Web Data Spaces

Now that broader understanding of the Semantic Data Web is emerging, I would like to revisit the issue of "Data Spaces".

A Data Space is a place where Data Resides. It isn't inherently bound to a specific Data Model (Concept Oriented, Relational, Hierarchical etc..). Neither is it implicitly an access point to Data, Information, or Knowledge (the perception is purely determined through the experiences of the user agents interacting with the Data Space.

A Web Data Space is a Web accessible Data Space.

Real world example:

Today we increasing perform one of more of the following tasks as part of our professional and personal interactions on the Web:

  1. Blog via many service providers or personally managed weblog platforms
  2. Create Event Calendars via Upcoming.com and Eventful
  3. Maintain and participate in Social Networks (e.g. Facebook, Orkut, MySpace)
  4. Create and Participate in Discussions (note: when you comment on blogs or wikis for instance, you are participating in, or creating, a conversation)
  5. Track news by subscribing to RSS 1.0, RSS 2.0, or Atom Feeds
  6. Share Bookmarks & Tags via Del.icio.us and other Services
  7. Share Photos via Flickr
  8. Buy, Review, or Search for books via Amazon
  9. Participates in auctions via eBay
  10. Search for data via Google (of course!)

John Breslin has nice a animation depicting the creation of Web Data Spaces that drives home the point.

Web Data Space Silos

Unfortunately, what isn't as obvious to many netizens, is the fact that each of the activities above results in the creation of data that is put into some context by you the user. Even worse, you eventually realize that the service providers aren't particularly willing, or capable of, giving you unfettered access to your own data. Of course, this isn't always by design as the infrastructure behind the service can make this a nightmare from security and/or load balancing perspectives. Irrespective of cause, we end up creating our own "Data Spaces" all over the Web without a coherent mechanism for accessing and meshing these "Data Spaces".

What are Semantic Web Data Spaces?

Data Spaces on the Web that provide granular access to RDF Data.

What's OpenLink Data Spaces (ODS) About?

Short History

In anticipation of this the "Web Data Silo" challenge (an issue that we tackled within internal enterprise networks for years) we commenced the development (circa. 2001) of a distributed collaborative application suite called OpenLink Data Spaces (ODS). The project was never released to the public since the problems associated with the deliberate or inadvertent creation of Web Data silos hadn't really materialized (silos only emerged in concreted form after the emergence of the Blogosphere and Web 2.0). In addition, there wasn't a clear standard Query Language for the RDF based Web Data Model (i.e. the SPARQL Query Language didn't exist).

Today, ODS is delivered as a packaged solution (in Open Source and Commercial flavors) that alleviates the pain associated with Data Space Silos that exist on the Web and/or behind corporate firewalls. In either scenario, ODS simply allows you to create Open and Secure Data Spaces (via it's suite of applications) that expose data via SQL, RDF, XML oriented data access and data management technologies. Of course it also enables you to integrates transparently with existing 3rd party data space generators (Blogs, Wikis, Shared Bookmrks, Discussion etc. services) by supporting industry standards that cover:

  1. Content Publishing - Atom, Moveable Type, MetaWeblog, Blogger protocols
  2. Content Syndication Formats - RSS 1.0, RSS 2.0, Atom, OPML etc.
  3. Data Management - SQL, RDF, XML, Free Text
  4. Data Access - SQL, SPARQL, GData, Web Services (SOAP or REST styles), WebDAV/HTTP
  5. Semantic Data Web Middleware - GRDDL, XSLT, SPARQL, XPath/XQuery, HTTP (Content Negotiation) for producing RDF from non RDF Data ((X)HTML, Microformats, XML, Web Services Response Data etc).

Thus, by installing ODS on your Desktop, Workgroup, Enterprise, or public Web Server, you end up with a very powerful solution for creating Open Data access oriented presence on the "Semantic Data Web" without incurring any of the typically assumed "RDF Tax".

Naturally, ODS is built atop Virtuoso and of course it exploits Virtuoso's feature-set to the max. It's also beginning to exploit functionality offered by the OpenLink Ajax Toolkit (OAT).