Courtesy of Nova Spivack's post titled: Tagging and the Semantic Web: Tags as Objects, I stumbled across a related post by John Clarke titled: Tagging and the Semantic Web. Both of these posts use the common practice of tagging to shed light on the increasing realization that "The Pursuit of Context" is the fusion point between the current Web and its evolution into a structured Web of Linked Data.

How Semantic Tagging Works (from a 1000 feet)

When tagging a document, the semantic tagging service passes the content of a target document through a processing pipeline (a distillation process of sorts) that results in automagic extraction of the following:

-- Subject matter Entities (Subject matter Concepts reflecting topics covered by the document

Once the extraction phase is completed, a user is presented with a list of "suggested tags" using a variety of user interaction techniques. The literal values of elected Tags are then associated with one or more Tag and Tag Meaning Data Objects, with each Object type endowed with a unique Identifier.

Issues to Note

Broad acceptance that: "Context is king", is gradually taking shape. That said, "Context" landlocked within Literal values offers little over what we have right now (e.g. at Del.icio.us or Technorati), long term. By this I mean: if the end product of semantically enhanced tagging leaves us with: Literal Tag values only, Tags associated with Tag Data Objects endowed with platform specific Identifiers, or Tag Data Objects with any other Identity scheme that excludes HTTP, the ability of Web users to discern or derive multiple perspectives from the base Context (exposed by semantically enhanced Tags) will be lost, or severely impeded at best.

The shape, form, and quality of the lookup substrate that underlies semantic tagging services, ultimately affects "context fidelity" matters such as Entity Disambiguation. The importance of quality lookup infrastructure on the burgeoning Linked Data Web is the reason why OpenLink Software is intimately involved with the DBpedia and UMBEL projects.

Conclusions

I am immensely happy to see that the Web 2.0 and Semantic Web communities are beginning to coalesce around the issue of "Context". This was the case at the WWW2008 Linked Data Workshop, I am feeling a similar vibe emerging from the Semantic Web Technologies conference currently nearing completion in San Jose. Of course, I will be talking about, and demonstrating practical utility of all of this, at the upcoming Linked Data Planet conference.

Related

Faviki (note: this service needs to expose Linked Data compliant Tag URIs)