XML-BASED RDF QUERY LANGUAGE (XRQL)

AND ITS IMPLEMENTATION

Norihiro Ishikawa, Takeshi Kato,

Hiromitsu Sumino

NTT DoCoMo Inc, 3-5 Hikarino-oka, Yokosuka, Kanagawa, Japan

Johan Hjelm

Ericsson Research, Torshamnsgatan 23, Kista SE-16480, Stockholm, Sweden

Kazuhiro Miyatsu

Ericsson Research Japan,1-4-14 Koraku, Bunkyo-ku Tokyo, Japan

Keywords: Mobile phone, RDF, XML, Query Language, Metadata, Semantic Web

Abstract: Resource Description Framework (RDF) is a language which represents information about resources. In

order to search RDF resource descriptions, several RDF query languages such as RQL and SquishQL

have been proposed. However, these RDF query languages do not use XML syntax and they have

limited functionality. xRQL is proposed to solve these issues by defining an XML-based RDF query

language with enhanced manipulations of the RDF metadata. xRQL is a logic language relying on a

functional approach. It consists of an operator declaration, a RDF data description and a result

description. Based on RDF graph data model, xRQL defines a graphical path expression with variables,

which is similar to GOQL for describing RDF data. It also adopts the object-oriented model for

creation, modification and deletion operations of RDF data. Users can define their favorite XML-

compliant result descriptions by themselves, which is similar to XQuery. In addition, a set of RDF

operations for RDF schema is defined to manipulate the class and property hierarchies in RDF schema.

xRQL has been implemented as an RDF query language over a native RDF database management

system. This paper also briefly describes the implementation status.

1 INTRODUCTION

The World Wide Web was originally built for human

consumption. It contains such a large volume of

information that it’s impossible to manage manually.

The use of metadata to describe the information can

achieve automatic processing. RDF (Resource

Description Framework) (Ora, 1999, Dan, 2002) is a

W3C recommendation for describing information

objects and processing metadata.

For querying RDF resource description, several

RDF query languages (G.Karvounarakis, 2002, Mike,

2002, L.Miller,2001, Flavius,2002, Joachim,2002,

Edutella, 2002) have been developed. However, in

the context of processing within an XML-based

distributed database system, and with respect to an

RDF and logic language, there are some problems to

be solved:

z Non-XML Syntax

Since the existing RDF query languages use SQL-

like syntax to describe an RDF query, a special

parser other than an XML parser is required.

SQL-like syntax has limitations in expressing a

complicated RDF query, compared with XML-

based syntax. It is also difficult to describe the

hierarchical RDF data model with a plain text

syntax language.

z Limitation of RDF Metadata

Manipulations

The existing RDF query languages only support

RDF metadata search operations. Other functions,

such as Creation, Modification and Deletion of

319

Ishikawa N., Kato T., Sumino H., Hjelm J. and Miyatsu K. (2005).

XML-BASED RDF QUERY LANGUAGE (XRQL) AND ITS IMPLEMENTATION.

In Proceedings of the First International Conference on Web Information Systems and Technologies, pages 319-326

DOI: 10.5220/0001229203190326

 SciTePress

RDF metadata should be considered to support

metadata manipulation operations, as well as the

similar functions are supported by SQL in a

relational database system.

z Lack of Result Format Declaration

The existing RDF query languages retrieve the

search result in an unordered and ambiguous way.

It takes more time for machine to understand and

process the retrieval results. It will be effective

and powerful if an expected result format is

declared in the RDF query language.

As a solution to the above problems, we have

proposed an XML-based RDF query language,

named xRQL, to manipulate RDF resource

description and RDF schemas. It relies on the RDF

graph data model that captures the RDF modeling

primitives and permits the interpretation of

superimposed resource description. xRQL adopts the

functionality of an XML query language (D.

Chamberlin, 2001), several object-oriented query

languages (ODMG,1998, Serge, 1989, Zeki, 1996,

Flavius, 2002) and a graph query language

(Jeonghee, 1999, R.Fikes, 2001). So it’s expected to

be a standard query language for manipulation of

RDF resource description stored in the distributed

RDF repositories.

Section 2 summarizes the requirements that an

RDF query language should satisfy. In Section 3, we

depict xRQL design concepts and xRQL syntax in

details, and give some typical examples. In Section

4, a native RDF management system that supports

xRQL is briefly introduced. Finally, we summarize

our contributions and show our considerations in the

future research in Section 5.

