Geek Speak: Ron Lake, creator of Geography Markup Language

Ron Lake is a pioneer in the building of the GeoWeb. The 60-year-old founder and chief executive officer of Vancouver-based Galdos Systems created Geography Markup Language, a Extensible Markup Language-based standard for encoding geographic data.

Lake wrote the first version of GML, which was published in 2000 by the Open Geospatial Consortium, and most of GML 2.0, which became a standard in 2001. GML 3.2.1 is the latest specification.

From July 27 to 31, about 220 people from 20 countries attended the GeoWeb 2009 conference at Simon Fraser University’s Morris J. Wosk Centre for Dialogue. Lake and his company started the annual event in 2002 as GML Dev Days.

The Georgia Straight reached Lake by phone at his downtown office.

How do you describe what your company does?

That’s sometimes difficult. So, basically, we provide software—either software solutions or we actually do have some products—that automatically aggregate and/or distribute geographic information or, you could say, facilitate the sharing of geographic information.

Why would you want to do that? Well, many organizations collect information about the world. Obviously, in the city, you’ve got organizations that might be like a national police organization, you might have a regional water company, you might have a national postal authority, et cetera, et cetera. Any of these groups need information from one another. Today, that happens in a kind of ad hoc fashion. Somebody phones somebody and says, “You have any data on such and such?” What we’re trying to do is automate that information distribution and aggregation process.

If you think of something like Google or Microsoft—like Virtual Earth or Google Earth—the process by which they build that globe is still quite manual. We’re trying to automate it. We’d like to be able to automate it right down to, when a city changes something in the city, it automatically appears in places like that globe or in a regional atlas or national atlas. Atlas is not really the right word, but some kind of national or regional aggregation of that information.

What is the Geographic Web?

GeoWeb means a lot of different things. But, basically, it’s the evolving network of information about the world—but in a way that is geographically organized. So, if you want to find out things about Vancouver or you want to find out things about a building or you want to find out things about the fires in the Okanagan, then you can rapidly access, integrate that information.

What is Geography Markup Language used for?

Today, Geography Markup Language is intended to be kind of a general way of encoding information—and transmitting it around—that involves geography. That is a little bit hard for general people to grasp. But it means about the shapes of things, their spatial extent, how they’re connected to one another.

So, GML is usually used to create other languages that use geography. Some of the major ones: there’s a language called AIXM, which is basically a language for communicating information for commercial aviation. So, this is in the process of becoming realized, but it’s already an international standard. So, organizations like FAA, the U.S. Federal Aviation Administration; Eurocontrol; eventually Nav Canada; eventually all the airplanes in the world will be communicating both with one another but also with ground resources.

That information—a lot of it—is spatial in nature, like where the airplane’s going; what the allowed fight zone of the airplane is; the location of navigation aids and their state, because if aids go off-line, then they stop working; taxiways on runways; obstacles like cranes or buildings that might be arranged on the approach path at an airport. So, it describes all that information, and it does so in a way that allows it to be automatically propagated between different systems, including the aircraft.

Today, you might be surprised, but the management of information around commercial aircraft is relatively manual—manual in the sense that it’s basically notification messages but they’re text. They’re not easily readable by machine. There’s maybe no guarantee that what the pilot knows is the same as what the aircraft’s electronics system knows. Of course, that’s not a good situation. So, there’s big moves to go to a new generation of air traffic control—AIXM, which is written in GML. That’s one such application.

Another one quite different, also discussed extensively at this conference, is something called CityGML. CityGML, again, is written in GML. But now it’s about modelling cities, not the flight of aircraft. But, again, the same kind of issues: you need to describe the geometry of buildings; what buildings are next to which; the characteristics of those buildings, maybe from the point of view of heat transfer, so that you can not so much produce pretty pictures of the building but people can do studies for planning whole cities or whole parts of cities. CityGML has got a fair following in Europe—Germany in particular. It’s part of an approach to urban design that is becoming both more analytical. We’re calculating more things, I guess, because of environmental issues, and people are looking at larger areas. So, that’s another application.

Another one is something called GeoSciML, where it’s being applied in the geological sciences.

What was the most interesting thing you heard at GeoWeb 2009?

To me, the most interesting thing was seeing very clearly the different perspectives that people had on what the Web was and how that influenced the positions that they took on various kinds of arguments. Just to illustrate that, I think there were sort of at least three different ideas of what the Web is.

