I’ve arrived in Amsterdam and already I’ve bumped into a lot of lovely people – Tom Coates, Paul Hammond, Simon Willison (who were all on the train from the airport, and all Yahoo!s: proof), Matt Patterson, Edd Dumbill, Mark Nottingham, Uche Ogbuji, Bijan Parsia. While I typed this, Libby Miller and Damian “I’m in the Guardian” Steer just walked past to register.
and now back to writing the slides for my talk later this week: Semantics Through the Tag on Friday.
Ooh, we are suddenly distracted by the new Apple x86 Mac Books being launched.
Technorati tag: xtech for Planet XTech
XTech 2006: Open Street Map (OSM), Steve Coast
Presentation at XTech 2006 in Amsterdam
The Open Street Map/ (OSM) approach is to have both Free Data and Free Software and to add to the data in the wiki style.
There are various problems with using UK Ordnance Survey (OS) data including IPR (even with academic licenses), leaving little choice but to use out of copyright maps, the last open ones are from 1944, pre-WWII and before motorways were constructed.
Steve showed an animation of tracks recorded by couriers in London, used to build up vectors of the roads in london. The tracks became rather thick when aggregated due to inaccuracy of GPS but this sodes build up lots of data.
The OSM site provides RESTian APis to the data – both read and write of data in XML form, along with monthly dumps planet.osm. He shows examples of the data in the UK that has been collected over the last 2 years, although not all of the data is yet shown in the OSM frontpage.
There are several third-party applications that are making use of OSM data including GpsDrive free software route tracker, the FlightGear flight simulator (who built a free flight simulator but forgot about terrain) and it can even be used on some GPS units as source of data for routing, the example given was a Garmin device. The XML data itself has been increasingly turned into prettier output such as using XSLT into SVG, giving much more accessible maps.
The current OSM focus is the UK but there is data collected elsewhere in Europe. The UK status is that all UK motorways are now mapped, several major cities and even a forest – the New Forest, so it is not just for the drivers in the world, but for walkers over paths and public rights of ways, bridleways. (I thought maybe the UK Rambler’s Association should be brought in here).
In comparison to the US, it has a glut of data due to the US copyright law that the governement cannot own copyright on map data they create – the Tiger/line data. This can be loaded into OSM and annotated just like for any other country.
Recently, the OSM project ran a mapping for the UK Isle of Wight island in the south of England. The last usable map that could be used as a basis was the copyright-free map from 1944 which gave a rought outline of the land and areas but is still rather crude. 30-40 volunteers walked, drove and cycled the island and covered 90%+ of the roads in two days taking pictures/recording the roadnames for traces and turning it into annoated traces and then vectors, loaded into the OSM itself. Not all of this is yet online. Two weeks ago there was another mapping even in Manchester – a built up area which causes technical problems – but feedback from the first event caused more volunteers to turn up than planned who needed training as they expected to be given GPS units when they arrived. There may also be a need for more expensive gps units to handle city mapping.
As an aside, Steve pointed to the Isle of Man which is a blank island for roads in Google Maps as it is a separate country with different licensing and copyright laws, so has no licensed road maps. OSM does have maps for it since there are contributors who live there.
There has been fallout from the two events after publicity in an article in the UK Guardian Newspaper and on blogs such as Boing Boing. The Guardian article ementioned a trap street in Bristol: Lye Close which was made up to see when maps were stolen. Or in Steve’s words: “they give you crappy data and make you pay for it”. Steve suggested that cartographers are the very purist – to make the perfect map and prevent you scribbling on it. OSM doesn’t have to do this as it has no profit/business model
An OS spokesman was quoted on their work and explained that maps are expensive to update at the level of detail they work at, however Steve said a lot of people don’t want that data, just something at a rough scale, a personal scale, not down to the millimeter.
Finally he mentioned UK Postcodes (aka Zip Codes in the US) which are completely Copyrighted by the UK Post Office and you can’t touch them. Annoying. So OSm have been collecting GPS points + post codes one-by-one and although this isn’t going to be complete as there are 19M+ UK postcodes, it will give the 100m accuracy level most people carea about: the first part + the number of the second part of the post code.
Technorati tag: xtech