The Geospatial Web: A Call to Action
by Mike Liebhold, contributor to O'Reilly's upcoming Mapping Hacks
What We Still Need to Build for an Insanely Cool Open Geospatial Web
Editor's note: In this article, Mike Liebhold writes about what we need to do to tap the as yet unharvested business opportunities in a geospatial web. This is just the type of topic we'll be exploring at the O'Reilly conference Where 2.0, coming to San Francisco in late June. Join us to learn how vendors, application developers, and consumer web companies are using GPS, RFID, WLAN, cellular networks, and networked sensors in new ways to solve old business problems.
Beyond a growing commercial interest in mobile GIS and location services, there's deep geek fascination with web mapping and location hacking. After several years of early experiments by a first generation of geohackers, locative media artists, and psychogeographers, a second, larger wave of hackers are demonstrating some amazing tricks with Google Maps, Flickr, and del.icio.us. Meanwhile, a growing international cadre of open source digital geographers and frontier semantic hackers have been building first-generation working versions of powerful new open source web mapping service tools based on open standards like WMS (web map services) and WFS (web feature services), all built on GML (geographic markup language) and XML. You can see for yourself: hundreds of new geospatial web-related links from perhaps thousands of geeks and users are aggregated daily at del.icio.us/inbox/starhill_blend.
Out of this teeming ecosystem, we can see the beginning shapes of a true geospatial web, inhabited by spatially tagged hypermedia as well as digital map geodata. Google Maps is just one more layer among all the invisible cartographic attributes and user annotations on every centimeter of a place and attached to every physical thing, visible and useful, in context, on low-cost, easy-to-use mobile devices. In a recent email, Nat Torkington, organizer of the upcoming Where 2.0 conference, said it this way: "Everything is somewhere. Whether you're talking about assets, people, phone calls, pets, earthquakes, fire sales, bank robberies, or famous gravestones, they all have a location attached. And everything we touch in our lives, from groceries to digital photos, could have a location. From these locations we could learn a lot more about ourselves and build new economies."
We'll be able to view and utilize tags, pages, and graphic data about places far richer than the current yellow pages and commercial POIs (points of interest): environmental details; live, in-place cultural information, entertainment, and games; history, mythology, and social information about people nearby; geodemographics; micro-local commercial information; safety information based on data about health, accidents, and crime; political data, facilities details, and local public services; physical objects' material composition and origins; links to manufacturers' sites and help desks; instructions on uses; stories about prior use; user annotations; digital graffiti; user-created art; and location-tagged messages.
The first-generation internet and Web generated a huge amount of economic energy, and so will a geospatial web. So it needs the same high level of early, informed financial support from forward-thinking public servants, geospatial service entrepreneurs, and philanthropists.
However, while it is interesting to entertain ideas of early financial returns from geospatial web services, we all need to take a deep breath and perform a sober and unhyped assessment of where we are, and what we still need to do to enjoy the economic and creative benefits of a geospatial web. As Nat observes, "There are problems getting data. The mobile phone (the computer most people carry with them every day) is unnecessarily difficult to develop on. The underlying map data needed for even simple visualizations is not free in most of the world (and of dubious quality where it is free). And software and hardware builders are still thinking about shipping software and not building the infrastructure to power this revolution. With these and other obstacles out of the way, entire industries would be built around this ubiquitous location metadata."
We can't afford a second dot-bust; investments and developments have to be smarter this time. Here's what we need in order to build a sustainable geospatial web:
1. Cheap, reliable mobile broadband IP access over commercial wireless services.
One of the reasons the wired Web works so well is that IP network services are dependably always there, just like a dial tone. On the other hand, mobile IP internet access over commercial carriers is still limited, expensive, and mostly awkward to use on mobile phone devices.
Carriers hoping to reap big rewards in mobile commerce and other services might instead see bored users and subscriber churn by limiting user choices to their narrow "walled garden" service offerings. The situation is reminiscent of the days before the Web, when giant services like Prodigy, CompuServe, and AOL were operating competing, noninteroperable, closed-wall services. These massive efforts failed as the open-standard Web unfolded. Perhaps today's wireless carriers might better profit and recover the costs of broadband deployment by letting a more natural IP services ecosystem grow: including their own mobile commerce services, which they can offer in fair competition with myriad other providers, just like on the Web; and offering tiered or guaranteed levels of broadband quality of service that starts with a default (like a dial tone), low-cost, open IP access to the Net for all wireless users.