One: what I will call the conventional or original idea of the Web as this massive collection of hyperlinked documents—you see a document, you click on it, you see another document or piece of a document—which is clearly very, very important.

But we also saw people who see the Web—obviously, they see that hyperlinked thing—but they also see the Web as evolving as a kind of mechanism to deliver services: computational services, viewing data-access services. So, maybe the people like the air-traffic-control people that were there were thinking in those kinds of terms.

Then we have the Web as a real-time collaborative vehicle. We saw David Boloker’s talk, where there was an example of telemedicine, where multiple radiologists were talking to one another, seeing one another in high definition, and manipulating the same application to mark up and measure things on a radiological image—an X-ray or CAT scan.

I think the fusion of all these different views of the Web is obviously critical to the evolution of the GeoWeb. Real-time collaboration and non-real-time collaboration is clearly important for urban planning. I think Ken Greenberg’s talk at the end of the conference showed that very clearly. But that also requires kind of point-to-point access to a lots of services, even if a person’s not accessing—that might be machine to machine.

Meanwhile, the conventional Web of documents is also very important. But I think that seeing those different points of view more clearly than maybe we’d seen them in the past was quite interesting to me, and seeing how it informed people’s, you know, what they said next was kind of exciting.

Where do you think the Geographic Web is headed in the coming decade?

I think there’s a lot of ways things can develop. But one of the ones that, I think, will become apparent is the GeoWeb as kind of reflecting the state of the world. I think I gave the example before that if you go to the Internet in general, or the Web in general, there’s this massive collection of documents. Basically, it’s like going into a library to do research. If I wanted to know the state of the climate or the state of polar-bear habitat or the state of traffic in downtown Vancouver and how that might be impacted by the bicycle bridge or closing lanes, there’s no place you can really quickly go and see those things.

I think that’s one important direction that the GeoWeb will evolve toward is this kind of summary of the state of things at various levels—not that it won’t be subject to argument as to whether it’s correct, much like Wikipedia. The articles in Wikipedia, you could say, are they completely right or not? I think that kind of debate would continue. But something where when you saw these visualizations of things relating to the state of various things in the world that you would know that and believe, I think, to some certain extent that was tied back to real information. If things change in the city, they’re reflected in it.

I believe that’s almost kind of a social evolution too. Well, first, we use the word we to mean maybe the whole globe. We, in that sense, are concerned with where things are going. I think that this idea of the GeoWeb as kind of summarizing the state of things in various critical areas—not by any one person, by thousands or millions of people—I think is a very important direction. I expect we’ll start thinking of it that way in the next five years.

What are your thoughts on the Semantic Web?

The Semantic Web, to me, means a body of technologies to add more meaning into the Web. As you know, if you do a search through Google or Microsoft or Yahoo or whoever, that search will return you all kinds of things that are not related to what you’re looking for, because the search engine doesn’t understand what you’re looking for. The search, while it’s very simple, it’s not sort of natural-language-like.

So, if you type the word “bar” into the search window, say “bar Vancouver”, then you’ll get a list probably of bars in Vancouver—places to drink. But maybe you’re looking for suppliers of, you know, metal bars or chocolate bars. You can start qualifying the word string by saying “Vancouver bar metal”, but it could be a long process in some cases. Where there is meaning ambiguity, it may take you quite a long time to find what you want.

So, people are trying to use so-called semantic technologies, ontologies to represent some of the meaning about what you’re looking for, so that the robots—the computer programs that build the indexes that are used by the search engines—can respond and give you more accurate kinds of searches, and also so you can phrase questions that are more English or natural-language-like—not necessarily English.

Those same kinds of technologies are obviously very important in the context of GeoWeb, because people are going to expect that there’s an understanding of meaning within the GeoWeb. As well, of course, the GeoWeb—geographic information—is a key part of semantics in general, a key part of our understanding of meaning. I think at this point one could say, “Is the GeoWeb part of the Semantic Web?” Probably. And maybe also the other way around. I guess that’s the way I would look at it right now.

Every Friday, Geek Speak catches up with someone in Vancouver’s technology sector, video-game industry, or social-media scene. Who should we interview next? Tell Stephen Hui on Twitter at twitter.com/stephenhui.