EL 29,1. Zora Konjović Faculty of Technical Sciences, Institute of Computing and Automatics, Novi Sad, Serbia

Transcription

1 The current issue and full text archive of this journal is available at wwwemeraldinsightcom/ htm EL 52 Received 6 November 2009 Revised 25 December 2009 Accepted 31 December 2009 CERIF compatible data model based on MARC 21 format Dragan Ivanović Faculty of Technical Sciences, Institute of Computing and Automatics, Novi Sad, Serbia Dušan Surla Department of Mathematics and Informatics, Faculty of Sciences, Novi Sad, Serbia, and Zora Konjović Faculty of Technical Sciences, Institute of Computing and Automatics, Novi Sad, Serbia The Electronic Library Vol 29 No 1, 2011 pp q Emerald Group Publishing Limited DOI / Abstract Purpose The purpose of this research is to observe all data from the Common European Research Information Format (CERIF) data model that can be described using bibliographic standards and move those data to a data model of bibliographic standard Design/methodology/approach Analysis of the CERIF data model and the MARC 21 format has shown that some elements of the CERIF data model could be mapped to the MARC 21 bibliographic record A CERIF compatible data model based on the MARC 21 format is proposed The data model was created using PowerDesigner CASE tool The proposed data model is represented using a physical data model in the conceptual notation that is adopted in the literature for representing the CERIF data model Findings A CERIF compatible data model based on the MARC 21 format is proposed The proposed model contains all the data from the CERIF2008 data model The central part of the proposed model is MARC 21 data model that is used as a replacement for 27 entities of the CERIF data model, including all their attributes as well as part of the attributes in entities related to organisational unit The mappings between attributes of entities of the CERIF data model and the data model of the MARC 21 format are described Research limitations/implications The CERIF compatible data model based on the MARC 21 format does not support all restrictions on data types, which are defined by the CERIF data model This means that such restrictions have to be controlled by software Practical implications The central part of the proposed CERIF compatible data model is a data model of MARC 21 format It means that most of the data are modelled according to bibliographic standard, which is very widespread worldwide This implies that the proposed CERIF model can be easily implemented within the existing library infrastructure In addition, the proposed model can be used for other purposes, such as the evaluation of scientific research results, generating bibliographies of researchers, and institutions, the citations etc A research management system based on the proposed model is implemented Also, this system is verified and tested on data about published results of researchers employed at University of Novi Sad, Serbia Originality/value A new data model compatible with the CERIF data model is proposed The basic idea is to map part of the CERIF data model related to published results of scientific research to some well-known bibliographic standard It was shown that this part of the data model could be mapped to the MARC 21 data model It can be mapped to data models of any other MARC standards in a similar way Keywords Europe, Libraries, Research Paper type Research paper

2 1 Introduction In an era of rapid development of science and ICT (information and communication technologies) the scientific research information systems take significant role in the development of science (Zimmerman, 2002) Standardization of data used in those systems is necessary to enable scientists and researchers to find the desired data in various scientific research information systems and to enable exchanging data between those systems CERIF (Common European Research Information Format) (Asserson et al, 2002) describes a formal data model that enables interoperability between scientific research information systems that contain information about people, projects, organisations, publications, patents, equipments, etc The first version of a CERIF was published in 1991 and current version is the CERIF2008 that was published in 2008 CERIF describes a physical data model (Jörg et al, 2009a) and XML data exchange format (Jörg et al, 2009b) There are different aspects of the CERIF data model extensions The papers ( Jeffery et al, 2000; Jeffery, 2000) describe an extension that uses the formalized Dublin Core for description of published results of scientific research Furthermore, the papers (Asserson et al, 2002; Jeffery et al, 2000) describe a data model extension that was created in order to satisfy requirements of research information system that was developed at the University of Bergen Moreover, the paper (Asserson et al, 2002) describes a data model extension that was created in order to satisfy requirements of scientific research information system that was developed on Science and Technology Facilities Council A data model extension that was created in order to satisfy requirements of the IST World portal is described in the papers (Jörg et al, 2005, 2006; Kiryakov et al, 2005; Ferlež, 2005) There are data models that can be mapped to subset of the CERIF data model Math-Net International information and communication system[1] has such a data model Mapping between the Math-Net and CERIF data model is described in CERIF TaskGroup ( Jeffery et al, 2002) Common European format for Curricula Vitae[2] describes a person and her or his personal qualifications as it is defined according to the CERIF data model A non-profit organisation eurocris[3] whose task is to promote CRIS (Current Research Information System) has been responsible for the development of CERIF since 2002 CRIS is an information system of scientific research projects, researchers and research institutions Such a system has been developed within the BISIS library system The BISIS system has been developing since 1993 at the University of Novi Sad, Serbia and the current version 4 is based on XML technology Within this version, an XML editor for cataloguing in the UNIMARC and MARC 21 format (Dimić and Surla, 2009; Dimić et al, 2010) that uses the component for generating catalogue cards (Rad enović et al, 2009) has been developed The papers (Boberić and Surla, 2009; Milosavljević et al, 2010) describe the XML editor for searching and downloading bibliographic records in accordance with Z3950 standard, and the papers (Tešendić et al, 2009; Milosavljević and Tesendić, 2010) describe the library circulation system Additionally, the subject matter of the paper (Rudic and Surla, 2009) is the application of XML for conversion of bibliographic records in MARC 21 format The conversion is implemented by XSLT technology CERIF model based on MARC 21 format 53