2. A cheap, reliable, privacy-observant standard geolocation infrastructure.
Let's break this down into two parts.
First: free, reliable, and privacy-observant standard geolocation services over commercial wireless services.
Most service carriers have or soon will have the theoretical capability to provide geolocation coordinates to user applications. That seems like a reasonable idea; the software on my mobile should be able to ask the Net where the device is, if the network knows. Obvious. Unfortunately, access to user geolocation APIs is carefully guarded and shared only with internal developers and service partners. Where carriers do offer users' location, the tariffs are high, last reported at pennies per query and quickly amounting to dollars for even the casual browsing of local services while in motion. Expensive geocoordinate services prohibit user interest. Why not offer a location-free feature, which will allow thousands of services to bloom and driving demand for carrier network services?
"Tear down the walled gardens!" is easy to say but problematic. The real challenge is developing vivid IP geoweb scenarios—backed up by some credible economics to help carriers understand the economic benefits of low-cost IP Net access—and free, private geolocation coordinates to every user. On the other hand, we could just write flaming rants back and forth to each other, and nothing will change. And if the carriers fail, who will run the network? Homeland.gov?
Second: a DNS for wireless waypoint MAC address locations for ubiquitous Wi-Fi and other wireless signal-strength geolocation.
Developers at Intel's placelab.org and elsewhere are demonstrating that a wireless device or a mobile phone can "know" where it is by simply "listening" to the signal strength of cellular towers and Wi-Fi base stations' MAC addresses, and comparing it with a cached portable database of the known geolocation (latitude and longitude) of these wireless transmitters. It works. Unfortunately, Wi-Fi base stations are ephemeral, on, off, moved ... hence the need for some kind of P2P, constantly updating, DNS-like service providing currently known locations of MAC addresses for 802.x base stations and cell, FM, and TV towers. The best news about these systems for locating users' devices is that users do not have to query the Net for geolocation, which would disclose their location to network operators. Instead, users' privacy of movement is preserved.
3. Better functions and form factors for low-cost, handheld mobile devices.All we want is built-in GPS and other geolocation; a standard Java runtime environment; better, cheaper displays; better keypads and pointers; multiband reception such as cellular, Wi-Fi, and Bluetooth; and obviously, more power, memory, storage, and battery life. The breakthrough adoption of mobile geoweb services clearly depends on a stable, common-client software platform running on a high-powered, low-cost, easy-to-use "tricorder." I know, I know, we have to be patient for maybe two or three more years for great mobile devices. But meanwhile, we still have lots of software to build.
Join us at the first Where 2.0 Conference June 29 and 30, 2005, in beautiful San Francisco. Location-based services and mapping are becoming mainstream technologies. Explore the emerging consumer and enterprise ecosystems around location-aware technologies—ecosystems that increasingly impact the way we work and play. Need more reasons to attend?
4. A vigorous and productive codevelopment community of software tool and system builders and mobile platform developers.
We're off to a great start: there's a huge number of open source geosoftware communities (well documented in the upcoming O'Reilly book Mapping Hacks) working on various pieces of the puzzle: OpenMappers, GPSmappers, MapServers, and GeoServers; RDF mappers and terrain mappers; geobloggers; and GMAP hackers, locatives, and WorldWinds. Pick one and dive in!
5. An active, productive community of standards and policy developers focusing on privacy and technical interoperability.
We're not doing too bad here either; lots of good work has been going on quietly for the last ten years:
- Standard location, presence, and privacy protocols (IETF Geopriv final draft)—rules for noting the locations of users and documents
- Standard databases for Wi-Fi base station locations—in process
- Standard database of Wi-Fi beacons for geolocation—needed
- Standard location API for Java mobile edition (J2ME)—implemented unevenly on many mobile platforms
- Standard carrier API for E911 location information (OPENLS)—implemented unevenly on many mobile networks
- Standard geographic markup language (GML/XML)
- Standard descriptions for maps—allow digital map data to be exchanged easily
- Standard web feature services (OpenGIS WFS)—rules for servers to serve GML-standard data
- Standard web map services (OpenGIS WMS)—rules for servers to serve
standard raster maps of combined layers
... and lots more for GPS, image processing, metadata, and beyond.
