Markup Language for books XML and Semantic Web Approach Rakhi Gupta Department of Computer Science Loyola University Chicago Chicago, IL 60640 Email: [email protected]

Abstract—There is no such thing as an independent book store any more. Books, like wine, are a mass commodity exploding at a dizzying rate across the world as more and more people are driving up demand for even the most obscure titles no matter where they are published. It is possible to find a little known Brazilian author in a remote shop in Goa as it is for the newest book about Harry Potter. Book club recommendations by national figures like Oprah and Dr Phil also cause sharp spikes in demand and both publishers and sellers have to be able to deal with this information overload. Not only bookstores but other auxiliary sellers and consumers of publishing data have to have a way to tap into this network for optimum efficiency like libraries and medical and academic journals. To facilitate this exchange of information between different vendors there has to be a common medium. By sharing common structures like XML, it becomes easier and less expensive for them to exchange information with their partners and clients.

I. I NTRODUCTION I will start with creating a markup language for books using XML information throughout, from storing details about books in a database to drawing XML using XQuery from catalog or to output sorted information on prices using XSLT, XPath and XQuery. Then I will explain how the markup language would be transformed to sematic web using RDF, RDFS , OWL. It will allow data merging with other partners without losing any details that does not match and hence will ensure the logical hierarchy. It will also have the flexibility to add additional data that consumers of the information might need. I will also discuss some of the applications where Semantic Web could be used to provide intelligent access to heterogeneous, distributed information, enabling software products (agents) to mediate between user needs and the information sources available. The source code can be found at the following Google Code repository: http://code.google.com/p/booksmarkuplanguage/

II. XML AND ITS TECHNOLOGIES I started with developing a markup language for books which defines various elements and attributes of the books in the form of xml tags and then I created a document structure or information model which identifies various elements and their attributes along with the values and type of content they can hold. I used XSLT to perform various transformations using XPath to navigate the document tree structure and locate the nodes. Using XQuery, I queried and retrieved data from multiple source documents and executed certain queries using FLWOR expression which supports iteration and binding of variables to intermediate results. FLWOR expression stands for a number of keywords that are used in conjunction to perform complex queries and helps in precise restructuring of nodes. It is composed of For, Let, Where, Order and Return statements and this query is very similar to SQL [1] query. For clause sets up a loop that is executed one time for each member of the sequence; Let clause defines variables that can be used later in the query; Where clause filters the values to be processes; Order clause helps in sorting results and finally the Return clause returns the result of an overall the expression. XQuery offers a strong combination of programming power and ease of use. It also helps in improving the performance in terms of data integration and querying XML database. It is better oriented towards problems involving multiple documents and also is great at expressing joins and sorts; manipulating sequence of values and nodes. On the other hand, XSLT [2] is better oriented towards transformation problems. Both use XPath[1], [3] to navigate through the tree hierarchy and helps in locating specific nodes. XQuery is basically an extension to XPath in terms that it has the capability to merge different source documents while XPath [4] can just handle a single document. But we cannot use XQuery [5], [1] without involving XPath. So, XPath is important and makes it easy to read and locate data anywhere in the XML tree structure.

One of the limitations of XQuery is that it does not have the concept of relevance which is quite important for both the text data (text-based information retrieval) and non-text data for example, fuzzy search. So, queries like ”Find a book about information retrieval” or ”find a book which is about 30 dollars” cannot be performed using XQuery.

I created one of the basic XForms where in user is asked to enter the title, author and review comments for the book and using the repeat property of the XForms, user can fill in multiple book reviews on the same form and once the data is submitted, it generates the output in the form of an XML file with these input fields as labels. So, XForms model with the element ”model” belongs in the ”head” section of the XHTML document. Form controls and user interface components belong in the body of the XHTML document. Every form control element has a required ”label” child element which contains the label associated with it. Also, each input has a ref attribute which helps in uniquely identifying it as an XForms input elements and this output can be used for querying and transforming later.