3 EL 54 The European Union encourages the development of national CRIS systems in accordance with CERIF standard because the EU wants to achieve maximum competitiveness of Europe at all levels of research activity To achieve this goal there is a need for making CERIF compatible data models that would: support the relevant international standards For example, the bibliographic standards Those standards include a number of metadata, which would enable a more detailed description of entities of scientific research information systems; have a smaller physical model, but at the same time a more modular physical model, which would enable easier implementation of scientific research information systems; contain metadata about other types of publications that are not supported by CERIF standard such as technical documents, theses, dissertations, reports; be less sensitive to CERIF data model changes; and support the specific requirements of individual countries in the field of scientific research information systems Based on our experience gained from developing the BISIS system this paper gives a proposition of CERIF compatible data model based on the MARC 21 format The basic idea is to move all data from the CERIF model that can be described using MARC 21 format to the MARC 21 data model used within the BISIS system A proposed data model for the MARC 21 format is shown in section 2, a proposed CERIF compatible data model based on the MARC 21 format is shown in section 3 Data models were created using PowerDesigner CASE tool[4] The other sections describe the mapping of data about publications, patents, products, people, organisational units, projects and events from CERIF data model to MARC 21 format 2 Data model of MARC 21 format The MARC 21 standard consists of five documents: (1) MARC 21 Format for Bibliographic data[5] (2) MARC 21 Format for Authority Data[6] (3) MARC 21 Format for Holdings Data[7] (4) MARC 21 Format for Classification[8] (5) MARC 21 Format for Community Information[9] The MARC 21 standard defines the MARC record structure A MARC record consists of four parts: leader, directory, set of control fields and set of data fields The leader appears as the first part of the record and consists of 24 characters Each data in the record leader is a coded value located on the appropriate character position and carries the information necessary for record processing The directory contains information about length and start location for every control field and data field of the record Each control field and data field has a name Control fields contain unique data about publication or set of data about publication where each data are a coded value located on the appropriate character position Data fields contain up to two indicators The role of indicators is to enable more precise interpretation of data field content Indicator values are interpreted separately, and two indicators in the data field are mutually

4 independent Data fields contain a set of subfields containing name and content Subfield names can be lowercase alphabetic letter or numerical symbols Contents of subfields are primary carriers of information about the publication to which the record relates The data model described in this paper is based on the MARC 21 formats for bibliographic and authority data The MARC 21 format for bibliographic data contains bibliographic information about printed and manuscript text materials, computer files, maps, music, continuing resources, visual materials, and mixed materials Bibliographic data commonly includes titles, names, subjects, notes, publication data, and information about the physical description of an item The MARC 21 Format for Authority Data is designed to be a carrier for information concerning the authorized forms of names, subjects, and subject subdivisions to be used in constructing access points in MARC records, the forms of these names, subjects, and subject subdivisions that should be used as references to the authorized forms, and the interrelationships among these forms The data model shown in Figure 1 correspond to records that are in accordance with the MARC 21 format for bibliographic data and the MARC 21 format for authority data, because these two formats have the same structure The recordtype attribute of the Leader entity contains information whether the record containing Leader is bibliographic or authority The starting point for creating this data model was the object model of MARC 21 record from (Dimić and Surla, 2009) The record directory is omitted in the data model shown in Figure 1 because it can be generated based on values stored in record leader, control fields and data fields The MARC21Record entity has two attributes The first attribute is recordid, which holds the unique identifier of a MARC 21 record The second attribute is type which purpose is to connect MARC 21 records with entities of CERIF data model According to MARC 21 format control fields and data fields within the record can be repeatable The controlfieldordinalnumber attribute of the ControlField entity holds ordinal number of control field in an array of control fields with the same name The datafieldordinalnumber attribute of the DataField entity holds ordinal number of data field in an array of data fields with the same name Also, subfields within the data fields can be repeatable The subfieldordinalnumber attribute of the Subfield entity holds ordinal number of subfield in an array of subfields that belong to the same data field and have the same name The content attributes of the ControlField and Subfield entities contain information about the publication or authorised form to which the record relates CERIF model based on MARC 21 format 55 3 CERIF compatible data model Analyzing the CERIF data model it was observed that part of data model could be replaced with data model of MARC 21 format (see Figure 1) All attributes of entities related to publications, products, patents, persons, projects, events and set of attributes of entities related to organisational unit are replaced with data model presented in the figure The mappings between MARC 21 bibliographic records and attributes that belong to entities related to publication, products and patents are described in section 4 Section 5 describes the mappings between MARC 21 authority records and attributes that belong to entities related to persons, projects, events and set of attributes of entities related to organisational unit

