OMB

This week, we’re going to be working on OSM, all week long. Check out the ZXV blog for more details.

It was great to spend the day in Southampton yesterday at the Ordnance Survey, to preview the new OpenSpace API:

The API has been a long time in the making, having first been announced at the OS Mashup Event last year. The API is based on the OpenLayers Javascript library, an interesting decision in itself and one which should see bug fixes and improvements fed back into OpenLayers.

Essentially, OpenSpace is just another slippy map API, with most of the features everyone’s come to expect and a few features that make it stand out. It uses the OSGB (OSTN02) datum projected onto the British National Grid (transverse mercator) rather than using WGS84 Latitudes and Longitudes and a mercator projection, as most other web mapping APIs use. To ensure compatibility with the growing volumes of geodata based on WGS84, OpenSpace includes Javascript libraries to transform from WGS84 to projected OSGB - which it seems to do pretty accurately.

Another notable point is the cartography. The Ordnance Survery have a far richer dataset at their disposal, so its not surprising that they can produce map tiles like this:

Whilst the “Street View” (the name for the highest zoom levels) cartography has been criticised by some specialists, its undoubtedly an improvement over Google’s. Its going to be interesting to see if the new map content spurs new types of mash-ups - people can now pinpoint the exact position of their houses and countryside users now have, well, something compared to the nothingness of the Google countryside. I spent at least an hour yesterday just browsing around the map - just having a nice, fullscreen slippy map of the UK is really cool. Take a look here

The license is a license, so some people are going to love it, some will hate it and most will just get on with hacking. One thing that the guys from the OS were emphasising yesterday, was that they really want people to consult with them. That’s why we were there yesterday - so that the people who made OpenSpace could see what we thought. I think the desire of the OpenSpace team to listen to people’s feedback and act on it is a genuine one, so maybe we could all try some constructive criticism before trashing it. But hey, this is the internet.

The most interesting thing about OpenSpace is that the OS own the data. Google, Yahoo, MicrosoftMultimap etc all license their data from other providers, which is why they do things like this. Google don’t want you to hate them, they want you to love them and use their services but are restricted from doing so by licensing regimes with data suppliers. The OS shouldn’t have this problem - its their data so they shouldn’t suffer from the problems that Google et al blame on their providers.

During my first play with the OpenSpace API, I stuck a pin on a map to invite you all to the ZXV christmas party, and grabbed an NPE WMS feed from GetMapping, so you can overlay 50 year old mapping onto the latest OpenSpace tiles. Take a look here.

(continued from the last post

The fact that organisations like the AGI are trying to get to grips with neo-geography is a good thing. Their members have amassed a serious amount of experience working with geoghraphic information and neo-geographers can learn a lot. But in order for this to be a successful two-way exchange, both parties need to understand what neo-geography is. A good starting point would be Andrew Turner’s O’Reilly Shortcut: Introduction to Neogeography. I’m going to try and address a few of the misconceptions and confusions surrounding neo-geography that I noticed at the AGI 2007

Neo-geography is not the same as open source

There’s a lot of cross over between the worlds of neo-geography and the worlds of Free and open source software (FOSS) and neo-geography, but they are not the same thing. FOSS applications exist that fill every space of the geo-stack - from heavyweight databases to desktop GIS to rendering engines. I talked about FOSS for geo-informatics at UCL earlier this year and have a brief list of applications here. FOSS exists to provide free alternatives to proprietary software and to create new software that might not exist if a proprietary form. FOSS projects usually have a focus on democracy and equal access and are based on the fundamental principal that everyone should have access to the source code.

There is no such belief within neo-geography. Projects like OpenSteetMap, which license all of their data under an open license are few and far between. Look through the listings of this year’s Where2.0 conference and you won’t find many open source projects - most companies do not want to give their data away. What you will find is a lot of innovation, ranging from visualisation to 3D modeling. The flagships of neogeography are neither open source nor open content. Applications like ShapeWiki, which lets users derive polygons by tracing over Google Maps do provide their content under open licenses, but are built upon a proprietary platform - Google Maps.

Further more, sites which allow a user to place a pin on a map, or draw a line or polygon on a map, may or may not be infringing the copyright of data-providers. Case law surrounding these issues is far from convincing, but the lack of clarity is enough to question exactly how open the content of sites likes these can possibly be.

