http://www.openlinksw.com:443/blog/kidehen@openlinksw.com/blog/?id=1360 Nova Spivack has just penned a post titled: On the Difference Between &quot;Semantic&quot; and &quot;Semantic Web&quot;, where he covers the fundamental difference between &quot;Semantic&quot; (what I call &quot;Semantics Inside&quot;) and &quot;Semantic Web&quot; applications. I would like to extend the distinctions further by adding the &quot;Linked Data Web&quot; distinctions to the developing discourse. The Linked Data Web (aka. Linked Data) describes RDF data injected into the Web, where the Data Object Identifiers (URIs) in an RDF graph (collection of RDF triples) are endowed with HTTP based URIs. The net effect of this approach to Data Object Identity is that it facilitates &quot;Open Data Access by Reference&quot; on the Web (aka data dereferencing). If you recall pre Web ubiquity, in the enterprise realm for instance, Open Database Connectivity (ODBC) emerged as a mechanism for separating Data Access and Data Management in the database oriented Client-Sever model. Although ODBC gave you access to data, the data access entry point took the form of a data access specific naming mechanism called a &quot;Data Source Name&quot; (DSN). ODBC DSNs typically exposed Tables or Views. The same thing applies to JDBC where a non HTTP based URN scheme applies. Zip forward to where we are today on the Web; the Web is evolving from a Document centric Database to a Distributed Object Database, and you should see that in Linked Data we are now truly looking at the best of all worlds: Web Open Database Connectivity (WODBC) with the following advantages: - direct Access to a single Record (an Entity) or Record Sets (RDF based Entity Sets) by reference over HTTP across disparate Data Spaces on the Web - the ability to mesh disparate data sources without being impeded by back-end DBMS engine model, vendor, host operating development frameworks, or host operating system specificity - an opportunity to learn from the enterprise DBMS market and Client-Server markets of yore with regards to the shape and form of next generation Linked Data Web oriented solutions. To conclude, we now have &quot;Semantics Inside&quot; (RDF or non RDF), &quot;Semantic Web&quot; (RDF graphs with Object Identifiers that may or may not be HTTP based), and &quot;Linked Data Web&quot; (RDF graphs with Object Identifiers that must be HTTP based and dereferencable) oriented applications, in the emerging landscape associated with the &quot;Semantics&quot; moniker. As per usual, this post is a record in my Blog oriented Data Space on the Web. The permalink of this post is a URI constructed with Giant Global Graph enrichment in mind :-) Thu, 15 May 2008 14:11:13 GMTVirtuoso Universal Server 08.03.3334Kingsley Uyi Idehenhttp://www.openlinksw.com:443/weblog/public/images/vbloglogo.gifhttp://www.openlinksw.com:443/blog/kidehen@openlinksw.com/blog/?id=1360I have seen the future and it's full of Linked Data! :-)8831 http://www.openlinksw.com:443/blog/kidehen@openlinksw.com/blog/?id=1360#4567http://www.openlinksw.com:443/blog/kidehen@openlinksw.com/blog/?id=1360#4567kidehen@openlinksw.comThu, 15 May 2008 18:23:00 GMTGraham, A bug in my glossary that drives automatic hyperlinking in my data space. If the machine generated hyperlinks are scoped to a space that is clearly defined, then there simply isn't an issue :-) It's a claim from a space on the Web, it's a point of view and nothing more. Machines, like humans should be able to : Agree to Disagree :-)http://www.openlinksw.com:443/blog/kidehen@openlinksw.com/blog/?id=1360#4566http://www.openlinksw.com:443/blog/kidehen@openlinksw.com/blog/?id=1360#4566gjh@bel-epa.comThu, 15 May 2008 17:48:45 GMTHmm. The link in "The net effect of this approach" points to "http://dbpedia.org/resource/.NET_Framework", which clearly isn't contextually relevant. Some other links are also contextually dubious, many are simply superfluous. <p> I'm /really/ not looking forward to an onslaught of machine-generated hypertext links contaminated with pseudo-semantic category errors. It's tantamount to semantic phishing. Have you given any thought to the human usability issues engendered by this "machine-friendly" approach?