An Ontology for Portability and Interoperability Digital Documents

An Approach in Document Engineering using Ontologies

Erika Guetti Suca and Flávio Soares Corrêa da Silva

Institute of Mathematics and Statistics, University of São Paulo, São Paulo, Brazil

Keywords:

Document Engineering, Document Interoperability, Document Portability, OOXML (Ofﬁce Open XML),

ODF (Open Document Format), Ontologies.

Abstract:

Organizations need to exchange information simple and efﬁcient, with costs as low as possible. Such infor-

mation is ussually presented as documents with pre-deﬁned content. These documents may be equivalent or

almost equivalent but quite distinct in different organizations. The same document can be different depending

on the historical context. Also, organizations do not always use the same technology to generate your docu-

ments. The purpose of this work is to enable interoperability of documents and achieve portability of digital

documents through the reuse of content and format in different plausible combinations. We propose the char-

acterization of digital documents using ontologies as a solution to the problem of lack of interoperability in

the implementations of document formats. As proof of concept we consider the portability between OOXML

and ODF document formats.

1 INTRODUCTION

Governments are interested in the development of

policies, processes, standards in Information and

Communication Technology (ICT), mounting struc-

tures dedicated to achieving interoperability. Mainly

in digital preservation, the challenge of interoperabil-

ity in addition to a technical is a social and insti-

tutional problem, as it depends on institutions that

pass through changes of direction, mission, admin-

istration and funding sources. Therefore, the Brazil-

ian government established a set of interoperability

standards called e-PING (Padrões de Interoperabili-

dade de Governo Eletrônico)

. Concerning the way

storing information, E-PING adopts document for-

mat Open Document Format(ODF) to transmit gov-

ernment information among public and private sec-

tors maintaining privacy and security. Digital doc-

uments are considered ofﬁcial records and are man-

aged according to laws and standards that understand

entire lifecycle of these materials(Bretas and do So-

corro Ferreira Mesquita, 2010).

Choosing a common standard format for exchang-

ing information is excellent, but still continues depen-

dent on a speciﬁc format, even being free. The main

problem with digital documents is to ensure access to

those documents in the long term. So it is necessary

http://www.governoeletronico.gov.br/

to overcome technical barriers associated with docu-

ment formats. The content, structure and context of

documents must be associated with software features

that preserves its representations and relationships en-

abling their reconstruction. The purpose of this work

is to facilitate the distribution of documents, over-

coming the problem of formats with which they were

created. Besides, we aim to enable documents inter-

operability and achieve documents portability simply

through the reuse of content and formats in different

plausible combinations.

Ontologies can assist us in this work. They pro-

vide a shared understanding of terms allowing inter-

operability and means for an intelligent integration

of information(Uschold, 1998). This work uses on-

tologies as a solution to the lack of problem of inter-

operability in implementations of document formats.

The proposal is to represent digital documents based

on two ontologies: (1) format ontology, that charac-

terizes the digital documents structure and presenta-

tion, independently of speciﬁc encoding of each soft-

ware product and (2) context ontology, that repre-

sents the information contexts of businesses. Figure

1 shows our proposal. Documents are offered based

on generic representations centralized, through me-

diators among ontologies, presentation systems and

document editing. As proof of concept will be consid-

ered interoperability between document formats ODF

and OOXML.

373

Guetti Suca E. and Soares Corrêa da Silva F..

An Ontology for Portability and Interoperability Digital Documents - An Approach in Document Engineering using Ontologies.

DOI: 10.5220/0004547503730380

In Proceedings of the International Conference on Knowledge Discovery and Information Retrieval and the International Conference on Knowledge

Management and Information Sharing (KMIS-2013), pages 373-380

ISBN: 978-989-8565-75-4

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

Renderer

Text

…

Document Ontology

ODF Translator OOXML Translator PDF Translator

Format

Ontology

Context

Ontology

Figure 1: Document generation from format ontology and

content ontology.

The paper is organized as follows: in section 2

we explain the problem of preserving digital docu-

ments; in section 3 we talk about document formats

based XML. in section 4 we introduce the fundamen-

tal concepts of document engineering; In section 5 we

summarize some related work, and in section 6 we ex-

plain our proposed method. Then, in the section 7, we

present the results. Finally, in section 8 we point out

some closing remarks and future works.