III. XF ORMS Web forms are everywhere. They can be commonly found in search engines and e-commerce websites and they also exist in almost every web application. Though HTML have made the forms successful but they have certain limited features: they mix purpose and presentation; they run only on the web browser and even the simplest form based tasks are dependent on scripting. So with XForms, the forms are separated into two separate components: XForms model which describes the purpose, and XForms [6], [7] user interface which describes how the form is to be presented. They use XML events which are a general purpose event framework described in XML and this also helps the triggering events to be script free during the user interaction. Once the user interaction is finished, instance data is serialized as XML and sent to the server. Hence, they help in building powerful user interfaces quickly because they handle data, model and presentation separately which also allows maximum reusability of the data.

IV. S EMANTIC W EB XML defines organization of data but it does not say anything about what this data means. For example, lets take two cases where we define some elements using XML in different forms as follows: Example1:

Advantages:

Example2:

1) Multiple Environments: XForms can support multiple environments like mobile phone browsers, audio browsers etc and help in communication between different processes. The labels attached to particular fields in the markup allow other programs to identify them. 2) Internationalization and localization: There are no encoding problems in XForms as they use XML as underlying serialization form [7] and situations where different forms of languages are used; separation of presentation from content comes to rescue because nothing in XForms say how the fields should be laid out. 3) Input Validation: One can validate these forms both on the client as well as the server side as all the client data submitted can be tested against schema to make sure that it confirms to the model. We have come across many situations wherein we have been asked to fill in the email address twice just for verification purposes but XForms makes these checks a lot easier to write by making them more declarative and more direct. 4) Avoid round trips: XForms overcome one of the limitations of HTML in which every calculated change requires a round trip back to the server [7] . In XForms, adding calculation fields that depend on the content of other fields enables a more declarative approach in XML and these fields could also be used as pure output fields which only clients could see.

If we look at the above examples closely, one can see that the second example is more informative but for a machine, the information content is equivalent in both the examples. So, there is a need of something which could help machines understand the data structures we define. Solution to this is the Semantic Web which helps in describing things in a way that a computer could understand and deduce the logic from the data. Semantic web provides us with a framework which allows data to be shared and reused across various applications, enterprises and different boundaries. It provides an efficient way to represent data on the World Wide Web and it acts as a database which could be globally linked in a manner understandable by machines to the content of documents on the web. Semantic Web approach: XML is the syntactic foundation layer of Semantic Web. All the technologies providing features of the Semantic Web are built on top of the XML [8] thereby providing guarantee of interoperability at the base level. 2

Semantic Web helps in data integration by mapping various data onto an abstract data representation. And then one can merge the resulting different representations and makes it easier to query them on the whole. It [9] helps in providing metadata to the existing XML files which make them machine understandable.Various semantic technologies exist like RDF, RDFS, SPARQL (Query language for RDF) and OWL Ontology.

xmlns:dc= "http://www.purl.org/metadata/dublin-core#">

V. R ESOURCE D ESCRIPTION F RAMEWORK RDF helps in describing and interchanging metadata over internet. Its framework pairs the resources with a specific item or location on the web so that the computer could exactly know where the resource is. RDF [10] provides a standard way of interpreting XML encoded description of things or resources. Basic Elements of RDF are: 1) Resource: RDF offers a standard way of specifying data about something known as a resource and the resource could be anything like a Web Page, a part of the web page or even a whole web-site and it could even be a real world object such as a book, a dog etc. RDF identifies the resource as a URI as the name should be global and unique. For example, the following URI uniquely identifies the resource: "http://www.bml.net/books/catalog #SemanticWeb" 2) Property: Property is a resource that has a name and can be used as a property and it can even be used to describe some specific aspect, characteristic, attribute or relation of a given resource. "http://www.bml.net/books/catalog#ISBN" 3) Statement: An RDF statement is used to describe properties of resources. Property value can be a string literal or a resource here. So, we have the following format: ”subject” has a property ”predicate” whose value is ”object”. These elements represent the RDF-Triples [11] and correspond to subject, predicate and an object. The knowledge expressed in RDF- Triples helps machines do useful operations. Example of RDF document: ]>

25.25 23.50 15 Rakhi Gupta 2009-07-25 /rdf:Description> rdf:RDF defines the URI namespace of rdf document and xmlns=http://wwww.bml.net/books/Catalog# is the namespace I created for the books catalog where Catalog signifies the main Class to which books belong. rdf:Description describes the resource and rdf:about provides an absolute URI for the resource. Next we have a resource called Semantic Web which is an instance of class Books. So, the subject node Semantic Web has a rdf:type predicate indicating that they are instances of some type.


