mSpace mobile pulls together a number of current technologies in innovative ways, and wraps them in "next generation" semantic web technologies to support rapid, constextual exploration of spaces and places relevant to you.
mSpace Mobile works across a series of Web Services. Web Services are used to provide fixed protocols across the architecture described below. All information transfer from query requests and Semantic results are organised through .NET web services. Currently we have three: our MK Server, our MQ server and our Buddy Ratings Server.
The machines that process Web pages don't know anything about their content. Search engines look for documents that match up words to produce lists of results (their cleverness is in no small part how they rank what goes at the top of the page).
With the Semantic Web, data gets added to pages that helps the machines make connections between pages, and even to reason about pages. For instance, you may want to know about all the researchers in a university who have ever done projects in Human Computer Interarction (HCI). Today, if some person's web pages doesn't explicity use those terms to describe what they've done, a search engine won't find the projects, even if all these projects are about HCI. With the Semantic Web, the project name in one page can be connected through a set of relations, to a description of those projects somewhere else that says "that project is in the area of HCI" and thus tell the searcher about relevant information.
But more than just returning more and better lists of search results, the Semantic Web can be used to make associations, and this making connections is what mSpace leverages in particular. mSpace is able to associate information with meaningful and multiple categories: when exploring locations, these kinds of groups may be pubs, restaraunts, landmarks, events. Things can shift contexts: looking at a concert listing may stir associated questions about not just the event, but about who are the composers? what is their music like?
Web Scale mSpace leverages the association building power of the Semantic Web to support easy movement between these kinds of related explorations at "web scale." The Semantic Web provides mechanisms for anyone to publish information that can be picked up by other Semantic Web applications. That's a powerful potential source of information. To take advantage of all these Web sources, mSpace needs a way to locate them and coordinate them. To that end, we've designed a distributed architecture to make it possible for anyone either to publish information about anything that an mSpace might use, or for anyone to wrap an mSpace around any area of interest.
Three-Layer Design mSpace implements a three-layer architecture, partly to abstract and distribute the Semantic processing.
MK Every datastore within the mSpace model is handled by an associated mSpace Knowledge Server (MK). Here we see that London Local, amongst others, has a dedicated and powerful server designed to handle its dissemination of information. This abstraction level allows access to any chosen Semantic storage, by presenting with a fixed protocol.
MQ mSpace Query Servers are powerful machines which handle the complicated and intense processing required to query across multiple large MKs; this is the function of this abstraction. This distribution also allows less powerful clients (MAs) to access required Semantic information efficiently. Queries are constructed here and sent over fixed protocols to the MKs for results.
MA mSpace Mobile is an example of an mSpace Application (MA). Many mSpace applications can exist and may be conerned with various sources of knowledge. With the three-layer abstraction above, a less-powerful mobile device can easily access information via any discoverable MQ, which will produce and execute large complicated queries and return simple results.
Open Guide Many sources of Semantic information are developing across the world as we speak. Projects such as Wikipedia and the Open Guide are continuously allowing users to record new information, which is being stored in RDF. These types of developing resources can be easily accessed through an mSpace model. For mSpace mobile, the recorded GPS information within the Open Guide to London is directly applicable for the London Demonstrator. Open Guides are developing for numerous cities across the world, and here in Southampton, the mSpace team has begun developing an Open Guide to Southampton.
Web Scale Note that these developing open sources of Semantic Information relate to our vision for Web Scale. Anybody can set up a web page on the Internet if they know how. In the same way, as people learn to publish semantic information freely, the mSpace model can be applied to these knowledge sources and become readily available through MAs.