2 PRESERVATION OF DIGITAL

DOCUMENTS

Currently doing business by document exchange is

natural and intuitive. Documents are interfaces for

people and business processes. However, there is a

lot of ﬁle are formats incompatible. Many documents

written and stored with the same format can be un-

readable, inoperable or after some time be necessary

to migrate the data. Therefore organizations need to

manage their knowledge effectively in order to pre-

serve their intellectual capital. Organizations need to

provide documents independently of the software cre-

ated. Hence it, is important to enable documents inter-

operability within the business processes, i.e., to en-

able coherent exchange of information based on the

adoption of rules and communication standards that

allows comunication between heterogeneous systems.

When documents are exchanged, it is very impor-

tant to ensure documents authenticity. A document is

authentic if it can prove that there is a set of proper-

ties, considered signiﬁcants, that were correctly pre-

served along time. To achieve authenticity is funda-

mental to rightly record the provenance of the doc-

ument. Contextualize their existence, describe their

custodial history and testify to their integrity was not

compromised (Ferreira, 2006).

Preserving digital information is sometimes in de-

liberately modify or transform the digital document

that carries the message. For this transformation does

not produce a message disproportionately degraded,

it is essential to deﬁne what are the properties of mes-

sage should be ensured during the proceso transfor-

mation. In summary, to enable document interoper-

ability would mean not lose the data that explain the

semantics of the document. While document portabil-

ity would mean not losing the characteristics of for-

mat settings document presentation (Ferreira, 2006).

3 DOCUMENT FORMATS BASED

XML

Document format encapsulates a complete descrip-

tion for storing digital documents. For instance,

organizing elements such as text, fonts, graphics,

and other information needed to display a document.

Among the frequently used document formats are

Ofﬁce Open XML (OOXML) and Open Document

Format (ODF). They are main open standards based

XML for document formats. However, there are no

implementations that offer 100% of portability for

both, not even into the dominant implementations Mi-

crosoft Ofﬁce and Open Ofﬁce.org (Shah and Kesan,

2009). For example, to present some differences be-

tween them, Figure 2 shows a simple text encoded in

standards OOXML and ODF.

4 DOCUMENT ENGINEERING

Documents are purposeful representations and orga-

nizations of information, but they exhibit great vari-

ety. Document engineering analyse and design meth-

ods that yield precise speciﬁcations for the informa-

tion and rules that business processes require. This

mean, developing models that emphasize document

requirements and patterns of information exchange,

focusing the separation of content and presentation of

the document information in an inherent and desirable

way. A document must be represented using a shared

common conceptual model (Glushko and McGrath,

2008).

Document engineering deﬁne models of different

types of documents in a rigorous and unambiguous

way so that we can automate their process or ex-

change within or between applications. It needs to

be diligent and precise when deﬁnes the meaning of

any information produced and consumed by business

KMIS2013-InternationalConferenceonKnowledgeManagementandInformationSharing

374

Figure 2: Sample text encoded in formats OOXML and ODF.

domain

range

domain

ObjectProperty

has

range

ObjectProperty

has

range

ObjectProperty

has

range

ObjectProperty

has

range

domain

ObjectProperty

has

ObjectProperty

contains

domain

rangedomain

domain

range

ObjectProperty

isComposedOf

ObjectProperty

displayContent

ObjectProperty

displayContent

ObjectProperty

displayContent

isComposedOf

ObjectProperty

domain

range

SubClassOf

domain

ObjectProperty

has

ObjectProperty

has

ObjectProperty

has

ObjectProperty

has

range

domain

range

domain

range

SubClassOf

domain

range

domain

Class

Body

Class

Document

Class

Meta

Class

Paragraph

Class

Cell

range

Class

Graphics

Class

Text

Class

Font

Class

Table

Class

StyleGraphics

Class

StyleParagraph

Class

StyleText

Class

Style

Class

StyleTable

Class

StyleTableColumn

Class

StyleCell

Figure 3: Format Ontology.

applications. As electronic documents are ubiqui-

tous, document engineering emphasize that a docu-

ment must be deﬁned in a technology-neutral way as

a purposeful and self-contained collection of infor-

mation. When businesses exchange documents, they

must agree on what the documents mean and on the

business processes they expect each other to carry out

with them, but they do not need to agree on the tech-

nology they use (Glushko and McGrath, 2008).

