Daniel’s Blog

Data Spaces Philosophy: Why and How?

Data Spaces are the best way to look at user related/orientated data, and I am not just saying that because I work for OpenLink Software who invented OpenLink Data Spaces (ODS). I shall explain using diagrams that I created for explaining OpenLink Data Spaces.

The best way to visualise a Data Space is by imagine that you and your data are in a Solar System:

And your Solar System can build relationships with other Solar Systems, this can be:

• a contact/friend (foaf:knows)
• an identity which is really the same as you (owl:sameAs)
• or something else, such as a pointer to see something else (rdf:seeAlso)

You can start to picture that there are multiple solar systems in a galaxy and multiple galaxies in a universe - which is how we achieve congruency with the universe scale.

In this universe of relationships everything has its own name, just like in Stargate SG1 P3X-272 (coincidently a planet with a repository of knowledge), this name is the universal key, the Unique Identifier also known as a Uniform Resource Identifier (URI).

As you may have noticed, our solar system is made up of objects described in Resource Description Format (RDF):

It uses technologies which enable the Semantic Web Vision which was proposed by Berners-Lee, Hendler and Lassila in 2001 [1], and then extended by Shadbolt, Berners-Lee and Hall in 2006 [2]. A Data Space system is a bit like nature - everything is related somehow, this is achieved by making full use of Linked Data (which involves linking data across knowledge bases / domains of information to achieve semantics for human readers and machine readers). As you can see Linked Data is very useful within a system and across systems (e.g. across galaxies in our data space analogy). Kingsley Idehen and Michael Bergman have interesting blog posts about Linking Data titled “FOAF-ing Linked Data is quite SIOC-ing” and “Linked Data Comes of Age” respectively.

A Data Space will use vocabularies such as FOAF and SIOC to interconnect data:

Everything in a Data Space is accessible via a URI, and can be queried using a language such as SPARQL.

This is all available now through OpenLink Data Spaces which lays on top of OpenLink Virtuoso Universal Server, and I expect that this kind of Linked Data and Data Space philosophy will become more and more common over this year (2008).

{p.s. you may have noticed that this blog post is full of links, these are actually me manually Linking Open Data! By doing this I am actually expanding the Giant Global Graph. For some explicitly human readable references see below}

Human Links:

• [1] Tim Berners-Lee, James Hendler & Ora Lassila: “The Semantic Web“, Scientific American 284(5):34-43 (May 2001)
• [2] Nigel Shadbolt, Tim Berners-Lee and Wendy Hall “The Semantic Web Revisited.” IEEE Intelligent Systems, 21 (3). pp. 96-101. (2006) ISSN 1541-1672
• OpenLink Software
• OpenLink Data Spaces
• Kingsley Idehen - Blog
• Michael Bergman - Website
• RDF Project on the W3C Website
• FOAF (Friend of a Friend) Vocabulary - Website
• SIOC (Semantically Interconnected Online Communities) Vocabulary - Website
• SPARQL Specification
• Semantic Web Project on the W3C Website
• Linked Data Resources
• The Linking Open Data Project

Folksonomy Links:

Technorati Tags: dataspaces, openlink, linked, data, semanticweb, web30, web20, rdf, dataweb