2 REQUIREMENTS OF AN RDF

QUERY LANGUAGE

We summarize the general requirements of an RDF

query language as follows.

z Machine Readability: RDF query language

syntax should be human readable and

machine understandable for processing RDF

metadata easily and efficiently between the

distributed heterogeneous metadata

repositories.

z Support for RDF Data Model: RDF query

language should support the RDF data

model (e.g. resources, properties, values),

which is similar to the object-oriented data

model. This means RDF query language

should support concepts in the object-

oriented data model, such as class

hierarchies and inheritance.

z RDF Schema Query: Besides the

manipulation functions of the distributed

heterogeneous metadata represented in RDF,

the retrieval function of schema definitions

for the vocabularies used in any given block

of RDF data is needed. Since schema

definitions are defined by using the RDF

Schema (RDFS) Specification, which

provide information about relationships

between classes and properties which might

be used in queries, RDF query language

should be able to query RDF schemas which

are implicit in any RDF data model.

z Inference: Because of the hierarchical

vocabularies defined in RDFS specification,

such as subClassof and subPropertyof, an

RDF query language should support the

operations of inferring “implicit”

relationships between classes and properties.

z RDF Metadata Manipulation: In addition

to basic search functions, RDF query

language should support the operations of

creation, modification and deletion of RDF

metadata. These operations, not yet

supported by the current query languages,

will be attractive to the clients who want to

construct their own RDF metadata

repositories based on given RDF schemas.

z Definition of Search Result Format: In

order to achieve the efficient processing of

search results, the capability of defining

search result format should be available in

RDF query language. That will help the

clients to retrieve the search results in their

favorite formats.

3 XML-BASED RDF QUERY

LANGUAGE: xRQL

xRQL is a logic language relying on a functional

approach. It consists of an operator declaration, a

RDF data description and a result description. Based

on RDF graph data model, xRQL defines a graphical

path expression with variables, in the same way as

Graphical Object-Oriented Query Language (GOQL)

(L.Sheng, 1999) for describing RDF data. It also

adopts the object-oriented model (D.Chamberlin,

2001) for creation, modification and deletion

operations of RDF data. Users can define their

favorite XML-compliant result formats by

themselves, in the same way as XQuery (ODMG,

1998). In addition, a set of RDF schema operations

is defined to manipulate the class and property

hierarchies.

WEBIST 2005 - WEB INTERFACES AND APPLICATIONS

320

3.1 Design Concept

We describe the design concepts of xRQL by using

the following example of RDF resource for the web

content. Figure 1 depicts the contents published on a

certain web site. The lower part depicts the

descriptions of two persons (r1 and r4) and two

content of music and movie on the web (r2 and r5).

The middle part depicts a schema for contents

annotation. In this schema, property create is defined

with domain “person” and range “content”. The

upper part shows the well-known RDF data model.

Properties (e.g. create, managed) serve to represent

attributes (or characteristics) of resources as well as

relationships between resources. Furthermore, both

classes and properties include hierarchical semantics.

For example, class music (or movie) is a subclass of

content while property compose (or produce) is a

subproperty of create.

RDF Data Model

rdfs:class

rdfs:domain

rdfs:range

rdfs:property

RDF Schema

string

company

create managed

compose

string

produce

technique

address

name

tel

RDF Resource

washington

John smith

12345678

compose managed

r4 r5

produce

address

name

tel

r1 r2

person content

composer music

producer movie

subclassof

subpropertyof

typeof

Figure 1: A schema of RDF resource for the web content

3.1.1 A Layered Graph Data Model

From our perspectives, RDF metadata architecture

can be divided into three levels: RDF Resource

level, RDF Schema level and original RDF Data

Model level. (Figure 2) RDF Resource level defines

labeled and directional graphs whose nodes are

called resources or literals (such as r2 in Figure 1),

and edges are called properties. RDF Schema level

essentially defines vocabularies of labels for graph

nodes and edges, such as class person and property

create. RDF Data Model level includes the well-

known RDF data model vocabularies, such as

rdfs:class and rdfs:property, which describe

vocabularies of the RDF schema. Each level can be

viewed as an instantiation of the levels above it.

More formally, those resources in RDF Resource

level are seen as the instantiations of RDF Schema

level. In the meantime schemas in RDF Schema level

are seen as the instantiations of RDF Data Model

level.

RDF Data Model

RDF Schema

RDF Resource

Figure 2: The layered data model

Then the relations between these three levels are

ensured by a standard type interpretation. In Figure

1, the r1 in RDF Resource level has a directed type

label pointing to composer in RDF Schema level

whose type is rdfs:class of RDF Data Model level.

xRQL is designed based on this 3 layered data

model.

