interoperability between library software: a solution for iranian libraries

9
Interoperability between library software: a solution for Iranian libraries Mehdi Alipour Hafezi Science and Research Branch, Islamic Azad University (IAU), Tehran, Iran Abstract Purpose – The main purpose of this paper is to propose a solution for data interoperability between library programs in Iran. Design/methodology/approach – The research proceeded through expressing the essence of interoperability in library programs in order to exchange metadata, In this regard, the current situation is analyzed using a researcher-made checklist, and then the problems and shortcomings are highlighted in the field of the interoperability which in turn enables us to find some ways to overcome them. Findings – The majority of the library software in Iran do not respect data exchange. They mostly use ISO 2709 as an export format and rarely use other formats. Moreover, most of the library software use Z39.50 client to get information from Library of congress and also Iranian National Library. Therefore, none of them could exchange data between each other because of not using server side service. The proposed model tries to introduce harvesting metadata by OAI service provider and also searching the metadata records by SRU client-server model. Originality/value – The findings indicate that Iranian libraries should be aware of the essence of interoperability. Using the proposed model would help them to exchange metadata in a cost-efficient and cost-effective manner. Keywords Open systems, Libraries, Iran, Z39.50 Paper type Research paper Introduction Preparing library catalogues has usually been done by professional librarians. Each library has been trying to build catalogue cards for their own information materials. Despite considering the same principles in descriptive and analytical cataloguing, they encountered huge differences between their catalogues, especially in analytical cataloguing. Such a situation may relate to their library policies and also to their catalogers viewpoints. Moreover, the mentioned activity was time and cost consuming job. For more than half a century, technological developments have let libraries use computers and specially networks to carry out functions in the libraries and information centers and to keep themselves and users in touch. Thus, computers have been used in libraries for their bibliographic data and afterwards as well as for library services. Networks help them not to be such an island – only accessible for those living there. The internet is a wide area network created by other networks that work with The current issue and full text archive of this journal is available at www.emeraldinsight.com/0264-0473.htm The author would like to thank Drs Ali Shiri and Yazdan Mansurian and Mr Alireza Esfandiari Mogaddam for their kind help in preparing the paper, and the studied programs’ managers for their assistance in aggregating needed information. TEL 26,5 726 Received 9 March 2007 Revised 7 September 2007 Accepted 19 September 2007 The Electronic Library Vol. 26 No. 5, 2008 pp. 726-734 q Emerald Group Publishing Limited 0264-0473 DOI 10.1108/02640470810910747

Upload: mehdi

Post on 19-Dec-2016

213 views

Category:

Documents


1 download

TRANSCRIPT

Interoperability between librarysoftware: a solutionfor Iranian libraries

Mehdi Alipour HafeziScience and Research Branch, Islamic Azad University (IAU), Tehran, Iran

Abstract

Purpose – The main purpose of this paper is to propose a solution for data interoperability betweenlibrary programs in Iran.

Design/methodology/approach – The research proceeded through expressing the essence ofinteroperability in library programs in order to exchange metadata, In this regard, the current situationis analyzed using a researcher-made checklist, and then the problems and shortcomings are highlightedin the field of the interoperability which in turn enables us to find some ways to overcome them.

Findings – The majority of the library software in Iran do not respect data exchange. They mostlyuse ISO 2709 as an export format and rarely use other formats. Moreover, most of the library softwareuse Z39.50 client to get information from Library of congress and also Iranian National Library.Therefore, none of them could exchange data between each other because of not using server sideservice. The proposed model tries to introduce harvesting metadata by OAI service provider and alsosearching the metadata records by SRU client-server model.

Originality/value – The findings indicate that Iranian libraries should be aware of the essence ofinteroperability. Using the proposed model would help them to exchange metadata in a cost-efficientand cost-effective manner.

Keywords Open systems, Libraries, Iran, Z39.50

Paper type Research paper

IntroductionPreparing library catalogues has usually been done by professional librarians. Eachlibrary has been trying to build catalogue cards for their own information materials.Despite considering the same principles in descriptive and analytical cataloguing, theyencountered huge differences between their catalogues, especially in analyticalcataloguing. Such a situation may relate to their library policies and also to theircatalogers viewpoints. Moreover, the mentioned activity was time and cost consuming job.