5 EL 56 Figure 1 Data model of MARC 21 format The CERIF compatible data model that is proposed in this paper has 93 entities and which is for 67 entities less than CERIF data model The proposed model is divided into two parts because of the better visibility and shown in the Figures 2 and 3 Entities that are inherited from the CERIF data model and not related to the MARC 21 data model are not shown Replacements of CERIF data model entities with appropriate entities presented in Figure 2 are: Entity MARC21Record_Class is replacement for entities ResPubl_Class, ResProd_Class, ResPat_Class, Person_Class, OrgUnit_Class, Project_Class and Event_Class Entity cfmorgunit is replacement for attributes cfcurrcode, cfturn and cfheadcount of entity cforgunit

6 CERIF model based on MARC 21 format 57 Figure 2 CERIF compatible data model (1)

7 EL 58 Figure 3 CERIF compatible data model (2) Entity MARC21Record_Keywords is replacement for entities PersonKeywords, OrgUnitKeywords, ProjectKeywords and EventKeywords Entity MARC21Record_MARC21Record is replacement for entities ResPubl_Event, ResPubl_ResPat, ResPubl_ResProd, ResPubl_ResPubl, Person_Person, Person_OrgUnit, Person_Event, Person_ResPat, OrgUnit_Event, OrgUnit_OrgUnit, OrgUnit_ResPat, Project_Project, Project_Event, Project_OrgUnit, Project_Person, Project_ResPat, Project_ResProd and Event_Event Also, entity MARC21Record_MARC21Record is generalization of entities MARC21Record_ResPubl and MARC21Record_ResProd

8 Entity MARC21Record_ResPubl is replacement for entities Person_ResPubl, OrgUnit_ResPubl and Project_ResPubl Entity MARC21Record_ResProd is replacement for entities Person_ResProd and OrgUnit_ResProd Entities cfclass, cfcurrency and cflanguage are inherited from the CERIF data model CERIF model based on MARC 21 format 59 Replacements of CERIF data model entities with appropriate entities presented in Figure 3 are: Entity MARC21Record_Srv is replacement for entities Person_Srv, OrgUnit_Srv and Project_Srv Entity MARC21Record_Equip is replacement for entities Person_Equip, OrgUnit_Equip and Project_Equip Entity MARC21Record_ExpSkills is replacement for entities Person_ExpSkills and OrgUnit_ExpSkills Entity MARC21Record_Facil is replacement for entities Person_Facil, OrgUnit_Facil and Project_Facil Entity MARC21Record_FundProg is replacement for entities ResPubl_FundProg, ResProd_FundProg, ResPat_FundProg, Person_FundProg, OrgUnit_FundProg, Project_FundProg and Event_FundProg Entity MARC21Record_EAddr is replacement for entities Person_EAddr and OrgUnit_EAddr Entity MARC21Record_PAddr is replacement for entities Person_Paddr and OrgUnit_PAddr Entity MARC21Record_Prize is replacement for entities Person_Prize, OrgUnit_Prize and Project_Prize Entity MARC21Record_CERIFEntity is generalization of entities MARC21Record_Srv, MARC21Record_Equip, MARC21Record_ExpSkills, MARC21Record_Facil, MARC21Record_FundProg, MARC21Record_EAddr and MARC21Record_PAddr Entities cfclass, cfcurrency, cfsrv, cfequip, cfexpskills, cffacil, cffundprog, cfeaddr, cfpaddr and cfprize are inherited from the CERIF data model All attributes of the CERIF data model entities are uniquely mapped to the proposed model This model preserves the existing references between the CERIF data model entities The model is based on the MARC 21 format that is rich in metadata and allows a more detailed description of the entities of scientific research information systems Information system based on the proposed data model can exchange data with other information systems using XML documents whose XML schemes are described in (Jörg et al, 2009b)

