RDF Abstract Syntax

Editors:

This version: June 3, 2001

Previous version: May 24, 2001

Introduction

This document defines an Abstract Syntax or Infoset for RDF [1]. RDF Statements are commonly refered to as triples of {predicate,subject,object}. In practice RDF Datastores may supplement the triple with associated information. This abstract syntax defines the supplemented information items.

Context

A Context is a list (ordered set) of Statements

Statement

A Statement is represented by the 6-tuple: <predicate,subject,object, context, index, asserted>

atom

An atom represents either an XML QName, according to XML Namespaces, or a quantified variable matching the EBNF:

atom := (quantvar | QName)

quantvar := ('?' Name) Name as defined in XML 1.0

predicate

A predicate is an atom.

subject

A subject is an atom.

object

An object is an atom. Note that literal strings may be represented by the data: URI scheme.

context

A context is represented by a QName

Purpose of the Context

The context allows separation of statements into lists which can be asserted as a group. Contexts allow 'coloring' of subgraphs. The document from which a set of statements arises forms a context (represented by its URI). Contexts can 'contain' other contexts.

index

The index gives the index of a statement within a context in document order. The index allows preservation of statement order as if each RDF <Description> defines a container, and folds the RDF concept of container directly into the RDF Abstract Syntax.

Purpose of the index

The index allows representation of the statements contained by a context as a list which is an ordered set of tuples. Index also allows reference to a statement within a context using the syntax "/" number, where the number refers to the particular statement. This is the syntax of the XPointer child sequence.

If we remove index, how do we represent lists? Also consider using a special statement to indicate the last statement in a list.

asserted

asserted flags whether a statement is asserted within the context. Asserted statements are the root statements of expressions formulated from more than one statement.

Purpose of asserted

An asserted statement is a root statement for expressions contained by a context.

If we remove asserted, how will we signal that the statement is the root statement of an expression? This seems to be the sole purpose of this flag.

Abstract Syntax

There exists a set called QNames
There exists a set called quantifiedVariables
There exists a set called Atoms where for all a in Atoms => (a in QNames) or (a in quantifiedVariables)
for all atom(x) <=> x in Atoms
There exists a set called Predicates where for all p in Predicates => atom(p)
There exists a set called Subjects where for all s in Subjects => atom(s)
There exists a set called Objects where for all o in Objects => atom(o)
There exists a set called Contexts where for all c in Contexts => atom(c)
There exists a set called Asserted where for all a in Asserted => boolean(a)
There exists a set of Statements := {<predicate,subject,object,context,index,asserted>}

Referring to Statements by Index

A statement may be referred to by index using a syntax consistent with the ChildSeq syntax of XPointer []. For example, the 2nd statement in a context is referred to as:

This syntax allows reference to a particular statement in a context as if the context were an <rdf:Seq> container. One can refer to a range of statements by inserting a hyphen between the start and ending index of the range, e.g.: /2-6.

Examples:

The statement:

(not (says jon (color sky blue))

is represented by the Statements:

<not,/2,true,ctx-foo:s1,1,true>
<says,jon,/3,ctx-foo:s1,2,false>
<color,sky,blue,ctx-foo:s1,3,false>

The Squish [2] query: (dc:creator, foo, ?x)(foaf:mbox, ?x, mailto:jonathan@openhealth.org)

is represented as:

<stmt,this,/2-3,ctx-query,1,false>
<dc:creator,foo,?x,ctx-query,2,false>
<foaf:mbox,?x,mailto:jonathan@openhealth.org,ctx-query,3,false>

S Expressions

An s-expression may be represented in XML:

(or a b c d)

becomes

<or>
 <a/>
 <b/>
 <c/>
 <d/>
</or>

In the case of representing an s-expression in RDF, the statement index is used to represent the list.

<rdf:type,this:or,or,ctx-stmt,1,true>
<a,this:or,true,ctx-stmt/1,2,false>
<b,this:or,true,ctx-stmt/1,3,false>
<c,this:or,true,ctx-stmt/1,4,false>
<d,this:or,true,ctx-stmt/1,5,false>