Slashdot Mirror
Who Makes MapQuest's Maps?
carpoolio writes "TechTV has an interesting story about the company that builds the mapping technology behind popular map services like Mapquest. The company, Navigation Technologies, is decidedly low-tech in its approach to making its maps: two people in a car drive around endlessly, inputting street information and landmarks into databases. Navtech's map databases are used in everything from Garmin GPS units to Alpine in-dash auto navigation systems. So next time you turn the wrong way down a one-way street, know that there are real people behind the controls."
With a GPS receiver in many cell phones we need to figure out how we all can collaborate on creating maps. Here is a map I created with the data from my cell phone over the course of a couple of months. If everyone contributed instead of the data from a few people driving around we could pool the collective data and have great, open maps. This service is free until the end of the year, if everyone who can signed up and we pool the locations we would have a great map (not to mention traffic info.)
Free cell phone tracking
On the other hand the census bureau is planning on having a new improved database for the 2010 census that includes every home in America with relative precision in the centimeter range and absolute precision in the meter range. Some of the tech that they use for this is VERY cool stuff.
You can start learning here.
The policy of the United States is worse than bad---it is insane. -- Ludwig von Mises, Economic Policy(1959)
I want Navtech to team up with a couple of the large carriers, like Schneider National, Werner Enterprises, JB Hunt and the other large trucking companies. These 3 companies, and many more, already have GPS transponders in the truck that track their locations and report back in realtime via Satellite.
Now, when a driver sees major road construction, etc, on major interstates they simply hit a button on their QualComm OmniTracs unit marking it as such. After so many drivers have done this, it marks the area as being under construction, with a little bit of info about what's going on (resurfacing, 3 lanes closed westbound from 9pm-4am at milemarker 139 to 177 until 12/16/03) and mapquest inturns adds that data to it's routing database.
This would be an excellent way for mapquest to add a pay-for service that I for one would definately use.
Looking for hardware (Currently need: Large Etch-a-Sketch) Have one? See my journal!
My subdivision is a little over 3 years old. We're on all the other maps, but not MapQuest. Repeated emails to them to get this corrected have gone nowhere.
Odd thing is, if I map to the Albertson's near my house and then scroll down in their map, I see our subdivision. If I map to our address, none of the streets display.
Kinda sucks when you tell someone you need to give them directions when they say they'll just map it using MapQuest.
At least MapBlast works. Whether or not it'll do better now that Microsoft owns them remains to be seen.
Their LineDrive maps are better anyway.
Cruising the internet on my TI-99/4A @ a whopping 300 baud!
Seriously, this is both interesting and disappointing. I've been working, as a hobby, on a Palm-based GPS mapping program. The reason I'm not making much progress is because even when I'm done it's not going to be very useful without map data which is probably not available for free. I had hoped there was some hi-tech way to snag decent map data (at least the roads themselves) perhaps by digitally analyzing satellite photos, etc. But this is a low-tech approach which certainly suggests to me that there's no realistic way I could come up with nationwide road data for my Palm app.
Oh well.
Sure there is. What we need is an "open source" map. Have anyone who wants upload their GPS "track" data to a central site. A little data massaging will be able to use the average of plots to determine major roads/highways, and a few volunteers could add names and addressing schemes. Maybe the individual users could even supply those if they wanted, with another averaging system to determine the correct name of the street based on percentages...
It could work. Would be a major, major project, though.
End of lesson. You may press the button.
The graph algorithms to calculate these routes are very mature. If there are problems, they're either from the algorithms being implemented improperly (unlikely, as they're actually quite simple algorithms), or there are issues with the data. Most likely it's the latter.
I can't imagine life without being able to get directions, and a custom-made map, to somewhere an hour and a half away that no one's ever heard of. I really don't think I could navigate with a paper map anymore.
But technology's most amusing when it all blows up. I wish I could find the link, but I distinctly remember reading about some lady who tried to plot an intracity voyage, and got routed through about 12 states -- even venturing into Canada for a while. (Does anyone else remember this?) And someone I know was talking about how on a recent trip, he tried navigating only by GPS; it worked perfectly, until it had him turn down onto what was a dead-end street. It turns out that the GPS assumed he could drive about 100' through the woods, up a steep embankbent, to get onto the highway. (I suppose it would have been a convenient shortcut, if only he had been in a Hummer and had a chainsaw for those pesky trees.)
________________________________________________
suwain_2
We're already way ahead of you. I'm one of the developers at project OneMap. We are currently building and serving one of the largest, free repositories in the world, completly built on open standards. We serve content in the fashion of GML and store everything internally as XML. We've integrated quite a few sources so far, both from a few custom norwegian sources and from the US TigerLine-files.
The main goal is to be able to update and review the content of our repository from within your own browser -- and we have the infrastructure to solve this. The biggest problem being that no-one has ever done anything like this in such a large scale, so we're kinda going along and feeling how the ground is all the way.
Our gateway (for viewing the maps) are currently built on SVG and utilizes the open, formatted GML response. The source is going to be opened up and everything is going to be available for free, but currently we're having a few issues we would like to solve before going public. As always, this is a work in progress. I'm probably doing my MSc with just this topic (updating a map by many individuals) and a way of making sure that only REAL changes go into the repository.
mats
One man's ceiling is another man's floor.