Perhaps web mapping has power not only to democratise cartography, but also to cartographise democracy. Spatial search helps people find things and services relevant and near to them; it also helps them find the individuals and organisations responsible through election for the maintenance of those services.
Rich Gibson, Schuyler Erle and I were all impelled to write a book of "Mapping Hacks" for interconnected but different reasons. Rich has a warm and folkish line that "maps tell stories", that the tools we hope to help people create allow a construction of a cartographic narrative of life, a kind of full-motion slideshow. He imagines his great-grandfather, journeying West on the Oregon Trail, with a GPS, a digital sound recorder; and wants to reconstruct that narrative; using maps to take the internal, and externalise it.
Schuyler talks of the "the democratization of cartography" with our tools. His urge to map is borne from social activism; a sense of the tremendous use value in collaborative mapping, shared civic planning and user controlled information. Putting mapmaking tools on the web, building interfaces to free and open source desktop tools for manipulating geographic information; aimed at delivering "Geographic Information Science" from the hands of a high priesthood still dependent o expensive and proprietary software stances.
I think, "Who controls the description of the world, controls the things in that description." Maps, in part and like the internet itself, are demilitarised techology. The experiments in location-aware art works and social tools of which my work has been a part were driven by the removal of selective availability from the Global Positioning (GPS) system in 2001. The geodata collected by state agencies has a military provenance. Indian government release of free satellite imagery, if that occurs, will be a technocratic show of strength.
At present people working to build online local information and spatial modelling services, are dependent on United States government published sources mostly for global data - the VMap0 and partially declassified VMap1 vector data sets; Landsat imagery and SRTM height models; the GeoNET gazetteer for global placenames... Though deeper, more accurate local sources of this information are available, they are kept by institutions, rented with no right to redistribution or modification from an agency such as the UK's Ordnance Survey. Philip of Spain kept his maps of Newfoundland under lock and key, state secrets, while the English maps were freely published: and those acquired the power to name things for the future. Collaborative mapping efforts offer a naming by consensus, a meaningfully local set of descriptions.
Over the last few years the climate around digital mapping and web mapping has changed. Amateur, home users can for the first time afford the many-gigabyte storage and capacious broadband internet connection, to handle very large bodies of spatial data - no more truck full of DAT tapes. This has helped emphasise a 'boom' in Open Source GIS, as hobbyist projects like Quantum GIS pick up more users and contributors, extend interfaces towards other software projects and find they have a place in an ecology.
Meanwhile, a generation of media arts and research projects are exploring the potential of combining a location sensing system like GPS with a two-way flow of bandwidth, and picking up on keywords like "mobile", "pervasive", "psychogeographic". Some projects are looking at cartography more critically, exploring its moral and metaphorical dimensions; others are finding new needs and uses for spatial information.
Projects like geourl.org and gpster.net start connecting things on the web to things in the world, coming up with their own annotation formats to tag things in space. Ad-hoc standards appear, and creators of experimental projects, of free software and open networks start to need infrastructure for making maps and creating and sharing data.
'Google Maps' appeared around the same time we were finishing pre-production for "Mapping Hacks". Having seen how key the local, mobile search market would be to the mega search engines, achieving branding, description even at tremedous data licensing expense in the web map space must seem of great importance.
Google's map site, in the UK at http://maps.google.co.uk/, is more of a map application service, and this is what has produced the 'buzz' about it. Rather than simply providing maps, Google provides a simple, web-programmable API, or application programmer's interface, that allows anyone to "remix" different information sources using a Google Maps base map subject to Google's terms and conditions. http://googlemapsmania.blogspot.com/ is a web log that collects examples of different 'google maps mashups' that have been made by people on the internet using Google's API and their own data sources - there are hundreds of examples, mostly in the US.
Why isn't Google Maps 'enough' for the kind of web mapping applications we want to make; enough of a democratised tool that any web page maker with basic knowledge of JavaScript should be able to customise? Google is taking on archiving and digitisation projects for major libraries and print publishers; why not outsource national mapping data collection and distribution to them?
Where there is free data available as part of Google's map, there's no way to separate the free from the non-free. Aerial imagery and vector data about streets and buildings are often derived from originally-free, government-published sources, but they are enhanced by commercial spatial databases which can't be re-used.
The maps are drawn at a great distance from the people who live in them. They're not yet available for most places outside the US, the UK and Japan. In areas of 'market failure' such as central Africa, they will never be made available while a technical-ecological context for them exists. While drawn at a distance, and with local variants in the 'ontology' of the map so great, features will be misdrawn, mislabelled, missing. Google's UK maps are just not very good -ugly projections, smaller features missing, labelling smaller places in preference to more 'recognised ones' on the ground, a very incomplete model of green space.
Anyone building a web service which they hope to maintain to a high level, that is dependent on Google Maps has to bear in mind what happens when Google needs to recoup its licensing costs by charging for API access or embedding adverts into the map imagery; or in the worst case, goes away entirely?