3.1.2 RDF Resource Manipulation

According to the data model in section 3.1.1, we

defined four basic operations to manipulate RDF

resource under the constraint of RDF Schema level.

These operations include search, deletion, creation

and modification.

The xRQL queries essentially provide the means of

manipulating RDF resource with the knowledge of

the given RDF schema. For search operations, we

proposed a path expression based on RDF graph data

model. In compliance with RDF graph (i.e. nodes

and edges) and human readability, the path

expression is defined as a sequence of nodes and

edges according to the given RDF schema, similar to

GOQL, which is an extension of Object Query

Language OQL (ODMG, 1998). A typical path

expression is the following:

L = C1 {p1} C2 {p2} C3 …

where p1,…,pn are names of properties and

C1,…,Cn are names of classes defined in the

designated RDF schema. A data path is a sequence

of nodes and edges (e.g. C1, p1, C2, p2, C3, …, pn,

Cn), where the Ci’ s are classes and, for each i, there

is an edge labeled pi between Ci+1 and Ci. Implicit

relation between C1 and Ci can be expressed as

{p1}C2…Ci-1{pi-1} sequence.

Furthermore the variables can be defined in the

path expression in order to combine the RDF schema

with the associated RDF resource smoothly. For

example, to retrieve the person resource that create

certain content in Figure 1, the path expression could

be written as:

person: x {create} content

Here, variable x represents the instance of a

person class that will be retrieved. If there are

subclasses defined under person class, the resource

XML-BASED RDF QUERY LANGUAGE (XRQL) AND ITS IMPLEMENTATION

321

description belonging to these subclasses will also be

retrieved. According to RDF schema defined in

Figure 1, the instances of composer and producer

subclasses will also be returned as the instances of

the person class.

For RDF metadata manipulation (i.e. creation,

deletion and modification operations), RDF resource

descriptions are defined with object-oriented

concepts, in the same way as OQL (ODMG, 1998).

A resource that belongs to a certain class defined by

RDF schema in the RDF repository is described as

an object identified by its unique ID (i.e. URI) with

properties and the associated values. In Figure 1, r1

is an object with name, tel, address and compose

properties. The values of name, tel and address

properties are defined as string while the value of

compose points to another object r2. r1 object

belongs to composer class and r2 object belongs to

music class according to the RDF schema in the

middle part of Figure 1. The RDF metadata

manipulation is defined as the operations of objects

which are constrained by the RDF schema. Figure 3

shows an object-oriented schema converted from the

original RDF schema in Figure 1.

Person

Create:

Name: St ri ng

Tel: St ri ng

Address: St ri ng

Cont en t

managed

Composer

Compose:

Pro ducer

Produce:

Musi c

Mo vi e

SubRelationaship

ClassValue (Range)

company

Figure 3: Object-oriented Schema

A deletion operation is used to remove the RDF

resource from the RDF repository. This operation

will remove an object completely. Not only the RDF

resource itself and its properties, but also the

associated statements will be removed. For example,

“Delete the company r3 from figure 1” will remove

the “r3” resource and all of its corresponding

relationships completely. Values of property

“managed” belonging to r2 and r5 will be changed to

Null after the deletion operation.

A creation operation is defined to create a new

RDF resource into RDF repository. As shown in

Figure 3, each of the RDF resources can be

represented as an object with a set of properties and

associated values. The value of a property could be

another object or literal. RDF resource inherits

properties defined by its parent class. For example,

the composer object inherits properties, such as

name, tel and address from its parent class, namely

person class.

A modification operation is defined to change the

value of a designated property. If the value is another

object, a relationship between the object that the

property belongs to and the value object is

established. As shown in Figure 4, this operation

could also be used to merge or separate sub RDF

graphs (e.g. right pattern in Figure 4).

stri ngr2

Figure 4: Modification operation based on object-

oriented concepts

3.1.3 Syntax of RDF Query Language

According to the World Wide Web Consortium, the

Extensible Markup language (XML) (T.Bray, 1998)

is now spreading out as the standard language for

machine understanding. Since RDF adopts the XML

syntax to describe its data model, XML parser

becomes a necessary tool for automatic machine

processing of RDF resource description. In this

approach, the use of XML syntax-based RDF query

language which is consistent with RDF syntax will

greatly lighten processing of the query analysis,

compared with those queries defined in non-XML

syntax, which often requires a specific parser.