Special rules: In above example, property price does not have a resource as its value and rdf:Description tag does not have anything like rdf:ID or rdf:about. This resource is an anonymous resource and its purpose is just to provide a context for the two properties to exist. These blank nodes use rdf:nodeID attribute along with rdf:Description element instead of rdf:about or rdf:ID attribute and a statement using a blank node as its object use a property element with an rdf:nodeID attribute instead of rdf:resource.

the relationships between them. It also defines the properties and associate them with classes and hence adds semantics to predicates and resources. Some of the core classes, properties and constraints of RDFS are: • Core classes: rdfs:Resource, rdfs:Property, rdfs:Class and rdfs:datatype where resource URI is verbatim of the subject; class describes the class and its properties; datatype predefines value property • Core properties: rdfs:subClassOf, rdfs:subPropertyOf • Core constraints: rdfs:range; rdfs:domain where rdfs: domain is used to specify which class the property being defined can be used with and rdfs:range specifies the possible values of a property being declared.

25.25 Characteristics of RDF: 1) Independence- Anyone can add properties anytime because of its decentralized functionality Suppose one of my friends also has the requirement 2) Interchange- One can easily convert a RDF file into an of the same resource then one could just add a XML file. new namespace to already existing resource as 3) Scalable: RDF is basically a three part record which con”xmlns:newp=http://www.myfriend.net/books/Catalog” in sists of properties, statements and values. Each of these the above example and then add the new properties like can be used as resources and have the ability to include retailPrice and salePrice for the books using the namespace metadata created by other people and organization. reference as shown in above example using: 4) Openly extensible: RDF is openly extensible in the sense that if one gets a description from someone and want to add new personal observations to it then it is easily 23.50 possible without having to agree to a change in schema. Also, the order of properties does not matter in RDF 15.00 unlike XML. 5) Decentralized: It works in a decentralized fashion RDF does not have a data type system of its own such because unique names need no context to qualify as data type for integers, real numbers, strings, dates etc. them and hence allows anyone anywhere to write a Instead, it borrows an external data type system and used description that involves anything anywhere. rdf:datatype tag to explicitly indicate which external data type system the document is using. VI. O NTOLOGIES 25.25

Ontologies is a description of relationship between ideas. An XML DTD is also a kind of ontology but it is a restrictive and limited to relationships like parent-of, child-of and attribute-of. A complete ontology can include any arbitrary relationships like is-akind-of, if-then, or even it can model specific ideas like published-by or is-afraid-of. They are described by a markup called RDF.

Dublin Core is a vocabulary used to describe documents or web pages. I used Dublin Core prdicates in above example to describe the creator as well as the creation date of the RDF document. RDF validator is being used to validate the RDF documents which is provided by www.w3.org at www.w3.org/RDF/validator and it helped me to know if the document is in proper RDF format or not. RDF Schema

Web ontology language used in Semantic Web is OWL [13]which comprises of RDF Schema for organizing information and additional new constructs which could express much more complex and richer relationships. It adds more domain knowledge with the use of cardinality constraints, set operators, various functional properties and enumerations etc. It also describes relationships between classes and use the logic to make deductions. OWL provides a reusable domain knowledge and one could also create new

As we have XML Schema for XML; similarly we have RDF Schema for Semantic Web. RDF Schema is an extension of RDF [12] and it provides a vocabulary which can be used by RDF instance statements. It helps in defining classes and 4

classes based on the existing information quite easily.



OWL: In RDFS, the root class is rdf:resource for everything we describe whereas in OWL, the owl:Thing class is the root of all classes.

expensive

For example:

xmlns:xsd= "http://www.w3.org/2001/XMLSchema#" "http://www.w3.org/2000/01/rdf-schema#" 1 "http://www.w3.org/2002/07/owl#" xmlns:base= "http://www.bml.net/Books.owl"> "http://www.bml.net/books/Books.owl"/> rdf:resource="#hasPublisher" /> nonNegativeInteger">1 rdf:resource= "http://www.yetAnotherOne.com#SemanticWeb"/> rdf:resource="#hasTitle" /> 1 rdf:resource="#SemanticWeb"/> "http:///www.w3.org/2001/XMLSchema#String"/> "http://www.w3.org/2001/XMLSchema#String"/> rdf:resource="#owned_by"/> has booktitle rdf:resource="#Advanced"/> Title of a book, 5

express that there must be some values of this property but atleast one value of this type has be owned by one Graduate student for example and owl:hasValue helps in constraining the values.

part of which is being cited.

