gotten better. At least the desktop maps also have some indication of trail quality, but it shows as a high quality trail the most direct trail from Explosive Ordinance Disposal Depot at the Quarry Bay rock quarry to Jardine's Lookout on the back side of the quarry. That trail includes a scramble up a rock face with the assistance of a cobbled-together sketchy rope, and trail blazing through bamboo leaves that will give you paper cuts. My wife now refers to that path as "I will kill you Hubby Mountain".
When we first moved to North Point, my wife and I were looking for an alternative route from our place up over the mountain range to Stanley. Google maps ended up taking us to the Morning Garden (fair enough), but then through a bunch of knee-length grass where the trail was apparently only virtual, down to the front face of the quarry in Quarry Bay, past the police shooting range and the EOD Depot, up a narrow concrete flash flood control ditch, and up a rock face assisted by a cobbled-together rope some kind soul left for others, and finally a scramble to meet Jardine's Lookout at the back side of the quarry. It wasn't that dangerous, but it was very rough and my wife wasn't happy about bamboo leaf cuts on her legs. Google maps really should have had us continue on Sir Cecil's Ride until it met the Wilson Trail behind the quarry, instead of taking a shortcut through Morning Garden.
We always take plenty of water and download the government map PDFs to our phones. I have plenty of experience navigating the wilderness on the US-Canadian border. I have plenty of orienteering / map reading as well as dead-reckoning experience. Still, it'd be nice to have a trustworthy navigating app to double-check my route.
I bought mine in the General Post Office in Central, near IFC. There's also the maps shop at the junction of Nathan Road and Gascoigne Road in Yau Ma Tei, Kowloon
1st it took me to the Paris - Castellane it took me to the French Alps. Using a park road. I guess the shortest route between 2 points is up and over the mountain.
2nd time we were in Sicily and at some point we ended up driving through a golf course, and a herd of goats. It was super strange.
There was also a 3rd incidence in Malta. The route was down a walking path. The road got narrower and eventual just wide enough for 2 humans.
Never used google maps for maps since, it's way too unreliable. The only use i have for it is to get business opening hours
I always review any route Google gives me and 95% of the time I make some sensible adjustments that get me there quicker and easier.
I often wish the app had a "simplest route" option.
While it’s a great product, I would not put my life in its hands.
https://www.theage.com.au/technology/online-map-linked-to-fa...
What a garbage story. An online mapping service could have caused this. But we don't actually know.
ADDED: And as I noted in another comment, they were in fact just using a paper map (which is what I remembered being the case at the time).
One of which was following what looked like a normal road out of town and ended up being a full day of crossing a mountain range, in the middle of nowhere, on a path where I had to push the motorbike up the hill for parts of it and cross a river half way through. Exhausting day, but amazing none the less.
Would you please review https://news.ycombinator.com/newsguidelines.html and post only in the intended spirit? We'd be grateful.
If you'd please review https://news.ycombinator.com/newsguidelines.html and stick to the rules when posting here, we'd be grateful.
Telling people how to have their own adventure seems like something a keyboard warrior would do...
The article appears to be almost devoid of information about Google.
My recollection from the time was that GPS mapping wasn't involved and, in fact, the Wikipedia article says: "Mrs. Kim told state police that they had used a paper road map, an account supported by the Oregon State Police, which reported that the Kims had used an official State of Oregon highway map. Mrs. Kim later recounted that, after they had been stuck for four days and were studying the map for help, both she and Mr. Kim noticed that a box in the corner of the map bore the message: "Not all Roads Advisable, Check Weather Conditions"."
I lived near a gorge where there was a nice modern road alongside an old horse track that was only intended for use by the residents connected to it. It was technically two-way (as making it one way would annoy the residents who might want to go either up or down) but barely wide enough for one car. If you met another car halfway, you had to reverse until you got to a driveway you could go into and let the other car past.
At some point Google Maps' algorithm changed and started suggesting that everyone use the old road. There were multiple crashes, and the local government begged Google to deprioritise it, but their answer was basically "our algorithm is always right". Five years on, Google Maps will still suggest it, so the local government added big signs at each end advising drivers to ignore Google Maps and to use the main road.
Although they already weren't suggesting that route, the other map providers (Apple, Here, TomTom etc) contacted by the local government were happy to proactively add a flag to the road saying that it is for residents only - so it will only route using it if the source/destination is that road.