9 EL 60 4 The mapping of attributes of CERIF data model entities to MARC 21 bibliographic record This section shows the mappings of attributes of entities that are related to publications, products and patents to attributes of the DataField entity and attributes of the Subfield entity of MARC 21 bibliographic record data model (see Figure 1) 41 data about publication An instance of the ResultPublication CERIF entity, together with connected entities (ResultPublicationBiblNote, ResultPublicationTitle, ResultPublicationSubtitle, ResultPublicationNameAbbrev, ResultPublicationKeywords and ResultPublication Abstract) is mapped to an instance of the MARC 21 bibliographic record Table I shows that the values of attributes of the ResultPublication entity are stored into the content attribute of the corresponding Subfield entity For example, the value of the cfuri attribute of the ResultPublication entity is stored into the content attribute of the Subfield entity which holds value u in the subfieldname attribute and is connected to the DataField entity which holds value 856 in the datafieldname attribute and holds value # in the attributes ind1 and ind2 The value of the first indicator defines an access method to the electronic resource, and the value of the second indicator defines relationship between the electronic resource and the item described in the record as a whole The value # indicates that the access method to the electronic resource and relationship between the electronic resource and the item described in the record as a whole are not defined which is in accordance with the CERIF data model which does not keep previously mentioned information The MARC 21 format offers possibility of a more detailed description of the URI that is associated with a publication The previous example in MARC terminology can be described as follows: the value of the cfuri attribute of the ResultPublication entity is stored in the subfield u of the data field 856, whose first indicator and second indicator hold value # Hereinafter this terminology is used to describe the mappings of CERIF data model entities attributes to MARC 21 record Entity: ResultPublication Entity: Subfield Entity: DataField Attributes subfield Name datafield Name Ind1 ind2 Table I ResultPublication entity attributes cfisbn A 020 # # cfissn A 022 # # cfedition A 250 # # cfissue B 260 # # cfrespubldate C cfseries A # CfNum V CfVol CfStartPage A 300 # # CfEndPage CfTotalPages CfURI U 856 # #

10 Values of the cfvol and cfnum attributes are stored in the subfield v of the data field 490 in the following format: vol value of the cfvol attribute, no value of the cfnum attribute Values of the cfstartpage, cfendpage and cftotalpages attributes are stored in the subfield a of the data field 300 in the following format: value of the cftotalpages attribute p ( value of the cfstartpage attribute - value of the cfendpage attribute ) CERIF model based on MARC 21 format 61 ResultPublicationBiblNote, ResultPublicationTitle, ResultPublicationSubtitle, ResultPublicationNameAbbrev, ResultPublicationKeywords and ResultPublication Abstract CERIF entities allow entry of multilingual data about the ResultPublication entity Table II shows mappings of attributes of these entities instances that contain data in the original language to attributes of the DataField and the Subfield entities of data model shown in Figure 1 The value of the cflangcode attribute of these entities is entered in the control field 008 on character positions The cftrans attribute has the value o (original) for data in original language The instances of entities from the previous table that contain translated data are mapped to subfields of the data field 880 For example, Table III shows values of attributes of the Subfield entity that illustrate storing of instances of the ResultPublicationBiblNote entity with original and translated data The relationship between the instance of the ResultPublicationBiblNote entity with original data and instances of the ResultPublicationBiblNote entity with translated data are made via the subfield 6 of the data field 500 The value should be MARC 21 CERIF Entity: Subfield Entity: DataField Entity Attributes subfield Name datafield Name ind1 ind2 ResultPublicationNameAbbrev cfnameabbrev a # ResultPublicationTitle CfTitle a ResultPublicationSubtitle CfSubtitle b ResultPublicationBiblNote CfBiblNote a 500 # # ResultPublicationAbstract CfAbstract a # ResultPublicationKeywords CfKeywords a 653 # # Table II attributes of entities instances with the original data Entity: Subfield Content subfield Name subfieldordinal Number datafield Name datafield OrdinalNumber recordid This is note a h-scc: a m-chi: a (OCoLC) Table III An example of multilingual data mappings