Also, there are certain cardinality constraints like minCardinality and maxCardinality to specify a range has publisher and owl:owned by property is used to define who owns which property. There are some richer properties which can be defined The publisher’s name. by OWL [14] like owl:ObjectProperty is used to connect a resource to another resource and owl:DataTypeProperty is Case where an used to connect a resource to an untyped value or a typed ObjectProperty might be value like a built in data type in XML schema. a better choice, where the range is some set of publisher Properties can also be symmetric, transitive, functional, names defined in another ontology. inverse-functional or it can even be an inverse of some other property. owl:disjointWith is by default a symmetric property which defines the classes that are not related in anyway. OWL ontology can be validated using a validator as well. Some of the validators available are mindswap, wonderweb etc. has ISBN VII. XML VS RDF XML lets everyone create their own tags – hidden labels The International Standard such as or that annotate Web pages or sections of text on a Book Number. page. It also allows users to add arbitrary structure to their documents but says nothing about what the structures mean. Meaning is expressed by RDF, which encodes it in sets of triples, each triple being rather like the subject, verb and object of an elementary sentence. These triples [11] can be written using XML tags [10]. In RDF, a document makes assertions that particular things like people, Web pages has price etc have properties such as ’is a sister of,’ ’is the author of’ having certain values in the form of another person or another Web page. Mostly, the data processed by machines The price of the document. follow the same structure. Subject and object are each identified by a Universal Resource Identifier (URI), just as used in a link on a Web page. The verbs are also identified by URIs, which enables anyone to define a new concept, a But there could be a situation where two different databases as ISBN. So, there should be some mechanism in which a being used to mean the same thing. To handle this scenario, Ontologies comes into picture and it formally defines the re lations among terms, classes and properties. Hence, Semantic So, OWL helps in adding more expressiveness to the defined classes.

web helps in data integration across diverse systems easily and then allow us to query the system on the whole.

owl:allValuesFrom in terms of above example helps in ensuring that Advanced level book is owned by at least one Gradaute student. owl:some-ValuesFrom could be used to

VIII. A DVANTAGES OF S EMANTIC W EB In large scale system architectures there are often many component systems which cannot talk to one another and 6

involve a lot of effort to make the systems talk to one another. RDF helps in enabling systems to understand one anothers metadata. Metadata ontologies could also be used to glue various diverse systems and data models together.

There is a need to enrich available information with machineunderstandable semantics and Semantic Web will provide intelligent access to heterogeneous, distributed information, enabling software products to mediate between user needs. Currently, the web is mainly a collection of information but does not yet provide support in processing this information, i.e., in using the computer as a computational device. Web services can be accessed and executed via the web. However, all these service descriptions are based on semi-formal natural language descriptions. In order to bring them to their full potential, it is required to combine them with semantic web technology which will provide mechanization in service identification, configuration, comparison and combination.

IX. A PPLICATIONS •

•

•

Swoogle: Swoogle is a semantic web search engine and it has already been used in so many research projects and applications. Swoogle’s Ontology Search interface allows a user to search for existing ontology documents distributed over the internet. Knowledge management: One very promising application area of semantic web technology is knowledge management. Traditional knowledge management [15] is facing new problems triggered by the web; information overload, inefficient keyword searching, heterogeneous information integration and geographically-distributed intranet problems. These problems could be tackled by using Semantic Web Technology. Web Services: We are at a stage where we need to produce active components that use the basic semantic web infrastructure to offer users intelligent services. Web services aim to support information access and e-business. Examples include UDDI, a repository for describing vendors, products and services. It uses WSDL [16] to describe its entries, and SOAP as a protocol to define how they can be accessed. None of these service description elements are based on semantic web technology currently thus requiring tremendous human effort to perform tasks like: searching for vendors, products and services; comparing and combining products etc. Semantic web-enabled services [17] can provide a higher level of service by automating many of these tasks.