For more than half a century, technological developments have let libraries usecomputers and specially networks to carry out functions in the libraries andinformation centers and to keep themselves and users in touch. Thus, computers havebeen used in libraries for their bibliographic data and afterwards as well as for libraryservices. Networks help them not to be such an island – only accessible for those livingthere. The internet is a wide area network created by other networks that work with

The current issue and full text archive of this journal is available at

www.emeraldinsight.com/0264-0473.htm

The author would like to thank Drs Ali Shiri and Yazdan Mansurian and Mr Alireza EsfandiariMogaddam for their kind help in preparing the paper, and the studied programs’ managers fortheir assistance in aggregating needed information.

TEL26,5

726

Received 9 March 2007Revised 7 September 2007Accepted 19 September2007

The Electronic LibraryVol. 26 No. 5, 2008pp. 726-734q Emerald Group Publishing Limited0264-0473DOI 10.1108/02640470810910747

each other by common protocols and it has made it possible for libraries to co-operatewith each other on a global basis. Equally, users are able to have wide access to librarycatalogues, services and also their full-text data in the latest kind of libraries, i.e. digitaland virtual libraries.

Libraries on the internet necessarily need to be interoperable in order to becost-effective and also cost-efficient. In fact, they need bibliographic data which havebeen previously produced by the other libraries. They may also need to exchangefull-text data or other kinds of data which they need in their daily activities. So, havingonline interoperability between library software is a valuable solution in suchcircumstances. The internet and its related protocols prepared such an environment, solibraries could connect to each other and exchange their data. Most libraries andinformation centers in the developed countries have been using these technologies andtheir related privileges for years in order to overcome their issues.

On the other hand, most of the libraries in developing countries have not yet made theirlibrary systems interoperable. Interoperability is more pragmatic than other approachessuch as attempting to make all systems compatible at the code level, focusing solely onadding new layers of middleware that try to make all systems look and act the same, orseeking to make different systems interchangeable (Gates, 2005, p. 13).

Libraries in Iran use various kinds of software for their catalogues and full-text datajust as in other countries. Indeed, there are a few public programs used by some of thelibraries. Programs, especially library software, have different architectures, and alsouse specific protocols in order to interact with others which use the same software. So,how do the different products from different companies interact with each other becauseof the incredible diversity of PC hardware and programs and also different architecture?

This question and some other related questions spurred the author to do someresearch in this area. In order to answer the questions, the current situation of thelibrary software in Iran was studied in the case of data repository, data exportstructure and data exchange by using a researcher-made checklist. Also, the currenttechnologies, and their perspective which have been using by the libraries in thedeveloped countries studied in order to propose a practical solution.

Why interoperability?Libraries are organizations which prepare their materials and services for those who arelooking for information. In fact, user’s information needs could be categorized in groupsin which each group has different specifications to the others. But, the item which isimportant in our discussion is the user’s struggle to overcome their information need.The libraries’ significance and highlighted role started from this point – from the user’sviewpoint. When they refer to a library – physically or virtually – they expect to findsome information or even appropriate assistance and directions to answer theirinformation request. Therefore, libraries and information centers should know theirusers and also aware from their information needs in each period of time.

Libraries – from ancient times to the present – try to be the best reference forpeople who encountered a problem in the process of research or find a gap in theirknowledge. Libraries factually prepare some information resources related to theirusers’ questions. But, like their users, they actually cannot collect all the informationwhich is published in the field, because of the space, time and financial restrictions.Therefore, libraries have resorted to building consortiums with other libraries in their

Interoperabilitybetween library

software

727

fields in order to make wider access to information resources. One of the libraryservices for making the library resources accessible is inter-library loan (ILL) service.In this case, libraries try to meet their user’s requests via other libraries provided thatthey do not access them in their own collection. Thus, it becomes practical to sharemetadata collections across libraries, each needing to use another’s collections in orderto answer their user’s requests.