11 EL 62 interpreted as follows: Translations of This is note are in subfields a of data fields 880 whose subfields 6 hold value The CERIF data model entities that allow entry of some data in original language and translations of this data have attribute cflangcode, which holds information about data language and attribute cftrans, which can have three values: (1) o shortened from original (2) h shortened from human translate (3) m shortened from machine translate Values of attributes cftrans, cflangcode and translated data (cftrans o) are entered in following format: value of the cftrans attribute - value of the cflangcode attribute : translated data For example, value h-scc: from the Table III is interpreted as follows: h indicates it is a human translate, scc indicates it is a translation to Serbian language and is translated data 42 data about patent An instance of the ResultPatent CERIF entity together with connected entities (ResultPatentTitle, ResultPatentKeywords and ResultPatentAbstract) is mapped to an instance of the MARC 21 bibliographic record Table IV shows the mappings of ResultPatent entity attributes The value of the cfregisterdate attribute is stored in the subfield d of the data field 013 followed by the subfield e, which holds value effective The value of the cfapprovdate attribute is stored in the subfield d of the data field 013 followed by the subfield e, which holds value granted Table V shows values of Subfield entities attributes that illustrate mappings of ResultPatent entity attributes cfregisterdate and cfapprovdate to data model shown in the Figure 1 ResultPatentTitle, ResultPatentKeywords and ResultPatentAbstract CERIF entities allow entry of multilingual data about the ResultPatent entity Table VI shows mappings of attributes of these entities instances that contain data in the original language to attributes of the DataField and the Subfield entities of data model shown in the Figure 1 The value of the cflangcode attribute of these entities is stored in the control field 008 on character positions The cftrans attribute has the value o Entity: ResultPatent Entity: Subfield Entity: DataField Attributes subfield Name datafield Name ind1 ind2 Table IV ResultPatent entity attributes cfpatentnum A 013 # # cfcountrycode B cfregisterdate D CfApprovDate D CfURI U 856 # #

12 (original) for data in original language Attributes of these entities instances that contain translated data are mapped to subfields of the data fields 880 as it was previously described in section 41 Subfields 6 of these 880 data fields hold appropriate value from datafieldname column of the Table VI, for instance 245, 520 or data about product An instance of the ResultProduct CERIF entity together with connected entities (ResultProductName, ResultProductKeywords and ResultProductDescription) is mapped to an instance of the MARC 21 bibliographic record ResultProduct entity attributes are shown in Table VII ResultProductName, ResultProductKeywords and ResultProductDescription CERIF entities allow entry of multilingual data about the ResultProduct entity The Table VIII shows mappings of attributes of these entities instances that contain data in the original language to attributes of the DataField and the Subfield entities of data model shown in the Figure 1 The value of the cflangcode attribute of these entities is entered in the control field 008 on character positions The cftrans attribute has the value o (original) for data in original language Attributes of these entities instances that contain translated data are mapped to subfields of the data fields 880 as it was CERIF model based on MARC 21 format 63 Entity: Subfield Content subfield Name subfieldordinal Number datafield Name datafield OrdinalNumber recordid d (OCoLC) Effective e d 02 Granted e 02 Table V An example of the mappings of ResultPatent entity attributes cfregisterdate and cfapprovdate MARC 21 CERIF Entity: Subfield Entity: DataField Entity Attributes Subfield Name datafield Name ind1 ind2 ResultPatentTitle cftitle a ResultPatentAbstract cfabstract a # ResultPatentKeywords cfkeywords a 653 # # Table VI attributes of entities instances with the original data Entity: ResultProduct Entity: Subfield Entity: DataField Attributes subfield Name datafield Name ind1 ind2 cfresprodinterid a 084 # # CfURI u 856 # # Table VII ResultProduct entity attributes

13 EL 64 previously described in section 41 Subfields 6 of these data fields 880 hold appropriate value from datafieldname column of Table VIII, for instance 245, 500 or The mapping of attributes of CERIF data model entities to MARC 21 authority record This section shows the mappings of attributes of entities that are related to the persons, projects, events and set of attributes of entities that are related to organisational units to attributes of the DataField entity and attributes of the Subfield entity of MARC 21 authority record data model (see Figure 1) 51 data about person An instance of the Person CERIF entity together with connected entities (PersonName, PersonName_Person and PersonResInt) are mapped to an instance of the MARC 21 authority record Table IX gives the mappings of Person entity attributes, while Table X shows the mappings of PersonName entity attributes Values of the cffamilynames, cfothernames and cffirstnames attributes are stored in the subfield a of the data field 100 in the following format: value of the cffamilynames attribute, value of the cfothernames attribute, value of the cffirstnames attribute Table XI shows the mappings of PersonName_Person entity attributes Table VIII attributes of entities instances with the original data MARC 21 CERIF Entity: Subfield Entity: DataField Entity Attributes subfield Name datafield Name ind1 ind2 ResultProductName cfname a ResultProductDescription cfdescription a 500 # # ResultProductKeywords cfkeywords a 653 # # Table IX Person entity attributes Entity: Person Entity: Subfield Entity: DataField Attributes subfield Name datafield Name ind1 ind2 CfSex a 375 # # CfURI u 856 # # Table X PersonName entity attributes Entity: PersonName Entity: Subfield Entity: DataField Attributes subfield Name datafield Name ind1 ind2 cffamilynames cfothernames cffirstnames a #