Neo-geography is about pragmatic solutions to real problems

Neo-geography represents the democratisation of access to geographic tools and geodata. Before the release of Google Maps in 2005, the options for the amatuer geographer were limited by the barriers to entry to the use of geographic software and in many countries, geographic data. An Arc GIS or MapInfo license is simply too expensive for most people. FOSS alternatives like GRASS exist, but are not easy to use. The proliferation of geo-apps that has characterised the rise of neo-geography has been facilitated by the fast code-fix-deploy cycle that applications like the Google Maps API, and programming languages like Javascript, Ruby and Python provide, combined with the network effect of developing web-based (as opposed to desktop or server) app.

Neo-geography’s reaction against the built-in complexity of traditional GIS has caused a penetration of sectors of society that would be unimaginable using traditional GIS tools. The often voiced criticism that neo-geography is too simplistic represents the divide between the providers of traditional GIS and the consumers of neo-geography. People like neo-geographic apps because they are simple, because they solve a few targeted problems at a time.

Building complexity to ensure exclusivity

There’s also a darker side to the complexity of traditional GIS. The fact that someone needs a masters degree in GIS, to work as a GIS Technician should set alarm bells ringing. By maintaining the complexity of GIS, vendors like ESRI or Oracle are able to justify the costs of their products and consultants are able to justify their high fees and trade organisations justify their existance. Meanwhile, many individuals and businesses that could be using geographic information to make better decisions and operate with increased efficiency are strongly discouraged from doing so. Geographic information will not save the world unless the world’s citizens have access to both the information and tools with which to use it.

The Association for Geographic Information (AGI) are a UK-centric trade association for the producers and consumers of geographic information and related hardware suppliers, software developers and consultants. They exist to promote the interests of their members, provide them with a forum for discussion, oversee professional development and organise events like this week’s annual conference, held in Stratford-upon-Avon, UK.

I was invited to talk about OSM and some of the neo-geographic development that ZXV undertake. The AGI conference is a lot like Where2.0 for people who use proprietary desktop GIS and all the trappings of the pre-Web2.0 era, something that the conference tried to address this year by the inclusion of several talks relating to neo-geography, and a conference theme: ‘Building a geo-community’.

It was a tough crowd. These are the guys for whom tagging, folksonomies and the democratisation of, well, almost anything is a scary idea. In a lot of ways talking to groups like this is what its all about - these are the people who use geographic information all day, everyday. Convincing an open software advocate of the benefits of open geodata is one thing, but selling the idea that a bunch of normal people without years of training and professional certification, can make a map that’s actually quite good, to a room full of people who have built their careers upon concept that they are the specialists, is quite a challenge. Its a challenge that I relish and I think the message is really getting across - there was a lot of interest in OpenStreetMap, open-geodata and neo-geography.

Neo-geography - it will never work

The criticisms leveled and OSM and neogeography were nothing new. The usual concerns about ‘how can a trust the data’ and ‘will it ever be complete’, were voiced, along with some misinformed comments from several of speakers and panelists, including David Maguire, VP ESRI UK and one gentleman who’s name I missed, sitting on the far right of the pannel. (I’d like to crowd source the name of the guy sitting on the far right of the panel during the debate on Thursday, so please leave his name as a comment).

Both of these delegates spun the line that the kind of pins-in-maps analysis that neo-geographers undertake is one thing, but if you want ‘proper’ geographic analysis, you of course need a ‘proper’ GIS. Spatatial analysis cannot, apparently be undertaken with open source tools and is certainly outside of the realms of neo-geography. Have these guys not heard of GeoCommons?. ‘They must have heard of PostGIS or GRASS?’, I thought. Of course they have - these guys are board level executives of companies like Oracle and ESRI. Right now, it would be insane for the VP of ESRI to stand up at the AGI and tell the delegates that yes, its true, 90% of the users of ESRI, could actually use QGIS quite happily. Its tantamount to Vanessa Lawrence standing up and telling everyone to use OpenStreetMap maps. These guys are going to keep on peddling their company line, but that doesn’t necessarily mean they don’t understand what’s going on.

Oracle will profess that open source databases are no threat, but they they’ll release a free version. The OS will innovate to keep up with developments driven by the open source and neo-geographic community. And I would be surprised if ESRI don’t release a free desktop GIS that has equivalent functionality to QGIS. The smart guys are already doing so.

This is far from all I’ve got to say about the conceptions of the established GIS community regarding neo-geograhpy.

OpenStreetMap is coming to Leeds. See this page for details.

Come and help map one of Britain’s most interesting cities. As they say in Leeds, come for the mapping, stay for the Naan

There’s only a few weeks to go until FOSS4G 2007 and OpenStreetMap’s Victoria (Canada) mapping party. Some important dates are:

OSM Mapping in Victoria - ‘Map as a party’

If you haven’t been to an OSM mapping party before, this is you opportunity to find out what open mapping is all about. Mapping parties began with the 2006 Isle of Wight workshop, when 30 OSM volunteers from across Europe descended on the Isle of Wight with the ambitious aim of mapping all of the island’s roads and footpaths in one weekend. 48 hours later, we had a pretty good free map of the island, and thanks to the dedication of local OSM volunteers, the map was soon completed.