Many applications need to support different phys-

ical interfaces. These imply many-to-many mappings

between the input and output interfaces for each ap-

plication. Many-to-many mappings can be avoided

by mapping all physical interfaces to a common con-

ceptual model. So, the best way to facilitate inter-

operability is allowing the participants to share the

same conceptual model. A common metamodel helps

aligning different models. Basing the user and ap-

plication interfaces on a common conceptual model

ensures that the documents they process are interop-

erable (Glushko and McGrath, 2008).

Document engineering entails the need for stan-

AnOntologyforPortabilityandInteroperabilityDigitalDocuments-AnApproachinDocumentEngineeringusing

Ontologies

375

range

domain

ObjectProperty

study

range

ObjectProperty

belongs

range

ObjectProperty

make

ObjectProperty

isEnrolledIn

ObjectProperty

make

domain

range

ObjectProperty

hasAdvisor

domain

range

ObjectProperty

work

ObjectProperty

has

domain

range

ObjectProperty

has

domain

range

ObjectProperty

participates

domain

Class

Subject

Class

Student

Class

Curse

Class

Institute

Class

Qualification

Class

Defense

Class

Committee

Class

Professor

ObjectProperty

formedIn

domain

range

Figure 4: Context Ontology for Student Sheet.

dardization of syntax, structure, and semantics of

business documents and their reusable components.

Document engineering search achieve documents in-

teroperability and improve documents portability. In

other words, document interoperability is the abil-

ity of businesses applications to extract information

contained in various kinds of documents and trans-

form it standardized XML structures. These XML

data ﬁles can then be exchanged between the various

systems and further processed (Schmidt et al., 2006).

Using standardized XML structures saves effort and

yields more consistent, compatible, and successful

designs. Moreover, document portability refers to the

exchangeability of documents as a whole, i.e. with

all the information they contain, formatting settings

and graphic information. Crucially, all stylistic and

graphical data (Schmidt et al., 2006). This infor-

mation could be grouped in four components: con-

tent, logical structure, layout structure and presenta-

tion(Barron, 1996).

Document portability considers the issue of visual

ﬁdelity to an original, makes requirements in terms of

optical appearance, stylistic elements and other such

matters. On the other hand, document interoperabil-

ity is exclusively concerned with the exchange of the

business information contained in documents. Docu-

ment interoperability could enable businesses appli-

cations to communicate directly with a wide range

of different eGovernment services, platforms and ad-

ministration applications. The document interoper-

ability shall enable business processes to be generated

from ofﬁce applications, and then to be integrated in

corresponding eGovernment processes(Schmidt et al.,

2006).

5 RELATED WORKS

An excellent work existent in the literature is based

on the framework UN/CEFACT CCTS

(Core Com-

ponent Technical Speciﬁcation). CCTS is an euro-

pean conceptual framework for modeling document

components in a syntax neutral and technology inde-

pendent manner. It permits handling different doc-

ument conﬁgurations imposed by divergent national

legislations. The conceptual model is transferred to

XML schema serving as a basis for Collaborative Web

Services for eGovernment(Vogel et al., 2008).

Another interesting work is designing XML doc-

uments from conceptual schemas and workload in-

formation for compliant to consensual information of

speciﬁc domains. The research presents a conver-

sion approach which considers data and query work-

load estimated for XML applications, in order to gen-

erate an XML schema from a conceptual schema.

Loaded information is used to produce XML schemas

which can respond well to the main queries of an

XML application. The work evaluates an approach

through a case study carried out on a native XML

database. Its experimental results demonstrate that

the XML schemas generated by the proposed method-

ology contribute to a better query performance than

related approaches(Schroeder and Mello, 2009).

Lastly, Universal Business Language (UBL) is a

library of standard electronic XML business docu-

ments such as purchase orders and invoices. It was

developed by OASIS. UBL is designed to provide

a universally understood and recognized commer-

cial syntax for legally binding business documents

and to operate within a standard business frame-

http://www.unece.org/cefact/index.html

KMIS2013-InternationalConferenceonKnowledgeManagementandInformationSharing

376

work such as ISO 15000 (ebXML) to provide a com-

plete, standards-based infrastructure that can extend

the beneﬁts of existing EDI systems to businesses of

all sizes. UBL Library is based on a conceptual model

of information components known as Business Infor-

mation Entities (BIEs). These components are assem-

bled into speciﬁc document models. One document is

a set of information components that are interchanged

as part of a business transaction; for example, in plac-

ing an order. This approach facilitates the creation of

UBL-based document types beyond those speciﬁed in

this release(Bosak et al., 2011).