14 Values of the cfclassid and cfclassschemeid attributes are stored in the subfield x of the data field 400 in the following format: value of the cfclassschemeid attribute : value of the cfclassid attribute Values of the cfstartdate and cfenddate attributes are stored in the subfield y of the data field 400 in the following format: value of the cfstartdate attribute value of the cfenddate attribute CERIF model based on MARC 21 format 65 PersonResInt CERIF entity allows entry of multilingual data about the Person entity Mappings of attributes of this entity instances that contain data in the original language to attributes of the DataField and the Subfield entities of data model shown in the Figure 1 are given in Table XII The value of the cflangcode attribute of this entity is entered in the subfield b of the data field 040 The cftrans attribute has the value o (original) for data in original language Attributes of these entities instances that contain translated data are mapped to subfields of the data fields 880 as it was previously described in section 41 Subfields 6 of these data fields 880 hold value data about organisational unit An instance of the OrgUnit CERIF entity together with connected entities (OrgUnitName and OrgUnitResAct) is mapped to an instance of the MARC 21 authority record and an instance of cfmorgunit entity shown in Figure 2 Table XIII shows the mappings of OrgUnit entity attributes cfacro and cfuri Entity: PersonName_Person Entity: Subfield Entity: DataField Attributes subfield Name datafield Name ind1 ind2 CfPersNameVar CfClassId CfClassSchemeId CfStartDate CfEndDate a x y # Table XI PersonName_ Person entity attributes Entity: PersonResInt Entity: Subfield Entity: DataField Attributes subfield Name datafield Name ind1 ind2 CfResInt a 372 # # Table XII attributes of PersonResInt entity instances with the original data Entity: OrgUnit Entity: Subfield Entity: DataField Attributes subfield Name datafield Name ind1 ind2 CfAcro a # CfURI u 856 # # Table XIII OrgUnit entity attributes

15 EL 66 OrgUnitName and OrgUnitResAct CERIF entities allow entry of multilingual data about the OrgUnit entity Table XIV shows mappings of attributes of these entities instances that contain data in the original language to attributes of the DataField and the Subfield entities of data model shown in Figure 1 The value of the cflangcode attribute of these entities is entered in the subfield b of the data field 040 The cftrans attribute has the value o (original) for data in original language Attributes of these entities instances that contain translated data are mapped to subfields of the data fields 880 as it was previously described in section 41 Subfields 6 of these data fields 880 hold appropriate value from datafieldname column of the table, for instance 110 or data about project An instance of the Project CERIF entity together with connected entities (ProjectName, ProjectAbstract) is mapped to an instance of the MARC 21 authority record Table XV provides the mappings of Project entity attributes Values of the cfstartdate and cfenddate attributes are stored in the subfield y of the data field 110 in the following format: value of the cfstartdate attribute - value of the cfenddate attribute ProjectName and ProjectAbstract CERIF entities allow entry of multilingual data about the Project entity In Table XVI are shown the mappings of attributes of these entities instances that contain data in the original language to attributes of the DataField and the Subfield entities of data model shown in Figure 1 The value of the Table XIV attributes of entities instances with the original data MARC 21 CERIF Entity: Subfield Entity: DataField Entity Attributes Subfield Name datafield Name ind1 ind2 OrgUnitName cfname a # OrgUnitResAct cfresact a 372 # # Entity: Project Entity: Subfield Entity: DataField Attributes subfield Name datafield Name ind1 ind2 Table XV Project entity attributes cfstartdate CfEndDate y # CfAcro a # CfURI u 856 # # Table XVI attributes of entities instances with the original data MARC 21 CERIF Entity: Subfield Entity: DataField Entity Attributes Subfield Name datafield Name ind1 ind2 ProjectName cfname a # ProjectAbstract cfresact i 680 # #

16 cflangcode attribute of these entities is entered in the subfield b of the data field 040 The cftrans attribute has the value o (original) for data in original language Attributes of these entities instances that contain translated data are mapped to subfields of the data fields 880 as it was previously described in section 41 Subfields 6 of these data fields 880 hold appropriate value from datafieldname column in Table XVI, for instance 110 or data about event An instance of the Event CERIF entity together with connected entities (EventName, EventDescription) is mapped to an instance of the MARC 21 authority record Table XVII shows the mappings of Event entity attributes Values of the cfstartdate and cfenddate attributes are stored in the subfield d of the data field 111 in the following format: value of the cfstartdate attribute - value of the cfenddate attribute CERIF model based on MARC 21 format 67 EventName and EventDescription CERIF entities allow entry of multilingual data about the Event entity Table XVIII shows mappings of attributes of these entities instances that contain data in the original language to attributes of the DataField and the Subfield entities of data model shown in the Figure 1 The value of the cflangcode attribute of these entities is stored in the subfield b of the data field 040 The cftrans attribute has the value o (original) for data in original language Attributes of these entities instances that contain translated data are mapped to subfields of the data fields 880 as it was previously described in section 41 Subfields 6 of these data fields 880 hold appropriate value from datafieldname column Table XVIII, for instance 111 or Conclusion The proposed data model is compatible with the CERIF data model It means that the information system based on this data model is interoperable with other information Entity: Event Entity: Subfield Entity: DataField Attributes subfield Name datafield Name ind1 ind2 CfLocation c CfStartDate CfEndDate d # CfFeeOrFree g CfURI u 856 # # Table XVII Event entity attributes MARC 21 CERIF Entity: Subfield Entity: DataField Entity Attributes subfield Name datafield Name ind1 ind2 EventName CfName a # EventDescription CfDescription i 680 # # Table XVIII attributes of entities instances with the original data

