Extract from http://www.fcw.com/geb/articles/2001/1203/web-gml-12-03-01.asp, December 3, 2001.
"Redrawing the map" [Government E-Business, Dec. 3, 2001]
The success of Extensible Markup Language, which is emerging as the universal language for e-government and other online transactions, has produced an offshoot — Geography Markup Language (GML) - that has the potential to change how state and local agencies share mapping information.
Like XML, GML is designed to serve as a common way of exchanging data between systems, whatever proprietary geographic information system (GIS) software the systems use. Typically, a mapping application can interpret data only if it was created with its own software. GML would remove that barrier, because any application that understands GML would be able to work with GML data.
It's not unlike translating at an international conference. Rather than providing a translation service for every participant, conference planners figure that most people will know one or two common languages. Just as English or French may make it possible for people from around the globe to exchange ideas, GML proponents believe this evolving technology could give GIS data broader currency.
"Interoperability is the bottom line of this GML activity," said Roger Harwell, mapping and GIS solutions product manager for GeoMedia WebMap products at Intergraph Corp. "It will allow interoperation among various vendors so users can grab data from anywhere they want in order to accomplish an application. It just makes a whole lot more data available to people."
XML allows digital data to be identified and "tagged" so that Web servers can talk to each other and pull similar kinds of data together from a number of sources and then display it in different forms on Web pages. It supports links that point to multiple documents, as opposed to the links provided through HTML, the current standard for building Web pages, which reference just one destination on the World Wide Web.
GML similarly tags geographic data so it can be collected from different sources and displayed as maps or in other forms using a standard Web browser. That's hard to do with current GIS tools, which use vendor-specific, proprietary data formats.
Many government agencies have avoided incompatibility problems by using a single vendor solution. Even there, however, GML could help by making GIS data available to many other people in government without the need to install and maintain software separately on desktop computers.
For instance, a person in a city or county planning department could access GIS data generated in another department as long as that data had been converted into GML format. The data could be pulled from the Web server and displayed using that person's desktop browser.
GML also could be used to splice data that doesn't already have an obvious connection or a simple way of being displayed together.
"I think GML has a big role to play in government, particularly with location-based services," said Joe Sewash, GIS analyst in the Tennessee Department of Finance and Administration's Office for Information Resources. "If you have an agency with a lot of field work, for example, it would tend to have a lot of different databases with explicit GIS links to them. GML can be used to link a lot of traditional nonspatial data with this spatial data."
Ultimately, GML could change not just how agencies share data but how they store it, making it possible to develop large databases of geographic information that could be used to support multiple applications.
Government agencies usually generate and store geographic data with only their own applications in mind. A city or county in Tennessee, for example, would have data on rivers and streams that would likely be generated and used differently than data in Oregon.
However, there are increasing numbers of applications that need access to this kind of data on a global basis for such things as large-scale analysis of geographic features, or for specific applications in areas such as environmental protection.
By storing locally generated data in GML format and using XML-based functions such as XLink and XPointer to provide links to these local databases, GML could be used to create globally distributed databases of geographic information.
But some GIS vendors do not expect GML to have such a broad impact.
For one thing, existing GIS software already can store data in various formats, said Dave Danko, senior consultant for Environmental Systems Research Institute Inc. Data sharing is GML's "main reason for being," he said.
There's also the power of familiarity. Users have become adept at manipulating and storing GIS data in its native format, said Bill Shelley, manager of GIS and remote sensing software for ERDAS Inc. "The real potential of GML is in Web-enabling geographic applications," he said. The development of the basic GML standard is being overseen by the Open GIS Consortium, which draws members from government agencies, research groups, oil companies and technology companies.
However, despite the initial interest in GML, it's still unclear just how much demand for GML-based Web services will develop, or how soon. Even those vendors working within the Open GIS Consortium who are familiar with the promise of GML are cautious about committing themselves too far in advance.
"ERDAS is not doing anything with it today because we want to see which direction the market takes," Shelley said. "We'll certainly start using it when we have customers that want us to support GML, but we haven't had a large contingent even ask us about it so far."
From what he's heard, Neil MacGaffey, assistant director of MassGIS, Massachusetts' Office of Geographic and Environmental Information, believes it's far too early to predict how things will go."This standard is still in development," he said, "and it is not clear that all GIS vendors are necessarily going to adopt it."
Tennessee's Sewash also thinks it's too early to tell because "it will take time for vendors to actually provide systems that take advantage of [GML]." Nevertheless, given that XML will probably have a big impact on traditional IT applications, the obvious synergy between GML and IT makes it likely that GML will be used, he said.
Even though Shelley is cautious about the future demand for GML, he thinks it stands a good chance of being widely adopted in industry. And, since it typically is a "background" technology that enables applications, it may eventually be employed without users being aware of it.
"I'm very optimistic about GML in the long run," Shelley said. "I suspect it will be only a couple of years at the most until we start supporting it in our products."
Robinson is a freelance journalist based in Portland, Ore. He can be reached at hullite@mindspring.com.