How Google Builds Its Maps
theatlantic.com
theatlantic.com
Meanwhile, there is another program to build the world's best maps, which is quite the opposite of secretive, involves hundreds of thousands of volunteers around the world, offers public dumps under a free license, and has already outperformed Google for some areas and use cases: http://www.openstreetmap.org/
OSM might not have the computer vision tools or the Street View imagery, but there's still plenty of work to do with the existing tools and data.
As an aside, are there any efforts afoot to decentralise OSM? It seems well suited to a geographically distributed database, with each country/area looking after it's own map, and the whole being drawn together by a common markup language. Assembling a global map would be a matter of crawling a network of servers, rather than downloading from a single source.
A simpler use case for computer vision I could think of with Openstreetmap would be to automatically extract roads and building shapes from satellite imagery (though I hope that the license to use Bing would extend to something like this).
> As an aside, are there any efforts afoot to decentralise OSM?
Well, even Wikipedia's not decentralized yet and has no obvious plans to, so...
The project is a great start, but has basically been abandoned.
But you're right—more would be better. I'm a little surprised MS went for it, but I doubt Google will.
Edit: I see that "my places" now has a "Home address" which isn't pre-populated with my google account confirmed home address for some reason.... But if I set it and then try to get directions and type in the first letter of my address my home address pops up in the auto complete. This is a big improvement, but of course auto setting the home address from your google account and remembering the last five in the auto complete would make a lot more users happy as that would work by default v.s. forcing the user to do something, most of which will not. The number of people that will set their home address is probably only a bit larger than the number that were using my grease monkey script to add a home button.
I get your point - if you live in North London, then getting directions from "London" is probably not the way to go. Postcodes seem to be the quickest way, for me.
http://www.theverge.com/2012/6/11/3078987/apple-tomtom-opens...
It's amazing that Apple has nearly matched Google in some parts (maps, directions) and far exceeded them in others (3D).
The Apple worship is truly mind-boggling.
OpenStreetMap is where the crucial data comes from. And if I'm not mistaken it's submitted by volunteers.
And at no point did I say that OSM wasn't responsible for the map data. Only that Apple added other pieces from other companies.
Call it like you see it. Minds may differ.
As for quality of maps it is about 80% as good as Google and has been getting noticeably better over time.
Actually Google and Apple have pretty similar systems for 3d maps. Apple are in beta with their new version for ios6, which will have slightly better resolution than (the current, been there for a while) Google Earth, but not hugely so. See for comparison http://www.iphoneincanada.ca/iphone-news/3d-maps-comparison-...
Nokia are producing similar 3d maps, although again slightly lower resolution than Google.
I think more importantly than resolution will be coverage. 3d maps of the centre of major cities are one thing, but the whole world is a very different challenge, and one we have no clear idea of a winner on.
A better comparison for their current license would be the LGPL - as long as you keep your proprietary data separate from the OSM data, you don't need to release it but you still need to release any changes to the OSM data itself.
[0] http://blog.osmfoundation.org/2012/09/06/your-first-odbl-pla...
The terms of the license agreement with Google for maps data likely didn't include the ability to perform turn-by-turn navigation. Another reason why they moved to their own maps backend for iOS 6.
Really now? As another anecdotal point, I freaking love it, and I'm saying this as a an almost-Apple-fanboy. I'm a pedestrian living in an Eastern European city of 2-million people, and I've used the maps app countless times.
Not that I wouldn't know my city (I've actually implemented 2 or 3 or map-related projects focused on the city I live in, just as a hobby), but because it is impossible to remember how exactly to locate the street bearing the name of an obscure 19-century music composer that just happens to be the address of a restaurant you're interested in. Or how when I got lost in the center of Athens, as a tourist, around Omonia Square ("I don't need a paper map, I can manage" I was saying to myself as I was leaving my hotel room), I heavily relied on the said map and on the kindness on strangers who had shared their home wifi networks to make my way around to Exarcheia.
I questioned why it was neutered the way it is now - not why it was created. I buy Apple products for wholesome user experience and the Google maps ain't one of them (neither is the mail app and so on). If you think this is a silly sentiment, then you must be pretty dumb about user experience issues in mobile situations.
