Down to Semantics

At this year's Web Directions South conference in Sydney, David Peterson presented "Semantic Web for Distributed Social Networks".

At this year's Web Directions South conference in Sydney, David Peterson presented "Semantic Web for Distributed Social Networks".

Semantic Web
Although many of the semantic web technologies are already here, according to Peterson we're still far from what Tim Berners-Lee had envisioned.

In the semantic web the information is supposed to be collection and object-based. The current generation of web pages are still only documents.

Distributed Social Network
The main characteristic of a distributed social network is that you own and control the data about yourself. You can edit or delete it and decide who can have access to it.

There have been initiatives towards a distributed social network -- one being the OpenSocial.

Three main principles of distributed social networks are shared understanding, central identity and giant global graph.

Shared understanding
Tagging information adds meaning, but a word can mean more than one thing. For instance, 'Paris' can mean the city Paris or Paris Hilton. Some websites attempt to group relevant information together by adding prefixes to tags. For instance, on all Web Directions South content has been marked with a WDS08 tag, so that it can be accessed from a single location.

Central Identity
The second principle of distributed social networks is that individuals should have a single identity. OpenID is a technology that facilitates a central identity. With OpenID you can use one identity to login to multiple websites. Another authentication method that has been proposed is RDFauth -- based on already established web standards like REST and SSL.

The Giant Global Graph
The third and final principle -- the gigant global graph represents connections between objects. The idea is that we should be moving away from a web of documents to a web of things.

Technologies underpinning the semantic web are:

  1. Microformats: The problem with the current HTML tags is that they don't reveal the meaning of content. Microformats are an attempt to use the HTML tags to show the metadata and other attributes. The end users can then know what category the information falls into (eg. contact, geographical, calendar etc).

    Some of the microformats available are:

    |> hCard – contact information -- adr for address and geo for geographical coordinates
    |> hCalendar - events
    |> hReview - reviews
    |> hResume - CVs
    |> XFN – human relationships
    |> rel-license – specification for copyright license
  2. Ontologies
    If the existing microformats aren't enough, you can create your own vocabularies called ontologies.
    Ontologies are vocabularies of descriptive words for concepts within a domain of knowledge and the relationships between those concepts.
  3. RDF
    Resource Description Framework (RDF) is an XML W3C standard for describing and exchanging metadata.
    RDF lets you add meaning to your web content, so that your documents can be processed by applications.
    SPARQL is an RDF query language that allows you to perform multiple queries on distributed data end points at one time.