3.1.4 Result Format Definition

The existing RDF query languages have no

consideration of the result format definition in their

syntaxes. For automatic processing of the retrieve

result, users have to develop a tool to convert the

result into machine-readable format. We think it is a

bottleneck in the wide spread of RDF technologies.

xRQL attempts to provide “format-able” result by

providing an element for the users to flexibly define

their favorite result format based on XML syntax,

similar to return in XQuery. For example, in Figure

1 RDF schema, assuming that we have the variable x

that has been defined in path expression:

person:x{create}content retrieve with <personinfo>,

the result format is as follows:

WEBIST 2005 - WEB INTERFACES AND APPLICATIONS

322

</result>

A set of person instances will be returned with tag

<personinfo>. Using XSLT, this result could be

easily converted into xHTML for displaying or

further processing.

3.2 xRQL Syntax Definition

xRQL is a logic language relying on a functional

approach. Its syntax is defined by XML syntax.

xRQL syntax consists of three functional parts:

operator declaration, RDF data description and result

format description. The first part declares user’s

request operations. The second part defines the RDF

resource or schemas using object-oriented concepts,

path expression and condition declaration. The last

part depicts user’s favorite XML-based format of the

result.

3.2.1 Operator Declaration

An operation declaration element is defined to

declare an xRQL operation. We define xRQL

operations for creation, search, modification and

deletion of RDF metadata. RDF resource could be

queried and modified with these four operations,

while only search operation is provided for RDF

schema.

3.2.2 RDF data description

RDF data description includes object element,

objectvalue element, location element, condition

element and namespace element.

z Object Element and ObjectValue Element

According to the consideration in section 3.1.2,

we defined two elements to describe RDF

resource under the object-oriented concepts.

Object element is used to declare RDF resource

class based on RDF schema. ObjectValue

Element is used to assign unique ID and property

values to RDF resource which belongs to the RDF

resource class declared in object element.

z Location Element

A location element is defined to represent path

expression, in which a sequence of classes and

properties of RDF schema are used to describe

RDF graph data model, described in section 3.1.2.

z Condition Element

A condition element is used to specify conditions

for search, modification and deletion operations

to RDF resource designated by the path

expression in a location element.

z Namespace Element

A namespace element is defined to support

multiple RDF schemas. The abbreviation of a

namespace can be defined by a namespace

element. For cases where the same class or

property names are used in different schemas, the

namespace element can be used to explicitly

resolve such naming conflicts.

3.2.3 Result Format Description

In the search operation, a result format description

element is provided to define the result format. It is

based on XML syntax so that users can get a RDF

query result in their favorite XML format for further

processing.

3.3 RDF Schema Operations

Besides the operations for RDF resource description

mentioned above, xRQL also supports RDF schema

search operations.

3.3.1 Schema Variable and Path Expression

To distinguish the schema variables in the path

expression, we proposed schema variables prefixed

by “$”. A schema path expression in which schema

variables are depicted independently is expressed as:

<location>$x {managed} $y </location>

According to the layered data model mentioned in

section 3.1.1, the RDF Schema level could be

viewed as a instance of the RDF Data Model level.

Hence, this schema path expression could be

interpreted that $x belonging to rdfs:Class should be

domain of managed property, while $y belonging to

rdfs:Class should be range of managed property.

3.3.2 Function Definition

xRQL offers the functions of retrieving relationships

of classes and properties from RDF schema, such as

subclassof(), subpropertyof(), domain(), range(),

typeof(), etc.

XML-BASED RDF QUERY LANGUAGE (XRQL) AND ITS IMPLEMENTATION

323

3.4 Examples of xRQL Operations

In this section, we depict the details of xRQL

operations using typical examples.

Figure 5 shows a search operation for Retrieving

persons whose works are managed in companies

<xrql:xrql>

<xrql:operator>select</ xrql:operator>

<xrql:location>

person:x{create}content{manage}company

</xrql:location>

< xrql:result>

</xrql:result>

<xrql:usingnamespace>

<xrql:item>

“ns1=www.docomo.co.jp/DE/metaschema”

</xrql:item>

</xrql:usingnamespace >

</xrql:xrql>

Figure 5: Example of search operation

In the location element, we use a data path

expression with three class names (i.e. person,

content, company) and two property names (i.e.

create, managed). The data variable x represents the

instances of the class person. Condition Element

isn’t used in this example. The Result Element

designates the output format. Figure 6 shows the

result of this search operation.

