Unlike other maps on the Web, OpenStreetMap lets you update and use the underlying free data the way you want. Project founder Steve Coast believes what happened with Encyclopaedias will repeat with maps
Founded in 2004, the OpenStreetMap Project has managed to draw a map of the world, ready to compete with traditional and commercial geo data vendors. The map is drawn by volunteers, mostly by using GPS or tracing aerial imagery provided by Yahoo.
You can put the changes that matter to on the map
Countries like Great Britain, the Netherlands or German are fairly complete, and while rural areas sometime need improvement, the data quality in urban regions often surpasses the quality of commercial data. While large commercial data providers employ some 800 people to collect data, OpenStreetMap boasts a rapidly growing community of 135.000 mappers. Naturally, the project is often described as "Wikipedia for map data".
Two things stand out: You can put the changes that matter to on the map - instantly. Think of corrections and additions to the map of your local neighbourhood. This also means that places no one else would map in detail appear on the map, be it Cuba, Baghdad or Palestine.
And: You can use the data, provided by a series of simple XML based interfaces in your own projects, commercial or private.
During the OpenStreetMap State of Map conference in Amsterdam, Alexander Holy had an opinion to talk to OpenStreetMap founder Steve Coast:
derStandard.at: Let's take a look back at the beginning of OpenStreetMap. I remember first hearing about OpenStreetMap in 2005 at an Innovators conference in the Netherlands. From your talk it sounded like an insanely great idea, but a quick look at the website revealed an almost blank map. So it looked like quite a leap of faith back then. Did you intend just to map Great Britain, or did you intend to map the World right from the start?
Steve Coast: The World! Yes, we did not want to restrict us to mapping London, or just mapping Britain or Europe or anything like that, it's much more valuable to take a bet and go for the bigger goal of mapping the whole world, because once you done the software and built the community for one place, it works the same everywhere. So there is no point in restricting yourself to just mapping London or something like that.
derStandard.at: Why did you care for free geo data in the first place?
Steve Coast: I used to use Linux a lot on my laptop, and now I use a Mac. But back I used a GPS, a USB GPS for my laptop. There were a few mapping applications available, but what there was had no access to vector data - and with no vector data there was very little you could do with the software. Access to vector data was very restrictive, so I figured why not make my own vector data - it's pretty much that simple.
derStandard.at: Did it start with the business idea or just the personal need for data?
Steve Coast: This is where all the great ideas come from, it is just scratching your own personal itch. Just like Larry Wall did (the creator of Perl). The basic idea is if you have a problem find a solution. I had a problem, so I found a solution and shared it - and here we are.
derStandard.at: Is OpenStreetMap something very much like or unlike Wikipedia?
Steve Coast: I would compare it to Wikipedia with the exception of a few subtle changes; one of them is that we are building a data set rather than book-like documents, a reference work if you like. Also we are building something which in many ways avoids the problems of Wikipedia, in terms of having a neutral point of view, because we are mapping things that are factual. There is a road there, or there isn't. You cannot have opinions about that. So it is subtle different, but then again also very like Wikipedia.
derStandard.at: In fact you end up with geo data in XML, a very technology driven thing. The question was raised at the conference if the project should invest more in easily accessible frontends to the map?
Steve Coast: I certainly think that the website (openstreetmap.org) should be a showcase of the best of our mapping, as well as our data. There is a really good case to be made that we should provide more good uses of that data. We don't have infinite resource but it is certainly something I personally think we should do.
derStandard.at: OpenStreetMap has some 135,000 mappers today, and there was a lot of talk today about reaching one million. What is next, just completing the map?
Steve Coast: There will be much more information, you are already seeing this in places like Germany, where the map is already complete, where people have mapped all of the roads, all the parks, all the buildings. They are now going to map every tree in the city, every lamp post. So the map data is just getting richer and richer and richer, that's pretty clear to me.
derStandard.at: Will OpenStreetMap put commercial providers of geo information out of business; will they react - just as it happened with encyclopaedias and Wikipedia?
Steve Coast: The will have to react, and I hope we will have a better relationship with them than Encyclopaedia Britannica had with Wikipedia, because it is very easy to just put your head in the sand and deny that all of this is happening. But I think there are good signs that some of the major providers are really happy with having a discussion, and we can work out a better relationship than what happened with Wikipedia.
derStandard.at: Recently, Nokia has acquired Navteq, TomTom has acquired TeleAtlas, just to make sure they have all the needed geo data. Could they rely on OpenStreetMap as an alternative?
Steve Coast: It is very difficult for these companies because they would have to change their entire business model, and they find that a little bit difficult. It's very interesting to see what happens. Our main concern is completing the map, a completed map is really what would cause them to change their attitudes and the way they work. TomTom and Nokia want to supply maps to their own customers, and on top of that they want to be a data supplier, and they want to own the map data itself. It is pretty much a defensive move more than anything else. I don't think they imagine themselves in that business forever. So they have a problem, because the shift to a model with open and free data would mean they would have to write down the cost of the assets they just bought, and that is a very difficult thing for them to do. Like Encyclopaedia Britannica it is easier for them to stick their fingers in the air and pretend it's not happening... It will be very interesting to find out what is happening. E.g., Microsoft has shut down encyclopaedia, but they have done it years after they should have done it. Nokia and TomTom they might as well decide to do good things with their data assets, but don't expect it to happen any time soon. It will not even happen after OpenStreetMap is as good as they are. They will still for many years plow their own paths.
derStandard.at: OpenStreetMap chose a very different, much simpler approach than traditional geo information systems. Intention, or did this just happen?
Steve Coast: That was totally on purpose. Top-down ontologies are just insane for something like OpenStreetMap, we needed to do something like simple key/value pairs, and it worked very well. OpenStreetMap can cope with a continual change of environment. It could today point at (the maps of) Cuba, or Baghdad or Afghanistan, or even - London - and look how amazing these things are. But, the maps do not only get better, this time next year it will be different places. The best thing about the project is that is constantly changing. I try to refrain from going "hey, now we are done", because next year there is some other cool thing has been done. The key thing is that it is always evolving that there is always a better thing coming along.
derStandard.at: You are founder of CloudMade, a venture capital funded company with locations in the US, UK and Ukraine. What is your plan to build a viable business on top of open data?
Steve Coast: To me it is just obvious. It is what has happened with Linux, what has happened with Wikipedia and so on - these things are just better. The value is going to move away from the data set and on to the services on top. I think it is pretty clear from the look at history of other open source projects what is going to happen. (The interview was done by Alexander Holy. He is CIO of Der Standard and standard.at)