Thierry Technologies

pierre@nothos.net

binary This specification describes a BULK serialization of RDF ontologies and graphs.

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119. Literal numerical values are provided in decimal or hexadecimal as appropriate. Hexadecimal literals are prefixed with 0x to distinguish them from decimal literals. BULK bytes sequences and expressions are described with the same conventions than used in the BULK 1.0 specification

The RDF namespace is an official namespace identified by the UUID (BULK, "http://www.w3.org/1999/02/22-rdf-syntax-ns#"). Its first purpose is to serialize both RDF ontologies and graphs, but as RDF is deeply rooted in the World Wide Web and the REST architecture, some of the names and forms defined in this specification are intended to be used in applications related to those. This specification defines the type URIString, which is an array whose content is a string conforming to the "URI-reference" production in (section 4.1).

0x1 (mnemonic: uriref ) ( uriref URIString ) This form makes it possible to differentiate a string containing a URI from a URI reference itself. Type: URIRef

0x2 (mnemonic: base ) ( base URIString ) This form changes the current base URI for URI resolution. If the given URI is relative, it must be resolved with the current base URI.

0x3 (mnemonic: prefix ) ( prefix {uri} ) EagerFunction EagerFunction a function that returns a URIRef whose content is the concatenation of {uri} and its only argument (or returns {uri} if called without arguments) {uri} can be of type URIString or URIRef. If it is of type URIRef, the above description should be taken as if the form is of shape ( prefix ( uriref {uri} ) ).

0x4 (mnemonic: blank ) ( blank ) EagerFunction Expr a new blank node. If a same blank node must be used in several forms, it should be made the value of a name.

Literals are arbitrary data that can be an object in an RDF triple. It is important to note that their BULK serialization doesn't need to denote the string of their lexical form. Any array used as object in an RDF triple has the semantics of a plain literal without language tag.

0x5 (mnemonic: plain ) ( plain {lang}:Array {literal}:Array ) This is a plain RDF literal. {lang} is the language tag.

0x6 (mnemonic: type ) ( type {id}:URIRef {literal} ) This is a typed RDF literal. {id} is the datatype URI. Each type defines what is legal as {literal}.

0x7 (mnemonic: xmlliteral ) ( uriref "http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral" ) A literal with this type SHOULD contain a BULK serialization of XML (cf. ). An application MUST NOT use an array for this literal for anyhting else than a string that belongs to the lexical space defined for .

0x8 (mnemonic: triples ) ( triples {triples} ) {triples} must be a sequence of ( {subj}:URIRef {pred}:URIRef {obj} ).

0x9 (mnemonic: turtle ) ( turtle {statements} ) {statements} MUST be a sequnce of ( {subj}:URIRef {preds} ), where {preds} MUST be a sequence of ( {pred}:URIRef {objs} ), where {objs} is a sequence.

The following names are associated with uriref forms, with a base URI of . Their mnemonic is the lowercased URI reference. type Property Statement subject predicate object Bag Seq Alt value List nil first rest PlainLiteral

0x20 (mnemonic: this-resource ) this-resource This expression is intended to help embedding RDF metadata alongside or within a resource. It has the semantics of a blank node that designates this resource. What makes a resource the current described resource is application-dependent.

Harvard University

sob@harvard.edu

W3C/MIT Day Software Adobe Systems Thierry Technologies

pierre@nothos.net

Nine by Nine

GK@NineByNine.org

Hewlett Packard Labs

jjc@hpl.hp.com

Thierry Technologies

pierre@nothos.net

Here is the text notation of a BULK stream defining references for each element in the RDF Schema vocabulary. The whole stream is 444 bytes in length. ( bulk:version 1 0 ) ( bulk:ns 32 rdf ) ( bulk:ns* 33 rdfs ) ( define rdfs:schema ( prefix "http://www.w3.org/2000/01/rdf-schema#" ) ) ( define rdfs:resource ( rdfs:schema "Resource" ) ) ( define rdfs:class ( rdfs:schema "Class" ) ) ( define rdfs:subclassof ( rdfs:schema "subClassOf" ) ) ( define rdfs:subpropertyof ( rdfs:schema "subPropertyOf" ) ) ( define rdfs:comment ( rdfs:schema "comment" ) ) ( define rdfs:label ( rdfs:schema "label" ) ) ( define rdfs:domain ( rdfs:schema "domain" ) ) ( define rdfs:range ( rdfs:schema "range" ) ) ( define rdfs:seealso ( rdfs:schema "seeAlso" ) ) ( define rdfs:isdefinedby ( rdfs:schema "isDefinedBy" ) ) ( define rdfs:literal ( rdfs:schema "Literal" ) ) ( define rdfs:container ( rdfs:schema "Container" ) ) ( define rdfs:containermembershipproperty ( rdfs:schema "ContainerMembershipProperty" ) ) ( define rdfs:member ( rdfs:schema "member" ) ) ( define rdfs:datatype ( rdfs:schema "Datatype" ) ) ( define rdfs:schema* ( rdfs:schema ) )

Here is the text notation of a BULK stream containg a small part of the RDF Schema ontology. The whole stream is 375 bytes in length (the corresponding RDF/XML document would be 1525 bytes in length). ( bulk:version 1 0 ) ( bulk:ns 32 rdf ) ( bulk:ns 33 rdfs ) ( turtle ( rdfs:resource ( rdf:type rdfs:class ) ( rdfs:isdefinedby rdfs:schema* ) ( rdfs:label "Resource" ) ( rdfs:comment "The class resource, everything." ) ) ( rdfs:class ( rdf:type rdfs:class ) ( rdfs:isdefinedby rdfs:schema* ) ( rdfs:label "Class" ) ( rdfs:comment "The class of classes." ) ( rdfs:subclassof rdfs:resource ) ) ( rdfs:subclassof ( rdf:type rdf:property ) ( rdfs:isdefinedby rdfs:schema* ) ( rdfs:label "subClassOf" ) ( rdfs:comment "The subject is a subclass of a class." ) ( rdfs:range rdfs:class ) ( rdfs:domain rdfs:class ) ) ( rdfs:subpropertyof ( rdf:type rdf:property ) ( rdfs:isdefinedby rdfs:schema* ) ( rdfs:label "subPropertyOf" ) ( rdfs:comment "The subject is a subproperty of a property." ) ( rdfs:range rdf:property ) ( rdfs:domain rdf:property ) ) )