<xrql:result>

<description>”www.person1.co.jp”</description>

<description>”www.person2.co.jp”</description>

<description>”www.person3.co.jp”</description>

</xrql:result>

Figure 6: Result of Figure 5 search operation

By using several path expressions and condition

requirements, more complicated queries for

navigating through RDF repository are possible.

Figure 7 shows a search operation to Retrieve

person and their names and telephone number

whose first name is “John Smith”.

In this query, we use two data path expressions.

Three variables are defined in these two path

expressions. Because either path expression starts

from same class person, the path expressions

implicit that both properties belong to the same class

person. A Condition Element is defined in this

example to specify a condition for data matching.

The result format is also defined to retrieve three

variables using different tags.

<xrql:xrql>

<xrql:operator>select</ xrql:operator>

<xrql:location>

person:x{name}STRING:y,

person{tel}STRING:z

</xrql:location>

<xrql:condition>

<xrql:item>y=”john smith” </xrql:item>

</xrql:condition >

< xrql:result>

</xrql:result>

<xrql:usingnamespace>

<xrql:item>

“ns1=www.docomo.co.jp/DE/metaschema”

</xrql:item>

</xrql:usingnamespace >

</xrql:xrql>

Figure 7: Example of complex search operation

4 IMPLEMENTATION STATUS

We have implemented the xRQL processor over a

native RDF management system. Unlike other RDF

management implementations which usually manage

RDF data by using relational databases, this system

manages RDF metadata in a text file format. Figure 8

shows the system overview. The native RDF

management system consists of an RDF parser,

public API, Graph Match Process, Result Output

Process and the Persistent Storage. We implement

xRQL system by using the public APIs.

WEBIST 2005 - WEB INTERFACES AND APPLICATIONS

324

Public API

Graph Match

Process

RDF Parser

Persistent Storage

User Control Interface (API)

Graphic User Interface (GUI)

xRQL Parser

xRQL Result

Format Transformation

Result

Output

Process

Wrapped API

In Memory Storage

(Triple Structure)

RDF Repository

(Data and Schema)

Figure 8: Overview of the native RDF Management

System

4.1 User Control Interface

User Control Interface provides a friendly GUI for

the users with convenient tools, such as query

creation assistant and data input and output function.

xRQL parser module and xRQL Result Format

Transformation module work with other modules

shown in Figure 8 to evaluate xRQL query,

controlled by the User Control Interface.

4.2 xRQL Parser

xRQL parser analyzes an xRQL query string coming

from User Control Interface and generates a query

object. This query object will be delivered to Graph

Match Process module to be evaluated. This parser

is composed of the following three components.

z Preprocessor

The preprocessor component converts the input

query string into standard format that could be

processed by lexical and syntax analyzer.

z Lexical Analyzer

A simple XML parser (SAX) (SAX, 2002) is

employed to perform the main parts of lexical

analysis. Because some expressions defined in

xRQL BNF are not completely encapsulated with

XML tags, a small parser developed by JavaCC is

used to parse these data.

z Syntax Analyzer

The syntax analyzer will extract tokens from

Token Queue one by one. A corresponding query

object will be generated once a syntax unit is

matched. Finally, all query objects corresponding

to sub syntax unit will be composed into a query

object which represents the whole query.

4.3 xRQL Result Format

Transformation

The result format information is analyzed by syntax

analyzer. A result format object is created to

preserve the information; the object is also integrated

into the final query object. This transformation

module extracts the result format object from the

query object and uses it to render the query result

received from Graph Match Process.

5 CONCLUSION AND FUTURE

WORKS

In this paper, we propose an XML-based RDF query

language to manipulate RDF schema and resource.

Benefiting from the XML syntax, xRQL provides a

machine-understand interface to the users in the

XML-based distributed environment. Based on

object-oriented concepts, manipulations of search,

deletion, creation and modification of RDF resource

are more pithy and effective. The result format

definition also makes the result processing more

flexible. xRQL has been implemented over a native

RDF management system. Other functions, such as

Aggregation and Sort functions will be considered in

the future. Furthermore, the support of ontology

inference (e.g. DAML+OIL) in xRQL is also an

interesting topic in the future.

REFERENCES

Ora Lassila and Ralph R. Swick, 1999. "Resource

Description Framework (RDF) Model and Syntax

Specification", W3C Recommendation.

Dan Brickley and R.V. Guha, 2002. " RDF Vocabulary

Description Language 1.0: RDF Schema", W3C

Working Draft.

