Dear Geospatial Web: WHERE are the Semantics?
Geospatial Semantic Web sounds great. You can organize a conference with this title, and everyone will think your a genious, or at least you’re a mad scientist.
Locations are increasing importance everywhere in the web. And it’s easily understandable, most of information has a geographic dimension. In the last two years we’ve added this dimension to our popular Web 2.0. By 01/01/2008, Wikipedia had 7,5 million articles and 800,000 of them were geotagged using geonames. 2,2 million photos were geotagged in Flickr, Just in December 2007!. And the list of social webs that allows resources geotagging, or collaborative mapping, grows every day, Youtube, Picasa, Panoramio, Wikimapia, OpenStreetMap etc.
As we can see in the image from google trends, although geospatial enabled applications (Panoramio, Wikimapia) are growing in popularity quite fast, more theorical related topics haven’t taken off already (semantic web, geospatial, geo web).
It’s amazing how hundreds of thousands of people are collaborating to these geo - apps for free, but for some of us, a geographically enabled Web 2.0 is not enough. We want more (just a bit more?).
Humans have fantastic geospatial reasoning skills. We can understand very fast where is something, if it’s near, next to something, within something… and sometimes these adverbs can mean different things.
To make a machine to deal with these human skills is a huge complex task, and one of my main research interests. Geotagged Web 2.0 is not enough to achieve this goal. To keep the data in a bit more complex format is needed. I’m talking obviously about Semantic Web.
There are geospatial ontologies or metadata standards which already can keep geospatial information with better accuracy, and therfore they able us to reason with information locations. We have:
- Geography Markup Language (GML), from the Open Geospatial Consortium, is a huge modeling language for geographic systems, and it aims to be an open standard, in coordination with ISO standards.
- Keyhole Markup Language (KML), is the popular format used by google apps (Google Earth) to annotate an visualize geographical information. It is focused mainly to visualization. It can be stated as OGC standard by 2008.
- GeoRSS, an emerging standard for encoding geographic location as part of RSS feeds. It has two enconding formats, GeoRSS - Simple, a lightweight format for basic geographic needs, and GeoRSS - GML, which is in fact a formal GML Open Application Profile.
- Geonames Ontology. The OWL encoded ontology used by this open geographical data base.
- The recently proposed GeoOWL ontology from the W3C Geo Incubator Group. This ontology closely matches GeoRSS vocabulary. But it’s been kept much simpler than GML model. They wanted to develope an ontology for the web, instead of an ontology for spatial databases.
So we’ve got the tools to improve our brand new GeoWeb. Now we have to increase the amount of information compatible with this standards, and develop services that make use of them. And that’s all folks!