We can make our own maps, and we can definitely build web mapping services with a Google-esque panache to them but that have much extra depth and rewritability. Creating proofs-of-concept is a lot easier in the United States, where geospatial data collected by the federal agencies (NASA, the Census Bureau) is available for free and in the public domain. At the state level, many (though not all) local authorities make a lot of civic information available - it's expected. "Available" here implies machine-readable, in an open, structured, standard and documented form; for example, Minneapolis provides structured XML feeds of the latest sensor data from its traffic information system.
The US Census Bureau's TIGER/Line street database is a tremendously useful body of open geodata. It can be used to construct a fairly accurate street map with a vast amount of metadata about what things are, how they are labelled and how they fit together. It can also be used to build a 'geocoder', a service which approximates a latitude-longitude location from a street address, within five or ten metres, the typical error range of GPS.
http://geocoder.us is our free US geocoding service written mostly by Schuyler Erle. Though it misses newer streets and fails in some cases where street and building names are ambiguous, it's useful enough to build free web applications with, mapping public domain data such as health inspection reports, and underpins many Google Map hacks. Google themselves don't offer a free geocoding service;their data licensors won't allow it because it is a profit making service for them.Hobbyists, academic and small business organisations use geocoder.us on a nonprofit basis, as do the State governments of Rhode Island and Maine. Its source code and data set are both open.
In the UK the interest in mapping and spatial annotation is supposedly strong, and for a long time we have been complaining about lack of access to base map data 'owned' by the state; so why haven't we really seen any really good UK google maps mashups? Partly, we don't have access to a free geocoder - given a list of addresses on some common theme - a very common form of data and application for it - we still don't have a free and redistributable means of encoding those addresses in space; that information, and implicitly the ability to do that transformation, is copyrighted by semicommercial state agencies.
Partly, we don't have open access to reusable, machine-readable public information in the same way, the UK does not have a tradition of it; even our freedom of information laws reinforce the idea that information is not available by default. Information is too sadly often published on local council websites as PDF or even word document 'views' of data and of map information. There is perceived 'value' in geodata, and the short-sighted pressure to commercialise it trickles down from the national mapping agency. Planning archives held by the National Land Information Service cost several pounds per property to access. All the very spatially meaningful information, that might offer a spur to shared planning software, to more localised control over resources, that is not available to us unless for sale, and then with no right to republish and redistribute it.
In New York, Andrew Rasiej's campaign for 'Public Advocate' involved a platform of universal local wireless net access, and completely open access by citizens to government information. His campaign was notable for including a Google Maps mashup as part of its activity, using the photo-sharing web service 'flickr' to help people annotate potholes and cracks in roads along common cycling routes.Rasiej asked: "Why can't we collectively see the data coming into the 311 [city local information] system in real time?". He polled 5% of the vote.
The New Orleans hurricane and aftermath created a surge in voyeuristic interest in Google Maps, and a few positive applications of it; sites using feeds of data "aggregated" from different relief-effort websites to plot where housing and support is offered in different locations. In using the web to haul different feeds of data about and layer them on top of one another, web mapping hackers are doing something that comes naturally in the domain of digital cartography.
In some areas access to mapping data is restricted, either because it is commercialised by the state, or because it is not collected at all, in areas of state and market failure. Community mapping projects have arisen, using patterns derived from GPS track logs to model streets and build software for drawing and annotating them. Pioneering projects were bicycle mapping projects, including http ://bbbike.sourceforge.net , one Berliner's labour of love, and http://bikemap.onearth.com.au , a project which spawned the now popular 'Community Mapbuilder' software. OpenStreetmap, a London-based project which has nodes throughout northern Europe, provides a service which collects GPS traces in XML ad allows people to draw a collaborative map, wiki-style, with a java-based web interface.
The Mumbai Free Map project - http://freemap.in/ - took a different track in an area of more relaxed intellectual property, hand-vectorising local authority development plans for a city of 18 million people, to building level, which are now available via a web map service as they work on web-based annotation interfaces with an aim of producing PDF and print maps for use by local housing rights activists. It's possible to do this on a shoestring because of access to free and open source software to process and manipulate digital map data. The PostGIS spatial database, mapserver and geoserver for web publishig of data, the amazing gdal/ogr tools for dealing with many kinds of proprietary data formats.
Open Source geospatial software offers a powerful toolkit, but without data, there is no stuff to work with our tools. The Open Geodata manifesto offers a description of ideal data licensing conditions that state-run and grassroots projects could share - http://okfn.org/geo/manifesto.php . Open access to geodata is just one 'slice' of access to public information in general, though it provides a context, a framework, to analyse other kinds of information (such as MPs travel expenses vs. distance from home address to Parliament).
With access to more civic information sources, we will combine them, and combine the combinations, to produce more interesting and powerful tools than could ever really have been "planned". TheyWorkForYou.com, the Hansard-republication service, is part of a whole network of analysis services about MPs behaviours and interests, all of which are made more meaningful when looked at via a postcode search from TheyWorkForYou. A myriad applications, simple yet powerful when combined, would arise just from an easy-to-use, free web-based geocoder for the UK. If you support this ideal, please consider signing the manifesto to show it!