Our proposal uses ontologies implemented in

OWL (Web Ontology Language). Advantages using

ontologies compared to previous models, more ro-

bust and worked our proposal, we could say: OWL

is a standard semantic markup language for publish-

ing and sharing ontologies on the World Wide Web

and the Semantic Web. We have freedom to reuse the

content ontology in other services together in a ser-

vice document generation, i.e., because it is indepen-

dent of format ontology, and ﬁnally we can harness

the power of ontologies to infer new knowledge do-

main.

6 USING ONTOLOGIES FOR

MODELING DIGITAL

DOCUMENTS

Ontologies are designed for enabling knowledge shar-

ing and reuse on some domain that can be commu-

nicated between people and computers. Therefore,

to enable the sharing and reuse of knowledge, it is

necessary a formal speciﬁcation of concepts. Ontolo-

gies deﬁne rules of relationships between concepts to

query, infer knowledge(Gruber, 1995).

The documents present information depending on

its purpose, and this information can be presented in

different ways. Our proposal is to build a model that

considers essential qualities of digital document. Fol-

lowing the approach of engineering documents, a doc-

ument is considered as combination of information

components and presentation components. Whereas

the information is independent of how it is presented.

This model is based on a integration of two ontolo-

gies. An ontology that represents the presentation

structure and an other that represents the information

according business context, i.e. , a format ontology

and a context ontology. These ontologies are indepen-

dent between each other. The objetive of format on-

tology is to achieve simple document portability. The

purpose of context ontology is to achieve document

interoperability. The mapping between the format on-

tology and content ontology for its physical interfaces

occurs through translators.

6.1 Format Ontology

Format ontology characterizes visual structure of doc-

ument, i.e., formatting settings and graphic infor-

mation. Format ontology speciﬁes formally docu-

ment components, i.e., presentation structure includ-

ing metadata, paragraphs, texts, tables, lists, enumer-

ations, images, styles, etc. The Metadata is the in-

formation associated with the document, for exam-

ple: creation date, last modiﬁed date, text language,

document author, pages number, etc. The document’s

layout is based on tables. The tables are composed of

cells, cells can contain paragraphs with images, text

ou maybe another table. Each paragraph of text has

a presentation style, i.e., color, color-font, font, size,

horizontal alignment, vertical alignment, etc. The for-

mat ontology is shown in Figure 3. From the format

ontology, a document can be created in an appropriate

format for its purpose.

6.2 Context Ontology

A business context is a scope in which a special-

ized vocabulary is employed, so the business con-

text deﬁnes the type of document information. The

context is used to organize and analyze requirements

and rules information presentation. For example, a

student sheet, a medical record, rental contract, etc.

show different scenarios. For proof of concept this

work takes the context of a student sheet. The main

objetive of a student sheet is to provide academic in-

formation, i.e., grades, attendance, subjects, advisor,

date of birth, date of admission, etc. The context on-

tology was created based on that context. Context on-

tology is shown in Figure 4.

6.3 Generating Documents

Documents are generated from the combination of

format ontology and context. Figure 1 summarizes

the process of document generation. Usually number

of instances to represent a document is big. For exam-

ple, Figure 5 shows a tree of instances of format on-

tology that characterizes the text shown in Figure 6.

Figure 7 shows interaction between one instance of

Text concept of format ontology and another instance

of Institute concept of content ontology.