Furthermore, libraries need to exchange their bibliographic data electronically,because of the high cost of original cataloging. In fact, original cataloging is somehowtreated as an expensive and time-consuming job in libraries. Therefore, most of thelibraries try to collect bibliographic data for their library materials from the referencecenters known as pioneers in the field which prepared previously the bibliographicdata. This means that each library can just catalogue its core materials and collectother data from specialist organizations. For instance, libraries in the USA participatein shared cataloguing networks like OCLC in order to find copy cataloguing for 91-99per cent of their foreign and domestic titles in a database like World Cat. Originalcataloguing in the mentioned case is only created for the remaining 1-9 per cent (Spice,2001, pp. 171-2). In contrast, in Central and Eastern Europe, studies by Borgman (1996)found that the proportions of copy and original cataloguing are nearly the reverse ofthose in the USA. In Iran, the situation is closer to that of the USA, because of theirresemblance and emphasis in three languages: Persian, English and Arabic.

Nowadays, digital libraries are encountered with some new issues in full text dataexchange between libraries. In fact, digital libraries, like their predecessors, need to beinteroperable so that they can import or export bibliographic and even full text data.A new feature of ILL services is that they are now faced with this type of library.Digital libraries and virtual libraries, because they existing in the same environment –digital environment – are actually in the same situation. In a nutshell, libraries need tobe dynamically related to each other. So, interoperable libraries have powerful abilitiesto satisfy their users. As Bill Gates mentioned, in the new technology world, time is tooshort because of the speed of the changing technology.

A long time ago, libraries did not have such an essence to be interoperable with otherinformation centers, but libraries in the current era strictly need to be interoperable.

What is the current situation?Library programs are firstly prepared to automate library activities. They shouldadapt themselves with technology changes. In fact, new technologies change thepublic’s appetite in order to use it in their daily lives. Therefore, if libraries want tokeep their current key status in societies, they should be continually adapt to and keepup with technological changes.

In order to study the interoperability of Iranian library software, three basicquestions are posed:

(1) How is the data repository architecture in Iranian library systems’ databases?The main point of this question is adaptability.

(2) How is the data export structure in library software’ databases? Taking care ofexport standard format is the core issue in this question.

(3) How do the library programs exchange their data? Being interoperable withother library programs is the main goal of this question.

TEL26,5

728

The above three basic questions are composed of 17 queries in a checklist for gatheringrelated information from the companies which produced library software in Iran. Thefirst basic question was not too effective in this research, because of the differentstorage structures used by programs. In fact, the structure of storage is arguablyrelated to their database architecture. Thereby, their repository diversity would notmake highlighted problems in having interoperable systems. Although the standardstructure naturally simplifies the process of interoperability between programs, buteach one could have its own architecture.

Among the research population, national library software (RASA) is the only oneconsistent with UNIMARC guidelines to data repository. The others use their ownspecific format in their database structure without respecting to internationalstandards. ISO 2709 is the common format for exchange underlying all MARC formats;it consists of a record label, a directory and data fields, with standard characters forseparation (de Carvalho et al., 2004, p. 131).

According to the export format – as presented in Table I – the majority ofprograms offer ISO 2709, but the other formats are solely presented. Unfortunately,some of the studied programs which used ISO 2709 for their export format use theirspecific model (manipulated MARC tags with their specific codes) without releasing itsmap on the web. On the other hand, most of the companies use ISO 2709 export formatlocally. In fact, it is presumably thought that they do not offer the export format on theweb. So, it is off-line and requires specific knowledge which is practical with theirprofessional employees working in the company office. Apparently, just two of themcould have had off-line data exchange with each other by ISO 2709.

It should be noted that because of the neutral value of the software’s name in thisresearch, in the Tables which follow characteristic numbering is used. But, because ofthe value of the information in the tables to make a decision in the libraries, the name ofthe software are mentioned here. Thus, A is used for Parsazaraksh software; B is usedfor National Library software; C is used for Elm o Sanat University’s digital librarysoftware; D is used for Simorgh software that is produced by Nosa Software Company;E is used for Kavosh software which widely used by school and mosque libraries; andF is used for Mava software.

According to the different MARC formats – as presented in Table II – the majorityof programs use UNIMARC. One of the studied programs uses MARCXML, 2006 andalso another one claimed that the software is presenting MARC’s different formats, butunfortunately does not work properly at the time of the research (December 2006). Mostof the studied companies claimed that they technically have the ability to makedifferent export formats of MARC or any format libraries need, but because of thelibraries needless or lack of request they have not prepared the interface.

Others MARC XML XML Text Label Tag ISO 2709 Software

PDF – U U U U U A– UNI U U U U U B– – U – – – U C– – – – – – U D– – – – – – – E– – U – – U U F

Table I.Software export format

Interoperabilitybetween library

software

729

Most of the library programs use Z39.50 client to get information from Library ofCongress and Iranian National Library. Moreover, one of them uses Z39.50 client/serverin order to exchange bibliographic data with systems which have such ability. Therefore,none of Iranian library software could exchange – sending and receiving – data byZ39.50 protocol because of the lack of programs which have this ability.

To sum up, the studied products do not have facility to exchange data directly andon-line. However, as they claimed, most of them could technically make it feasible aswell as operational.

What is the solution?The current situation of library software in the fields of preservation, extraction andinteroperability has been studied to propose a practical solution. Therefore, technicalaspects, economic issues and professional abilities are arguably considered. Sincelibrary programs in Iran and any other countries in the world are produced by differentsoftware companies or organizations, each one built their software as well as databaseswith their own specific algorithm. However, most of them do not have a standard way todeposit their data in the databases. Then each solution needed to build changes in thearchitecture of databases would not be practical because of wasting a large amount ofmoney and time as well. In this manner, companies would not agree with the proposedsolutions which do not pay attention to their restrictions. As a result, the solution wouldbe encountered with company’s disagreements and finally will not be practical.

On the other hand, the proposed solution should be simple, not need much more timeand cost, and also coordinate with the world standards in order to be cooperative withother software. Therefore, inevitably, taking care of related standards and simplifyingthe solution must have essential principles. Also, the solution should be flexible andco-ordinate with technology development. In summary, the proposed solution should be:

. independent and do not need to make changes in the architecture of databases;

. cost-benefit and cost-efficient and also not need far more money and time;

. simple and practical;

. coordinate with the world standards; and

. flexible and coordinate with technology development.

In respect to the above issues in Iranian libraries and other libraries every where in theworld which have the same story, using MARCXML[1] in extracting bibliographic datais essential. MARCXML makes use of a XML Schema[2] designed to represent MARCrecords. The importance of XML appears unquestionable simply by the fact that it is aMeta language[3] capable of representing complex structures in non-proprietary and

Others USMARC MARC21 UNIMARC Iran MARC Software

– – – U – AMARCXML – – U – B– – – U – C– U U U U D– – – – – E– – – – – F

Table II.Different MARC formats

TEL26,5

730

self-explanatory ways. On the other hand, both XML documents and MARC recordsare tree-like structures, composed of elements that can repeat themselves and furtherbe divided into other elements (nodes in XML, fields, indicators, sub fields in MARC)(de Carvalho and Cordeiro, 2002, pp. 3-4).

In order to use MARCXML in exchanging data between library software, threeareas should be considered:

(1) record exchange (data transportation level);

(2) record validation (data conformity level); and

(3) sharing services (application services level), not only among library informationsystems (with a role similar to that of Z39.50) but also with a potentially verywide range of applications (de Carvalho and Cordeiro, 2002, p. 4).

Some current proposals in the area of XML representation of bibliographic records tryto introduce it such as ISO 2709, while, as we know, ISO 2709 is only used for dataexchange. ISO 2709 uses a leader part to locate and represent data in a text format. Itdoes not check the validity of MARC data. Hence, we can have a valid ISO 2709 recordthat contains a totally invalid MARC record. Valid means records that are fullyconformant with the structure and rules defined in data formats such as MARC21 orUNIMARC. In this perspective, a good XML format specification is one that does notrepresent an invalid record.

In this context, the concept of valid record is decomposed into three different levels:readable, correct and adequate. A readable record is composed of a leader, a set ofcontrol fields and a set of data fields with the additional characteristics defined by theISO 2709 standard. So ISO 2709 standard is just a readable format. A correct record is areadable record that contains the required set of fields prescribed by the MARC – inour project I mean UNIMARC – standard to model a given type of bibliographicitem – or full-text in digital libraries – and the content of which follows the relevantcoding rules and vocabulary types, wherever applicable. Finally, an adequate record isa readable and correct record that is fit for the purpose defined for that record in itscontext (de Carvalho et al., 2004, p. 133; de Carvalho and Cordeiro, 2002, pp. 4-5).

In a nutshell, library software should consider transport, validation and servicesareas so as to be interoperable. Up to this point, we just prepare metadata standardbased on MARCXML. Factually, the next challenge in order to be interoperable isfinding bibliographic data. As discussed above, libraries use different software withdifferent architecture, thus bibliographic data can be spread in a wide scale, makingsearching and finding them a hard job. This implies we should search library softwareseparately in order to find requested metadata. So, this task is a time and costconsuming job. In order to overcome this challenge, we should harvest bibliographicdata. In this model, a service provider access on OAI metadata repository, which isbased on MARCXML, via the OAI protocol (OAI is an increasingly popular webservice standard that enables repositories to make the resource metadata available toany organization that wishes to harvest it (Wusteman, 2006, p. 5)). In this manner,library programs prepare an OAI metadata repository. Thereby an OAI serviceprovider aggregates data by OAI protocol. In fact, OAI service providers crawlalternatively in OAI metadata repositories in order to find new items that newly addedto the system or up-to-dated. As a matter of fact, there is another challenge here tosearch into the harvested metadata. The OAI protocol does not address how that

Interoperabilitybetween library

software

731

database might be searched. That is where search and retrieval via the URL (SRU)would come in. The service provider would interface an SRU server to the database ofmetadata records of an SRU client to access (Figure 1).

Some may prefer to use Z39.50 inverse of OAI and SRU in their systems. In fact, thesuggested method is much less technically complex than other interoperability standards,such as Z39.50, in part because it does not rely on a distributed searching model (Prom,2003, p. 202). By the way, the current research suggests using OAI and SRU in order toexchange metadata between library software, because of its significant superiority.

Thus, in order to install such services libraries should consider the requirementsand specifications of Iranian library programs that are proposed in the following steps:

(1) Preparing a MARCXML Schema according to UNIMARC with respect to thementioned issues. As a matter of fact, using MARCXML Schema prepared inlibrary of congress’ web site is useful to do our best. It helps us to be interoperablewith other information systems all over the world. Actually, in this way, we couldimprove our systems according to the new standards and technologies.

(2) Preparing two converters to translate their data structures into MARCXMLaccording to MARCXML Schema and vice versa.

(3) Preparing OAI metadata repository for their bibliographic data.

(4) Preparing OAI service provider in order to aggregate metadata records by adominant organization such as a National Library.

(5) Preparing SRU server in order to make search facility for SRU clients by adominant organization.

(6) Respecting to OAI and SRU protocols by the dominant organization and librarysoftware in order to exchange data.

(7) Preparing suitable interfaces so as to make their systems able to exchange data.

In fact, users can search and retrieve data by SRU standards (Tenant, 1999, pp. 23-4;Taylor and Dickmeiss, 2005, p. 3). Then with prepared interface they can be able toextract enter the retrieved data into their own systems without any false drop.

ConclusionThe MARC standard for representing catalogue records and the Z39.50 standard forlocating and retrieving them have facilitated interoperability in the library domain

Figure 1.OAI/SRU interface

TEL26,5

732

for more than a decade. With the increasing ubiquity of XML, these standards arebeing superseded by MARCXML for record representation and SRW/U for searchingand retrieval (Taylor and Dickmeiss, 2005, p. 1). Library systems, inevitably, need to beinteroperable because of lowering cost and satisfying their users. Unfortunately,Iranian library programs do not take it into consideration. So, most of them have notprepared ability for data exchange. Therefore, using MARCXML standard to makethem interoperable, according to their current situation, is essential.

In order to make the proposed solution practical, we should have a control center –for instance the National Library – to manage the process and force the softwarecompanies to respect the rules. On the other hand, most of the Iranian librarians maynot have sufficient knowledge in this case, or may not have encountered such issues, ordo not know the ability of technology to solve as well as overcome it. In this regard,teaching them could be one of the significant steps to improve their knowledge in thediscussed issues.

Notes

1. Because of the Persian language’s specification and usage of UNIMARC in Iranian NationalLibrary software, the proposed MARCXML is based on UNIMARC.

2. XML Schema can be used to express a schema: a set of rules to which an XML documentmust conform in order to be considered “valid” according to that schema (Wikipedia, 2006).

3. Meta language is a language which describes another one. For example, verb, noun, adverb,and adjective are words which are in structure of language and are used for describing otherwords. The advantage of a Meta language is its usage for describing different languages(Ashbacher, 2000, p. 8).

References

Ashbacher, C. (2000), Sams Teach Yourself XML in 24 Hours, Sams, Indianapolis, IN.

Borgman, C.L. (1996), “Automation is the answer, but what is the question? Progress andprospects for Central and Eastern European libraries”, Journal of documentation, Vol. 52No. 3, pp. 252-92.

de Carvalho, J.R. and Cordeiro, M.I. (2002), “XML and bibliographic data: the TVS (Transport,Validation and Services) model”, paper presented in 68 IFLA council and generalconference, August 18-24.

de Carvalho, J.R. et al. (2004), “Meta-information about MARC: an XML framework for validation,explanation and help systems”, Library Hi Tech, Vol. 22 No. 2, pp. 131-7.

Gates, B. (2005), “Building software that is interoperable by design”, available at: www.microsoft.com/mscorp/execmail/2005/02-03interoperability.mspx

MARCXML (2006), available at: www.loc.gov/standards/marcxml

Prom, C.J. (2003), “Reengineering archival access through the AOI protocols”, Library Hi Tech,Vol. 21 No. 2, pp. 199-209.

Spice, P.B. (2001), “Key barriers to international resource sharing and OCLC actions to helpremove them”, Interlending & Document Supply, Vol. 29 No. 4, pp. 169-74.

Taylor, M. and Dickmeiss, A. (2005), “Delivering MARC/XML records from the library ofcongress catalogue using the open protocols SRW/U and Z39.50”, paper presented at WorldLibrary and Information Congress: 71st IFLA General Conference and Council, Oslo.

Tenant, R. (1999), XML in Libraries, Neal-Schuman Publishers, New York, NY.

Interoperabilitybetween library

software

733

Wikipedia (2006), available at: http://en.wikipedia.org/wiki/XML_Schema

Wusteman, J. (2006), “Realizing the potential of web services”, OCLC Systems & Services:International Digital Library Perspectives, Vol. 22 No. 1, pp. 5-9.

Further reading

ISO/IEC JTC 1 N7409 (2004), Comments Received on JTC 1 N 7335, Responses on SophiaResolution #39-Development of a Solution for the Unambiguous Identification and InterWorking of Codes Representing Countries, Language and Currencies (prepared on behalfof SC 32/WG 1), available at: http://std.dkuug.dk/jtc1/sc2/wg2/docs/n2733

About the authorMehdi Alipour Hafezi obtained a BA degree in Library and Information Sciences from TabrizUniversity in 1998, then received an MA in Library and Information Sciences from Azad Universityin 2002. He started his PhD course in Library and Information Sciences at the Science and Researchbranch of the Islamic Azad University in Tehran, Iran in 2005. He has more than seven years ofexperience in information services and systems gained while working in the Information Center ofIran Telecommunication Research Center and also as the Manager Digital Library projects and Headof the Electronic Library Department in the Atomic Energy Organization of Iran. He teaches as aVisiting Lecturer in the Department of Library and Information Science at Allame Tabataba’eeUniversity. He is also a member of the Iranian Library and Information Sciences Association. Hisresearch interests include digital libraries – particularly XML and its usage in digital libraries, andinteroperability between digital library systems. Mehdi Alipour Hafezi can be contacted at: [email protected]

TEL26,5

734

To purchase reprints of this article please e-mail: [email protected] visit our web site for further details: www.emeraldinsight.com/reprints