R EFERENCES [1] G. Marks and M. Roantree, “Pattern based processing of xpath queries,” in IDEAS ’08: Proceedings of the 2008 international symposium on Database engineering & applications. New York, NY, USA: ACM, 2008, pp. 179–188. [2] W. W. W. Consortium, “Rdf/xml syntax specifications,” Website, 2004. [Online]. Available: http://www.w3.org/TR/rdf-syntax-grammar [3] P. Genev`es and N. Laya¨ıda, “A system for the static analysis of xpath,” ACM Trans. Inf. Syst., vol. 24, no. 4, pp. 475–502, 2006. [4] M. Benedikt and C. Koch, “Xpath leashed,” ACM Comput. Surv., vol. 41, no. 1, pp. 1–54, 2008. [5] W. W. W. Consortium, “An xml query language,” Website, 2007. [Online]. Available: http://www.w3.org/TR/xquery [6] ——, “Xforms,” Website, 2007. [Online]. Available: http://www.w3. org/TR/xforms [7] I. D. Works, “Why xforms?” website, 2006. [Online]. Available: http://www.ibm.com/developerworks/library/x-xformswhy.html [8] J. Sim´eon and P. Wadler, “The essence of xml,” in POPL ’03: Proceedings of the 30th ACM SIGPLAN-SIGACT symposium on Principles of programming languages. New York, NY, USA: ACM, 2003, pp. 1–13. [9] A. Renear, D. Dubin, and C. M. Sperberg-McQueen, “Towards a semantics for xml markup,” in DocEng ’02: Proceedings of the 2002 ACM symposium on Document engineering. New York, NY, USA: ACM, 2002, pp. 119–126. [10] B. Linse and A. Schroeder, “Beyond xml and rdf: the versatile web query language xcerpt,” in WWW ’06: Proceedings of the 15th international conference on World Wide Web. New York, NY, USA: ACM, 2006, pp. 1053–1054. [11] J. J. Carroll and P. Stickler, “Rdf triples in xml,” in WWW Alt. ’04: Proceedings of the 13th international World Wide Web conference on Alternate track papers & posters. New York, NY, USA: ACM, 2004, pp. 412–413. [12] W. W. W. Consortium, “Resource description framework,” Website, 2009. [Online]. Available: http://www.w3.org/RDF [13] I. Horrocks and P. F. Patel-Schneider, “A proposal for an owl rules language,” in WWW ’04: Proceedings of the 13th international conference on World Wide Web. New York, NY, USA: ACM, 2004, pp. 723–731. [14] W. W. W. Consortium, “Semantic web activity,” Website, 2009. [Online]. Available: http://www.w3.org/2001/sw [15] B. Antunes, N. Seco, and P. Gomes, “Knowledge management using semantic web technologies: an application in software development,” in K-CAP ’07: Proceedings of the 4th international conference on Knowledge capture. New York, NY, USA: ACM, 2007, pp. 187–188. [16] D. Oberle, S. Staab, R. Studer, and R. Volz, “Supporting application development in the semantic web,” ACM Trans. Internet Technol., vol. 5, no. 2, pp. 328–358, 2005. [17] X. Wang, T. Vitvar, M. Hauswirth, and D. Foxvog, “Building application ontologies from descriptions of semantic web services,” in WI ’07: Proceedings of the IEEE/WIC/ACM International Conference on Web Intelligence. Washington, DC, USA: IEEE Computer Society, 2007, pp. 337–343.

X. R ELATED R ESEARCH There is an on ongoing work in the fields of knowledge management and electronic commerce. There is also a study going on the semantic web-enabled web services which will help to bring the semantic web to its full potential. There is a recent study being done by SWAP [17] demonstrating that the power of Peer-to-Peer computing and the Semantic Web can actually be combined to support decentralized environments where participants can maintain individual views of the world, while sharing knowledge in ways such that administration efforts are low but knowledge sharing and finding is easy. Emergent semantics extracts shared ontologies for sets of individuals or groups of people. The study depicts that Intelligent tools will use such definitions to ensure that knowledge will be appropriately structured, so that it can be easily refound. XI. C ONCLUSIONS The easy information access based on the success of the web has made it increasingly difficult to find, present and maintain the information required by a wide variety of users. 7