Document is composed of paragraphs, in-

stances of Text concept are always inside a para-

graph. The string of the Figure 7, #insti-

AnOntologyforPortabilityandInteroperabilityDigitalDocuments-AnApproachinDocumentEngineeringusing

Ontologies

377

Student Sheet

body

metadata

par1_TblLogoHeader

par2_TblPersonalData

par3_space

par4_TblAcademicData

cel_11_ImgLogoJanus

cel_12_TxtJanus

cel_21_ImgLogoUSP

cel_22_TxtTitleUniversity

tblLogoHeader

tblPersonalData

txtSpace

tblAcademicData

contains

displayContent displayContent

has

isComposedOf isComposedOf

par1_cel11_TblLogoHeader

imgLogoJanus

isComposedOf

displayContent

styleBorderLess

has

styleTblLogoHeader

has

par1_cel12_TxtJanus

isComposedOf

has

styleLogo

has

“images_sheet_student/janus.jpeg”^^string

addressImage

txtTitleJanus

displayContent

styleParTitle

“System Administrative

Graduate”^^string

content

has

isComposedOf

…

“NONE”^^string

cellBorder

“LEFT”^^string

“CENTER”^^string

horizontalPos

textAligned

“200”^^string

spacedLines

“Erika Guetti Suca”^^string

author

contains

“2”^^unsignedShort

“1.0”^^double

contains contains

columns

rows

witdh

…

styleSubTitle

“12”^^short

“Times New Roman”^^string

has

fontSubTitle

size fontType

has

isComposedOf

par1_cel21_TxtJanus

isComposedOf

par1_cel22_TitleUniversity

isComposedOf

par2_cel22_TitleUniversity

par3_cel22_TitleUniversity

“14”^^short “Times New Roman”^^string

fontTitle

size

fonType

imgBrasaoJanus

txtNameUniversity

txtNameInstitute

txtTitleDocument

styleTitle

“images_sheet_student/brasaoUSP.gif”

^^string

styleLogo

has

displayContent

“University of

São Paulo”^^string

“Institute of Mathematics and

Statistics”^^string

“STUDENT SHEET”^^string

content

addressImage

has

has has

isComposedOf

Figure 5: Tree instances of ontology format corresponding to header’s student sheet.

Figure 6: Sample of header’s student sheet.

tute_EN_#nameInstitution is a reference to instance

of Institute and its attribute separated by the sym-

bol #. The instance name is institute_EN and its in-

stance attribute referenced is nameInstitution. The

Text concept rederizes in document the value of at-

tribute nameInstitution of instance institute_EN, for

the example is University of São Paulo.

Instances of format ontology refer to instances of

content ontology. The translators receives instances

of format and content ontology to mapper to objects

that represent components of docx and odt formats.

Finally, based on the representations of docx and odt

formats, a document is rendered in an appropriate for-

mat.

7 RESULTS

This study tested a small subset, basic word process-

ing features, of what is needed for multiple interop-

erable implementations. This work is not trying to

test extremely complex elements, but elements that

are routinely used.

The experiment was implemented in java using

KMIS2013-InternationalConferenceonKnowledgeManagementandInformationSharing

378

Figure 7: OWL code of an instance of concept Text of for-

mat ontology and an instance of concept Institute of content

ontology.

OWL Api

3.2, OpenDocument Odfdom

0.8.7 and

Apache POI

3.8. The documents generated in ODT

were tested using OpenOfﬁce.org3. In the case of

the format OOXML, the documents were tested using

Microsoft Ofﬁce 2010. Lastly, the ontologies were

created using Protégé

4.2.0.

The main difﬁculty in the experiment was main-

taining the aesthetic characteristics ﬁdelity of the doc-

ument. We have not considered features not com-

patibles between standars OOXML and ODF. It is

not possible to obtain always 100% of the translata-

bility between DOCX documents to ODT and vice

versa, due to the unique characteristics of standards

OOXML and ODF(Eckert et al., 2009).

The objective was to enable sharing documents

keeping the integrity of their information, i.e., to

achieve document interoperability while allowing

simple portability.

The format ontology and content ontology over-

come the problem of preserving digital documents,

eliminating the dependence of particular technologies

and enabling the central storage of documents. The

context ontology can be reused in other applications

and take full advantage of the power of expressivity

of ontologies. The format ontology can generate doc-

uments regardless of context ontology. Figures 8 and

9 show rebuilding the document from a sheet student

represented our model. Despite failing to reproduce

the 100% of the stylistic features equally in both for-

mats, the student information integrity has not been

http://owlapi.sourceforge.net/

http://incubator.apache.org/odftoolkit/odfdom/

index.html

http://poi.apache.org/

http://protege.stanford.edu/

Figure 8: Version generated for DOCX format.