Driving is not the only mode of transport. And SatNav is not the only type of useful map.
Paper maps are useful - an equivalent to a paper map that holds data for the entire world, can put a pin in my exact global location, and can tell me where the nearest Starbucks is? Yeah, I can't see any use for that at all.
Yes, NASA, Digital Globe, Et Alia, have built geographic information services (GIS), there are constraints, both on the data and on what you can do with the data. Google bought Keyhole [1] (the satellite imaging company) discovered that even if you had a satellite image you don't necessarily get to do what ever you want with it, and went out and funded their own satellite [2]. And of course you drive around cars and take pictures of everything, that seemed a little bit crazy at the time but folks like Trimble wouldn't let you use data about streets for doing directions, and then of course different data sets were all slightly 'off' from one another which is really illustrated in the article.
So various folks have invested in one part of the problem, but few have invested in solving the bigger problem, and assembling that solution for disjoint offerings is problematic and expensive at best, and impossible at worst.
That is what I considered the mind boggling investment. And of course the recent tit for tat maps announcements by Apple and Google just reminded me that both companies are pouring in a lot of resources. And yes, Microsoft is right up there as well.
Its quite the barrier to entry if you want to a map based startup. (or at least a huge dependency on an external parties continued pricing model)
[1] http://googlepress.blogspot.com/2004/10/google-acquires-keyh...
I remember when that happened, and was pretty sure that a bidding war would start involving Garmin and Google, first for Navteq, and then for Tele Atlas, when they both were acquired (Google Maps used data from them, and it seems like they still do for some countries according to their Terms of Service: [1] and [2] as as examples).
[1] Canada: http://www.google.com/intl/en_ca/help/terms_maps.html [2] Argentina: http://www.google.com/intl/es_ar/help/terms_maps.html
It will be interesting to see if Microsoft's amazing augmented reality research will end up in a mapping product some day soon as well.
So Nokia (symbian/windows phone) still holds the edge when it comes to navigation. When it comes to local search/places, the Nokia offering is nowhere near as good as the (online) Google Maps
What i'd LOVE to see is for Nissan to throw away its custom OS, and replace it with Android... along with that I want google maps (instead of whatever Nissan is licensing). Plus it would be nice to download apps for my car such as a better mileage tracker etc.
I am truly astounded by how quickly we are accelerating through this information era.
Google is merely processing massive data sets using manually crafted algorithims. This is a complete different project, which has little or nothing to do with artificial intelligence.
That is what Watson is. There's not really any "true" artificial intelligence since the term is so poorly defined, but you could make a good argument that your description above fits humans pretty well, too.
High resolution 3D modelling is already available in much of Norther Europe. It is very cool, but surprisingly difficult to actually make useful. In most applications comic book style generalisations actually work better.
'Not to detour too much, but what you see above is just the beginning of how Google is going to use Street View imagery. Think of them as the early web crawlers (remember those?) going out in the world, looking for the words on pages. That's what Street View is doing. One of its first uses is finding street signs (and addresses) so that Google's maps can better understand the logic of human transportation systems. But as computer vision and OCR improve, any word that is visible from a road will become a part of Google's index of the physical world. Later in the day, Google Maps VP Brian McClendon put it like this: "We can actually organize the world's physical written information if we can OCR it and place it," McClendon said. "We use that to create our maps right now by extracting street names and addresses, but there is a lot more there." More like what? "We already have what we call 'view codes' for 6 million businesses and 20 million addresses, where we know exactly what we're looking at," McClendon continued. "We're able to use logo matching and find out where are the Kentucky Fried Chicken signs... We're able to identify and make a semantic understanding of all the pixels we've acquired. That's fundamental to what we do."'
That is scary. They could read license plates and tie them to addresses. They could read bumper stickers and determine your politics, whether your kid was an honor student at what school, etc.
If you really wanted to keep your politics secret, you wouldn't put a political sticker on your car.
All anyone sees on my car is what radio station I listen to.
now, that's what i call competitive advantage.
Wait, what? By what measure?