6. Location privacy.
Despite ongoing work at the IETF (Internet Engineering Task Force) geopriv committee, an enormous amount still remains to be done to protect users' location privacy. Current plans allow network operators and public agencies to always know a user's location. Fearing a potential negative impact on location-based mobile commerce, industry associations are proactively drafting privacy guidelines while negotiating with regulators over mandatory restrictions. Meanwhile, some carriers are already offering services allowing users to track the location of "friends" and "family." Absent some early regulations, these services appear to be inviting abuse; there's already a serious stalking problem. One well-known location-based social game in New York has been plagued with problems of exes stalking their former partners. Fortunately, research groups such as Bell Labs and Intel Labs are already crafting first-generation privacy application toolkits that would allow users to selectively reveal their location to validated, authentic friends, family members, and authorized colleagues.
7. Open, free access to all of the public geodata in the world.
We take free geographic data for granted in the United States. The U.S. government has a long-standing policy of freely distributing massive amounts of digital cartographic data collected by census, by satellite, and by agricultural, environmental, and space agencies. We have access thousands of types of geodata including road maps, weather maps, and land-use maps. Geoweb developers in other nations are not so fortunate. Public geodata is hoarded, or sold at prohibitive prices. A few highly visible activists such as Jo Walsh, Schuler Erle, and Steve Coast at openstreetmap.org are making the problem visible in the United Kingdom, as are others, including the Opengeodata group, which is trying to get organized. However, much needs to be done globally. Even in America, much useful geodata is still obscured and unavailable. Why shouldn't my daughter, a runner, be able to know which running trail is safe and which one has a history of assaults, for example? Why shouldn't inner-city residents be able to know precisely where the environmental risks are in their own neighborhoods?
8. A comprehensive search engine for existing geodata on the internet.
So far, digital map data is available only from a baroque assortment of directories, gateways, repositories, and libraries. There's just no way to search and retrieve all of the attributes of any place on earth—in its full digital richness of features, maps, and aerial and satellite images.
Organizing the Web's hundreds of geodata directories, gateways, and repositories into a coherent, unified geospatial resource is a global collaborative challenge. Vast amounts of proprietary and domain-specific geodata produced in nonstandard formats, or hosted at web sites with nonstandard search capabilities, remain unsearched. That includes databases belonging to governments, public agencies, and private companies all over the world.
The Web doesn't have a single, comprehensive clearinghouse where you can find all of the data and domains of knowledge covering all geographies. Instead there are hundreds of national, regional, local, academic, private, and individual online gateways. Ideally, a simple search will let a user see all data as well as informally created spatial hypermedia.
So far, this problem remains unsolved. Very few geospatial information scientists are working on the challenge beyond the GCMD (Global Change Master Directory), whose database holds more than 15,000 descriptions of data sets and services covering all aspects of earth and environmental sciences. See http://gcmd.nasa.gov/Data/portals/gcmd/param_search/ HUMAN_DIMENSIONS.html.
A small number of companies are beginning to search for geodata. At one point, a company called Ionicsoft launched the AsktheSpider.com search for OpenGIS-conforming spatial web services such as WMS and WFS. Unfortunately, because OpenGIS standard data formats (GML, WFS, and WMS) are so new, this is a still minuscule resource; thus, the service is off the air.
9. Vigorous, well-funded, productive R&D focused on the prevention of geospatial spam.
Quite a few geobloogers have attempted to put geocoordinates in the head and meta sections of HTML documents. However, this approach hasn't been widely accepted, because pages are not searchable by location. The problem, while still hypothetical, is spatial spam. There's nothing to prohibit spammers from putting multiple locations in the HTML of a single page so that the page shows up in searches for different locations. Based on early, painful experience with meta tag spam, search engines just ignore meta tags altogether. Spatial spam is one problem we can anticipate before it happens. Let's figure out a solution.
10. Intelligent project funding for cool projects: research funds, art grants, and venture capital.
There are enormous unharvested business opportunities in a geospatial web, but patience is necessary because of the fundamental constraints discussed earlier. Although a real geoweb infrastructure is still a few years away, smart developers and investors might make some huge gains by tackling and solving some of the challenges listed here.
Mike Liebhold is a full-time senior researcher at the Institute for the Future, focusing on proactive, context-aware, and ubiquitous computing, and the social implications and technical evolution of a geospatial web.
Return to the O'Reilly Network.