G. Karvounarakis, S. Alexaki, V. Christophides, D.

Plexousakis, Michel Scholl. 2002. “RQL: A

Declarative Query Language for RDF”, WWW2002,

Honolulu, Hawaii, USA. ACM 1-58113-449-

5/02/0005.

Mike Olson, Uche Ogbuji. 2002. “RDF Query and

Inference: Versa”, Available at

http://rdfinference.org/versa.doc?xslt=db.xslt

L. Miller. 2001 “RDF Query using SquishQL”, Available

at http://swordfish.rdfweb.org/rdfquery/.

Lee Jonas, Stefan Kokkelink. “RDF Path Requirements”,

Available at: http://zoe.mathematik.uni-

osnabrueck.de/RDFPath/req/

XML-BASED RDF QUERY LANGUAGE (XRQL) AND ITS IMPLEMENTATION

325

ODMG. 1998. “Object Query Language (OQL). User

Manual”, Release 5.0.

D. Chamberlin, D. Florescu, J. Robie, J. Simeon, and M.

Stefanescu. 2001. “XQuery: A Query Language for

XML”, Working draft, World Wide Web Consortium.

Available at http://www.w3.org/TR/xquery/.

S. Abiteboul, D. Quass, J. McHugh, J. Widom, and J.

Wiener. 1997. “The Lorel Query Language for

Semistructured Data”, International Journal on Digital

Libraries, 1(1):68{88.

Jeen Broekstra, Arjohn Kampman. (Sesame), 1999.

“Query Language Definition”, EU-IST Project IST-

1999-10132 On-To-Knowledge.

Serge Abiteboul. 1989. “Towards a deductive object-

oriented database language”, the First International

Conference on Deductive and Object-Oriented

Databases,pages 453-472. North Holland, Amsterdam.

Simon Brown. 1999. “The semantic of object-oriented

databases”, The University of Sheffield

Department of Computer Science.

Daniel Kim Chung Chan. 1993. “Object-oriented query

language design and processing”, Doctor thesis.

UNIVERSITY OF WISCONSIN-MADISON.

Zeki O. Bayram, Barrett R. Bryant, Faik Hakan Bilgen.

1996. “A Deductive Declarative Object-Oriented Data

Model and Query Language based on Narrowing”,

Proceedings of ISCIS XI, the 11th International

Symposium on Computer and Information Sciences,

Antalya, Turkey.

Flavius Frasincar Geert-Jan Houben Richard Vdovjak

Peter Barna. 2002. “RAL: an Algebra for Querying

RDF”, The Third International Conference on Web

Information Systems Engineering(WISE'00) , p. 173.

Singapore.

Joachim Peer. 2002. “A Logic Programming Approach to

RDF Document And Query Transformation”,

Workshop on Knowledge Transformation for the

Semantic Web at the 15th European Conference on

Artificial Intelligence. Lyon, France.

Nitish Manocha, Diane J. Cook, and Lawrence B. Holder.

2002. “Structural Web Search Using a Graph-Based

Discovery System”, FLAIRS Conference: P133-137.

Jeonghee Kim, Taisook Han Kyu, Young Whang. 1999.

“Visualization of Path Expressions in a Visual Object-

Oriented Database Query Language”, 6th International

Conference on Database Systems for Advanced

Applications (DASFAA '99) Hsinchu,

Taiwan, p. 99.

Martin Erwig, Ralf Hartmut Güting. 1994. “Explicit

Graphs in a Functional Model for Spatial Databases”,

IEEE Transactions on Knowledge and Data

Engineering, Volume 6. Number 5, P.787-804.

R. Fikes. 2001. “DAML+OIL query language proposal”.

Available at http://www.daml.org/listarchive/joint-

committee/0572.html.

Edutella, 2002. “Edutella RDF Query Exchange

Language”, Available at: http://www.kbs.uni-

hannover.de/Diverses/edutella-

archive/discussion/pdf00004.pdf.

T. Bray, J. Paoli, and C.M. Sperberg-McQueen. 1998.

“Extensible markup language (XML) 1.0”,W3C

Recommendation. Available at

http://www.w3.org/TR/REC-xml/.

L. Sheng, Z. M. Özsoyoglu, G. Özsoyoglu. 1999. “A

Graph Query Language and its Query processing”,

IEEE ICDE Conf., Australia.

SAX, 2002 Available at: http://www.saxproject.org/

WEBIST 2005 - WEB INTERFACES AND APPLICATIONS

326