Figure 9: Version generated for ODT format.

compromised. May be tolerable minimal lost or dif-

ference of stylistic document, but not the the student

information. The two student sheets continue serving

their purpose, reporting the data and student perfor-

mance. The documents maintained its authenticity.

8 CONCLUSIONS

This work has shown that ontologies and simple map-

pings provide a good foundation for the creation of

digital documents allowing document interoperabil-

AnOntologyforPortabilityandInteroperabilityDigitalDocuments-AnApproachinDocumentEngineeringusing

Ontologies

379

ity and simple portability. In addition, based on cen-

tralized representations of documents, it is possible to

change the physical interface without changing all its

mapped physical interfaces. Even more, new physical

interfaces could be added, perhaps for new display

or output devices, without changing the conceptual

model. If technologies change, translators will also

change even though the underlying conceptual mod-

els will not.

Using the same base of conceptual model, multi-

ple publicationsand formats could be created, and dif-

ferent assemblies of document could share common

structures or patterns. The idea is to reuse standard

schema components wherever it is possible. We could

easily imagine applications that reuse documents and

processes in ways not anticipated by theirs creators.

For future work, we will improve the implementa-

tion of translators and include anothers formats, PDF,

HTML, etc. A future application might be generat-

ing ﬁlled forms with predetermined information. To

get a complete ﬂow of document interoperability, We

will need create instances of format ontology from the

reading of documents. This would imply creation of

tool that assists in automating creation and reading of

instances. Then we would have complete ﬂow of doc-

ument interoperability, information represented in our

conceptual model could be updated with information

from documents.

Finally, We will also develop a use case docu-

ment interoperability applied to electronic govern-

ment, where format ontology and content ontology

should play an important role in preserving and dis-

tributing digital documents efﬁciently.

REFERENCES

Barron, D. W. (1996). Portable documents: problems and

partial solutions. Department of Electronics and Com-

puter Science University of Southampton, 8:343–367.

Bosak, J., McGrath, T., and Holman, G. K. (2011). Univer-

sal business language v2.1. Organization for the Ad-

vancement of Structured Information Standards (OA-

SIS), Standard.

Bretas, N. L. and do Socorro Ferreira Mesquita, C. (2010).

Panorana da Interoperabilidade no Brasil. Ministério

do Planejamento, Orçamento e Gestão.

Eckert, K.-P., Ziesing, J., and Ishionwu, U. (2009). Docu-

ment Operabilitiy Open Document Format and Ofﬁce

Open XML. Fraunhofer Verlag, Germany, fokus edi-

tion.

Ferreira, M. (2006). Introdução e preservasão digital: Con-

ceitos, estratégias e actuais consensos. Escola de

Engenharia da Universidade do Minho, edição elec-

trônica edition.

Glushko, R. J. and McGrath, T. (2008). Document En-

gineering: Analyzing and Designing, Documents

for Business, Informatics and Web Services. Mas-

sachusetts Institute of Technology.

Gruber, T. R. (1995). Toward principles for the design of

ontologies used for knowledge sharing. Int. J. Hum.-

Comput. Stud., 43(5-6):907–928.

Schmidt, K.-U., Fox, O., Henckel, L., Holzmann-Kaiser,

U., Martin, P., and Tschichholz, M. (2006). Document

Interoperability for Use in eGovernment: Integration

of XML-based Document Content in Public Adminis-

tration Processes. FOKUS.

Schroeder, R. and Mello, R. D. S. (2009). Designing xml

documents from conceptual schemas and workload in-

formation. Multimedia Tools Appl., 43(3):303–326.

Shah, R. and Kesan, J. (2009). Interoperability challenge

for open standars: Odf and ooxml as examples. The

proceedings of the 10th International Digital Govern-

ment Research Conference.

Uschold, M. (1998). Knowledge level modelling: concepts

and terminology. The Knowledge Engineering Re-

view, 13:1:5–29. Printed in the United Kingdom.

Vogel, T., Schmidt, A., Lemm, A., and Österle, H. (2008).

Service and document based interoperability for eu-

ropean ecustoms solutions. J. Theor. Appl. Electron.

Commer. Res., 3(3):17–37.

KMIS2013-InternationalConferenceonKnowledgeManagementandInformationSharing

380