17 EL 68 systems based on the CERIF data model Advantages of the proposed CERIF compatible data model comparing with the CERIF data model are: The proposed model is more general and contains more metadata than the CERIF model MARC 21 format, which is the base for this model is rich in metadata and allows more detailed description of entities of scientific research information systems For example, MARC 21 format for bibliographic data contains field 510 that can hold information about who and where cited the publication described by MARC 21 record The proposed physical data model is smaller The CERIF data model has 160 entities, and the proposed data model has 93 entities Also, the proposed model is more modular The proposed model can contain metadata about other publication type that are not supported by the CERIF data model such as: technical documentations, theses, dissertations, reports The MARC 21 format for bibliographic data can describe previously mentioned publications The proposed model is less sensitive to the CERIF data model changes It is possible that a future extension of CERIF data model will add an entity or attributes to existing entities that could map to the MARC 21 format In this case, the proposed data model remains the same The proposed model contains repository of bibliographic data about scientific research results of individuals and institutions, which can be used for the evaluation of scientific research results, generating bibliographies of researchers and institutions, the citations etc A research management system based on the proposed model is implemented Also, this system is verified and tested on data about published results of researchers employed at University of Novi Sad Notes 1 Math-net an International Information and Communication System, available at: wwwmathnetde/ (accessed 3 October 2009) 2 Commission recommendation on a common European format for curricula vitae (CVs), available at: ¼ OJ:L:2002:079:0066: 0072:EN:PDF (accessed 13 October 2009) 3 eurocris, Current Research Information Systems, available at: wwweurocrisorg/ (accessed 13 October 2009) 4 PowerDesigner, available at: wwwsybasecom/products/modelingdevelopment/ powerdesigner (accessed 13 October 2009) 5 MARC 21 Format for Bibliographic Data: Table of Contents (Network Development and MARC Standards Office, Library of Congress), available at: wwwlocgov/marc/ bibliographic/ecbdhomehtml (accessed 13 October 2009) 6 MARC 21 Format for Authority Data: Table of Contents (Network Development and MARC Standards Office, Library of Congress), available at: wwwlocgov/marc/authority/ecadhome html (accessed 13 October 2009)

18 7 MARC 21 Format for Holdings Data: Table of Contents (Network Development and MARC Standards Office, Library of Congress), available at: wwwlocgov/marc/holdings/echdhome html (accessed 13 October 2009) 8 MARC 21 Format for Classification Data: Table of Contents (Network Development and MARC Standards Office, Library of Congress), available at: wwwlocgov/marc/ classification/eccdhomehtml (accessed 13 October 2009) 9 MARC 21 Format for Community Information: Table of Contents (Network Development and MARC Standards Office, Library of Congress), available at: wwwlocgov/marc/ community/eccihomehtml (accessed 13 October 2009) CERIF model based on MARC 21 format 69 References Asserson, A, Jeffery, K and Lopatenko, A (2002), CERIF: past, present and future: an overview, Proceedings of the 6th International Conference on Current Research Information Systems, University of Kassel, Kassel, August 29-31, pp Boberić, D and Surla, D (2009), XML editor for search and retrieval of bibliographic records in the Z3950 Standard, The Electronic Library, Vol 27 No 3, pp Dimić, B and Surla, D (2009), XML editor for UNIMARC and MARC21 cataloguing, The Electronic Library, Vol 27 No 3, pp Dimić, B, Milosavljević, B and Surla, D (2010), XML schema for UNIMARC and MARC 21 formats, The Electronic Library, Vol 28 No 2, pp Ferlež, J (2005), Public IST World Deliverable 13 Data Model for Representation of Expertise, p 12, available at: DataModelForRepresentationOfExpertisepdf (accessed 13 October 2009) Jeffery, K (2000), An architecture for grey literature in an R&D context, The International Journal on Grey Literature, Vol 1 No 2, pp Jeffery, K, Asserson, A and Revheim, J (2000), CRIS, grey literature and the knowledge society, Proceedings of CRIS-2000, Helsinki, available at: ftp://ftpcordislu/pub/cris2000/ docs/jeffery_fulltextpdf (accessed 13 October 2009) Jeffery, K, Lopatenko, A and Asserson, A (2002), Comparative study of metadata for scientific information: the place of CERIF in CRISs and scientific repositories, Proceedings of the 6th International Conference on Current Research Information Systems, University of Kassel, Kassel, August 29-31, pp Jörg, B, Ferlež, J and Grabczewski, E (2005), Public IST World Deliverable 12 Data Model for Knowledge Organisation, p 18, available at: deliverables/istworld_d12_datamodelforknowledgeorganisationpdf (accessed 13 October 2009) Jörg, B, Ferlež, J, Grabczewski, E and Jermol, M (2006), IST World: European RTD information and service portal, 8th International Conference on Current Research Information Systems: Enabling Interaction and Quality: Beyond the Hanseatic League (CRIS 2006), Bergen, p 10, available at: ISTWorld01pdf (accessed 13 October 2009) Jörg, B, Krast, O, Jeffery, K and Grootel, G (2009a), CERIF Full Data Model (FDM) Introduction and Specification, p 43, available at: wwweurocrisorg/fileadmin/cerif-2008/ CERIF2008_10_FDMpdf (accessed 13 October 2009) Jörg, B, Krast, O, Jeffery, K and Grootel, G (2009b), CERIF XML Data Exchange Format Specification, p 33, available at: wwweurocrisorg/fileadmin/cerif-2008/ CERIF2008_10_XMLpdf (accessed 13 October 2009) Kiryakov, A, Grabczewski, E, Ferlež, J, Uszkoreit, H and Jörg, B (2005), Public IST World Deliverable 11 Definition of the Central Data Structure, p 23, available at:

