BendyConf: The Internet Is Your New Database

In his talk The Internet Is Your New Database, Will laid bare the ideas behind the semantic web. While usually shrouded in misunderstanding, the semantic web has lots of cool ideas packed in it.

Triples

Will started with the simple notion of a triple:

1. Subject: Will “Fresh Prince” Smith
2. Predicate: from
3. Object: West Philadelphia

Or as I first thought of it “a thing has some relationship to another thing.” The cool part about RDF (Resource Description Framework) is that it makes this relationship much more rigorous. When I said thing before, it would be better to replace that nebulous idea with URI, or Uniform Resource Identifier. That is, when you’re talking about some noun, why not reference it explicitly? And when I said relationship before, we can do the same sort of trick; we concretely name a relationship from some available schema.

SPARQL

SPARQL is the whimsically named query language for the semantic web. What it lets you do is plumb the graph for interesting tidbits of information all while using a familiar(ish) language. Using Will’s example:

SELECT (?age) WHERE {
    ex:Fresh foaf:age ?age
}

Which translates to something like “find me all the tuples that match ex:Fresh with a foaf:age predicate and any object” or put more simply: “find the age of Will “Fresh Prince” Smith.”

Queries can involve intermediate variables that are not expressed in the output:

SELECT (?nick) WHERE {
    ex:Fresh ex:aunt ?aunt.
    ?aunt ex:nickname ?nick.
}

Here we travel through the intermediate ?aunt object in order to get the ex:nickname of whatever entity that is. Logically, this is like saying “find me an ?aunt such that it successfully completes the two parts of the join.”

Summary

Will had a lot more to say about these topics and also covered more ground than this summary. Check out the slide deck and video.