The Isle of Wight party set the standard for OSM mapping parties and whilst there are no hard and fast rules about how to have a mapping party, we nearly always follow a similar plan.

Day 1 AM - Meet up in a local coffee shop/community centre (hopefully with WiFi), meet the mappers, decide which areas we will each be mapping, head out to map.

Day 1 Lunch - Meet up at a local pub/cafe, get some lunch, take a look at the mornings traces.

Day 1 PM - Head back out for an afternoon’s mapping.

Day 1 - Evening - Meet up at a local pub, before heading out to a restaurant.

Day 2 - Repeat Day 1

The emphasis really is on inclusiveness and having fun. Anyone who is interested in mapping, GIS, OpenSource, GPS and so on is welcome to come along to the party. OSM have a load of GPS units that we can lend out to people for the day and we will give full training - so even if you have never used a GPS before, you’ll be mapping in no time and once you start you’ll find it hard to stop. There are no rigid rules - you don’t have to map the way I think you should (in-fact one of the great things about mapping parties is hearing about other people’s mapping techniques) and you don’t have to map a particular neighborhood - its your free time that you are giving to OSM, so its up to you what you do with it.

OSM at FOSS4G

There’ll be a few OSMrs at FOSS4G - Mikel, Andrew and Corey will all be in attendance. I’m planning on doing some informal mapping on Monday 24th, so if you are interested in joining in, get in touch. Friday’s Code Sprint could also provide a good opportunity for some OSM hacking - it would be especially cool to talk to some other FOSS4G developers about integrating OSM data and software with with other FOSS tools - why include a shapefile of OSM data in the QGIS binaries for example?. There’s also the BOF sessions that could provide great opportunities.

Meet the mappers

So there’s a few ideas - if you want to meet up to talk about OSM, free data or anything else during the week in Victoria, drop me an email. If you wanted to attend but can’t make it, keep tuned to OMB and OpenGeoData - where I’ll be blogging more about the mapping party and the conference.

On the 1st and 2nd September, the Isle of Man is going to be hosting its first OpenStreetMap mapping party, organised by Dan Karran, who’s been blogging quite a bit about the event. Dan’s latest post highlights the impact that OpenStreetMap maps of the Isle of Man can have. Right now, the best map of the Isle Of Man that you can find on the internet is OpenStreetMap’s. Its not complete yet, but it could be with a couple of days of effort.

You can find more details about the weekend on the OpenStreetMap wiki. If you are interested in joining in the mapping, get in touch with Dan, or me.

Yep, its OSM tiles on Multimap. More info on Mcknut’s blog.

If bright flashing lights are more your thing, take a look at this animation of OSM tile requests:

Songbird is an open source project to develop a cross platform media player to rival proprietary offerings like Apple’s iTunes or… oh well I guess iTunes is the only proprietary media player worth a mention really. The good news is the is now a “developer’s” binary of Songbird 0.2.5 available from here, so you can now have the joy of Songbird without the misery of worrying about C compiler flags. I’ve tested it on Ubuntu and Mac OS X (Intel):

One of the cool things about songbird is its seamless integration with network resources - once point it at the URL of a site that has MP3s (or almost any other audio format imaginable) on it, you can easily subscribe to podcasts and feeds and effortlessly drag and drop remote media into your library. A short but comprehinsive screencast will tell you everything you need to get going. If you listen to music on your computer, check out songbird.

I’m going to jump on the band waggon and say thanks to OSM’s Cartography Crew, who slogged it out before the Oxford meet up a couple of weeks ago and have re-vamped OSM’s cartography. The map tile show belown is produced using the Mapnik rendering engine - an open source (LGPL) GIS library that makes boring databases fun colourful. The downside of Mapnik is the 873 lines (and growing) of XML needed to produce the map. Hacking this level of config isn’t that much fun, so thanks again to those who took the time to work on it.

Central London, as seen in the OSM Slippy Map

If cartography like this isn’t enough to get you going, just think how far we’ve come. When I first found OSM a year ago, the map looked like this:

OSM - One year ago

So now we have Open geo-data that in certain areas rivals the completeness and even the existence of proprietary data. With this data, we can use a Free software library to make maps that look better than most of the other maps you see on the internet. Maybe there’s something in the whole ‘open’ thing then after all?

You want contours? We do contours too:

Produced by Robert Hart using Osmarender