Geospatial Semantic Web Blog

I enjoy using Flickr, especially its geotagging feature. Not only it helped me to organize photos based on location, but also helped me remember the context in which my photos were taken.

For example, during my recently trip to Paris and London, I took some photos at the Château de Versailles and the Windsor Castle. Using Flickr, I was able to geotag photos in a level of detail that I never thought would be possible. Without it, in the next few months, I probably would have forgotten the exact locations where some of my photos were taken (e.g., this, this and this).

See my photos taken at the Windsor Castle

See my photos taken at the Château de Versailles

Flickr announced a new feature that allows users to geotag photo — i.e., annotate photos with location information such as latitude and longitude coordinates. Unlike typical keyword annotation where users manually enter text strings of keywords, Flickr geotag allows users to annotate location information by simply drag-and-drop photos onto a digital map.

Since most Web users are comfortable with online mapping applications (Google Maps, Yahoo! Maps), for them to learn to use Flickr geotag shouldn’t be too difficult. Flickr even produced a short video that teaches users how to geotag photos and set privacy protection.

I think Flickr has done a great job building this new feature. First, there is little mention of GIS vocabularies in the whole tagging process (e.g, you don’t have to know what lat/lng is before you can geotag). Second, it has built-in privacy protection. This gives a greater comfort to those who put information on the Web. Flickr users will be in control of the information and not Flickr. Third, the geotag map UI looks better than the one in Yahoo! Maps. To me, it seems to have a cleaner design with less clutter.

Spotted on Anything Geospatial

People often ask the question:

• Is geospatial ontology just a bunch of tags (e.g., XML tags, technorati tags or flickr tags)?
• Why is it so important to differentiate tags from ontologies?

Before I try to answer these questions, let me say few words about the term “ontology” with respect to computing systems.

What’s Ontology?

The concept of ontology originates from the studies of philosophy. Today in computer science, especially in the field of Artificial Intelligence, ontology is a set of vocabularies and relations that people or computing machines must agree upon in order to communicate. Without sharing a common ontology, they won’t be able to share information effectively.

Why I think Tags are not Ontology

Let’s take the technorati tag “cell” as an example. The term cell in our pop culture could mean cellphone, the smallest basic unit of an animal of a plant, or a small residential accommodation in a prison. Let’s say you tagged one of your blog post with the tag “cell”. This blog post was later discovered by a technorati search bot. Now the question is how much information can the search bot learn from the tag “cell”?

No much. All it knows is that your post is tagged with an English word that spells ‘c’, ‘e’, ‘l’, ‘l’. It probably doesn’t even know your post is about cellphones, human cells or prison cells.

A natural question to ask is that can we do better? For example, can we help the search bot to know more about the content of our blog? In other words, can we help computing systems to understand the semantics of information?

I don’t know for sure, but I do believe it’s possible. In fact, many ongoing Semantic Web efforts in OGC and W3C are working to solve this kind of problems.

Coming Back to Geospatial Ontology

Once we understand tags are not ontology, we can immediately conclude that geospatial ontology is not just a bunch of tags. What makes geospatial ontology so interesting is that much of the information in our everyday life involves some kind geospatial ontology. If we are on the path to develop computing systems that could understand the semantics of our blog posts, for example, we must pay extra attention to geospatial ontology.

In the near future, I will discuss geospatial ontology development and its usage in real world applications.