19 EL 70 worlddfkide/downloads/deliverables/istworld_d11_centraldatastructurepdf (accessed 13 October 2009) Milosavljević, B and Tesendić, D (2010), Software architecture of distributed client/server library circulation, The Electronic Library, Vol 28 No 2, pp Milosavljević, B, Boberić, D and Surla, D (2010), Retrieval of bibliographic records using Apache Lucene, The Electronic Library, Vol 28 No 4, pp Rad enović, J, Milosavljević, B and Surla, D (2009), Modelling and implementation of catalogue cards using FreeMarker, Program: Electronic library and information systems, Vol 43 No 1, pp Rudić, G and Surla, D (2009), Conversion of bibliographic records to MARC 21 format, The Electronic Library, Vol 27 No 6, pp Tešendić, D, Milosavljević, B and Surla, D (2009), A library circulation system for city and special libraries, The Electronic Library, Vol 27 No 1, pp Zimmerman, E (2002), CRIS-cross: current research information systems at a crossroads, Proceedings of the 6th International Conference on Current Research Information Systems, University of Kassel, Kassel, August 29-31, pp About the authors Dragan Ivanović has worked in the Department of Computing and Automatics, Faculty of Technical Sciences, Novi Sad, Serbia as a Teaching Assistant since 2007 Dragan Ivanović received his Master s degree in 2006 in Computer Science from the University of Novi Sad, Faculty of Technical Sciences He holds teaching seminars on the subjects Web Programming and Network-based Systems in the Department of Computing and Automatics He has published seven scientific papers Dragan Ivanović is the corresponding author and can be contacted at: chenejac@unsacrs Dušan Surla has been a Full Professor in the Department of Mathematics and Informatics, Faculty of Science, Novi Sad, Serbia since 1991 Dušan Surla received his Bachelor s degree in Mathematics from the University of Novi Sad, Faculty of Philosophy in 1969, Master s degree in Robotics from the University of Novi Sad, Faculty of Mechanics in 1976, and PhD degree in Robotics from the University of Novi Sad, Faculty of Technical Sciences in 1980 He has been with the Faculty of Science in Novi Sad since 1976 Dušan Surla has participated in 14 science projects; in eight of which he was the project leader He has published more than 170 scientific and professional papers Zora Konjović has been a Full Professor in the Faculty of Technical Sciences, Novi Sad, Serbia since 2003 Zora Konjović received her Bachelor s degree in Mathematics from the University of Novi Sad, Faculty Science in 1973, Master s degree in Robotics from the University of Novi Sad, Faculty of Technical Sciences in 1985, and PhD degree in Robotics from the University of Novi Sad, Faculty of Technical Sciences in 1992 From 1973 until 1980, she was with the Faculty of Science in Novi Sad and, since 1980, she has been with the Faculty of Technical Sciences, University of Novi Sad Zora Konjović has participated in five scientific and more than 30 professional projects; in five she was the project leader She has published more than 150 scientific and professional papers To purchase reprints of this article please reprints@emeraldinsightcom Or visit our web site for further details: wwwemeraldinsightcom/reprints