Today I finished an installation of a PostgreSQL/PostGIS instance on one of our Red Hat development servers. It's populated with geo-spatial data from our campus, and is set up (as of now) so we can query it from our apps to determine three region levels: building, campus region, and lastly, the entire campus. With that we should be able to get much more meaningful description of a user's location. This first version really just has regions arbitrarily defined by me, but we can certainly refine those later.
What this means is that we can now report a user's location and receive in return a container name (that is, x,y is within this described polygon), as well as a URL that is linked to meta-data about that place (an RDF). The client must then do something appropriate. An intelligent client ideally should handle the RDF and extract the low-level details fully describing the place and acquire available services.
For visualization of location, I'm setting up a MapServer instance to provide relevant images of the place. At first these will be available through WMS.
So what we want near-term is the ability to know where one is (text), and and a visual representation of same (.png, etc.).
Related: tonight I was playing with OpenJUMP, and took this screenshot: