by Cal McElroy, Guest Author
Maps, and their cousins satellite images, are getting a lot of attention these days. In the past past two weeks there have been some interesting articles in Wired, New York Times and The Province, a daily newspaper in British Columbia, Canada (that covered the GeoWeb Conference last week, in Vancouver, BC). These stories all focus on a growing trend in user generated content appearing in online web mapping services like Google Maps, Google Earth, Microsoft Virtual Earth, and Yahoo Maps. Of course, we now take for granted the Google Earth-enabled “fly in” from space – to the location of a disaster or news incident – seen on CNN and Fox News.
With over 50,000 Google Map mashups since it’s launch in January 2005, there is certainly something important going on here. I would like to spend a little time in my last three posts in this series, discussing the implications.
First though, some useful context. Maps are a form of communication, that dates back to the Stone Age, and is believed to even predate written language. One of the oldest surviving maps, created in 6200 B.C., is painted on the wall in a Neolithic settlement, called Catal Huyuk, discovered in Turkey. Maps have always been about depicting the natural and man-made features of the world around us – as spatial orientation, or location is fundamental to the human experience.
Computerized maps appeared on the technology scene in the 1960’s from pioneers including Intergraph and ESRI. It was only in the mid-1980’s, with the advent of the Geographic Information System (GIS) that we started to link database records, documents and files to digital maps. Up till then, it was maps, for maps sake. GISs are expensive, complex, and unwieldy systems that often need an army of GIS experts to operate and maintain. Enterprise GIS projects often create a permanent new department, to feed and care for the GIS.
So “mapping for the masses” unleashed by services like Google’s My Maps, or Microsoft’s Live Search Maps has a lot of appeal. As a side note, the launch of these services has likely caused some angst for VC-backed Google Map mashup services like Platial, Flagr, Frapper and Plazes – Om Malik was the first to suggest the demise of the push pin “mashup gang”, in his popular blog.
The stories in the highlighted articles boil down to fundamental concepts – publishing, sharing and finding “user generated” content about, or in context of a location (with maps as a user interface). In fact, there are two approaches being discussed; 1) sharing map or geospatial data, and 2) sharing non-spatial content linked to a “pin” in a web map.
The story in The Province, declares the City of Nanaimo, BC in Canada, to be rulers of Google Earth (according to Google). Effectively, what they have done is use a data conversion tool, from Safe Software (a great little Canadian company), to publish geospatial or geodata (map content from the city’s GIS system), within Google Earth, using Google’s proprietary Keyhole Markup Language (KML) format. This is an interesting concept if it catches on with local government agencies. It could potentially enable Google to bypass the global map data aggregators, like Tele Atlas and Navteq. I am curious to know if Google has agreed to pay royalties to the City of Nanaimo, for use of their geospatial data that was created at significant taxpayer expense.
Historically, ESRI has dominated the market for GIS systems used by state and local governments. Government agencies have spent hundreds of millions of dollars on GIS systems and programs to create the most current, accurate and complete map and local geodata available. Yet, despite these longstanding relationships, you don’t hear much about ESRI’s web-based Geography Network that was designed to share geodata, using the proprietary .shp file format. If Google succeeds where ESRI did not they could tap the best source(s) of digital map data in world.
Finally, Fortius One, another private, VC-backed company has an interesting “clearing house” model for publishing and sharing geospatial data, using Google’s KML. Brad Forrest, of O’Reilly Radar fame describes the FortiousOne GeoCommons data repository as the Flickr/Swivel/YouTube/Sribd of geodata.
The second approach discussed in the articles is “push pin” map mashup applications, where communities of users link pictures (Google’s Panoramio or Flickr’s GeoTag service), notations, hiking or bike trail routes collected by GPS recievers, and a myriad of other user generated content to web maps. For the most part, this is not geospatial data (except for the GPS points collected on the bike ride), and needs a “pin” or “feature” as a geographically-aware object in the mapping system, in order to link in the content. I have written about the challenges with this in another post, and about the emerging problems with map spam. Simply put – mapping systems were never designed to be search engines and vice versa, and there is currently no concept like a unique internet address, or URL, to uniquely identify a place in the real world. We believe this is required to reliably organize, integrate and retrieve non-spatial content, based on location context. Lat/lon does not work – as geographic coordinates or “positions” do not uniquely identify a specific building, let alone a suite in an office tower or a store in a mall – and user generated push pins, content and “tags” don’t allow for adequate categorization, filtering and relevance ranking.
There are two very interesting references in the NY Times and Wired articles, that begin to shine a light on this issue and the future of what we have called PlaceSmart Search(tm) and PlaceSmart Information(tm) web services.
NY Times: “They (amateur map makers) are creating a new kind of atlas that is likely to be both richer and messier than any other… They are also turning the Web into a medium where maps will play a more central role in how information is organized and found.”
Wired: “Google Earth and Google Maps have long had search boxes, but you couldn’t find much… other than links to sites that Goolge had generated themselves, such as Yellow Pages.” To quote John Hanke, Google’s Director of Google Maps and Google Earth… “Now that you’ve got a lot of stuff out there (cluttered, unclassified push pins with linked content, or map spam – my paraphrase), it will become important to sift the wheat from the chaff…”
The web, paricularly in the transition to Web 3.0 or the Semantic Web, where context is king, requires a new approach. Simply sticking millions of pins in Google maps (with user generated tags) or layering a gazillion geodata layers on Google Earth, using KML – does not organize the worlds information based on location context. All it does, according to Google’s mapping chief, is create a chaotic pile of “stuff”, and of course lots of traffic to Google’s mapping services.
Now that the cost and complexity of GIS and mapping technologies have been virtually eliminated by Google, Yahoo and Microsoft, the next challenge is coming to the forefront – indexing, integrating and retrieving non-spatial data, for use in location-enabled and map-based applications, such as local search and personal navigation.
As I have said throughout this series on local search – we believe It’s about Place… and to solve this problem we need a universal, location-aware Place ID(tm) in order to organize, integrate and find content associated with places or geographic areas of interest.
To be continued…