Rencana-Rencana/ ArticlesThe MALMARC System - A Participant's Viewpoint

by UML Cataloguers

Abstract: The Malaysian Machine Readable Catalogue (MALMARC) Project is a pioneeringeffort at mechanised co-operative cataloguing by the National Library and major academiclibraries of Malaysia. Universiti Sains Malaysia Library functioned as the Co-ordinating andProcessing Centre. As a participant of the Project the cataloguers of University of MalayaLibrary faced certain unique problems stemming mainly from the lack of centralised cataloguingin the library system. The solving of these difficulties, the benefits derived from and thelimitations of the MALMARC Project are discussed.

Abstrak: Projek MALMAAC merupakan usaha perintis untuk mengkataJog dengan menggunakankomputer dan secara bersama oleh Perpustakaan Negara dan perpustakaan-perpustakaanakademik yang utama di Malaysia. Perpustakaan Universiti Sains Malaysia telah berfungsisebagai pusat penyelerasan dan pemerosesan. Sebagai peserta di dalam projek ini,pengkatalog-pengkatalog di Perpustakaan Universiti Malaya telah menghadapi beberapamasalah unik yang timbul kerana ketiadaan pengkatalogan secara berpusat dalam sistemperpustakaan. Usaha-usaha untuk mengatasi kerumitan-kerumitan ini; manfaat yang diperolehidan batasan-batasan Projek MALMARC dibincangkan di dalam makalah ini.

Background

The MALMARC (MALaysian-MAchine-Reada-ble-Catalogue) System is a computerised co-operative cataloguing system initiated andoriginally funded by UNESCO as a pilot projectto design a library network for university librariesin an Asian country. Malaysia's bid to host theproject was approved by UNESCO in June 1975.A consultant, Dr. Bjorn Tell was appointed tomake a "survey to evaluate the present state ofthe university library network of Malaysia. Theprimary purpose of that survey was to preparerecommendatIons which could bring aboutimprovements and be used in the preparation ofa long-term plan for the development of a univer-sity library network. "1 The resultant favourablereport recommended inter alia the use of MARCtapes to speed up the cataloguing process andavoid duplicate cataloguing of materials amongacademic libraries in Malaysia. Dr. Tell, had in thecourse of his survey evinced that a high percen-tage of acquisitions among the libraries werepublications in the English medium and co-operative usage of UK and LC MARC tapeswould be advantageous.

As a follow up to the above report, anotherconsultant, Stephen W. Massil, from the Universityof Birmingham Libraries Co-operative Mechaniza-tion Project was appointed by UNESCO "toconduct a feasibility study for the National Library

Kekal Abadi 9(1) Mac 1990

and the University Libraries on the use of MARCtapes for co-operative processing, the productionof catalogue cards or catalogue data on microfilmand access to the MARC database. The MARCdatabase at the British Library shall be used forpurposes of the feasibility study as a follow upto the conclusions of Bjorn Tell's mission toMalaysia. "2 Massil concluded that it was feasibleand economically viable to use MARC tapes inMalaysia and on his recommendation the Uni-versiti Sains Malaysia Library (USML) and theComputer Centre at that University were madethe Co-ordinating and Processing Centre forthe MALMARC Project. The USM librarians areto be lauded for their enterprising spirit in meetingthe challenge in undertaking this Project andcarrying it to fruition. Their tenacity and resource-fulness in catering to the needs of and balancingthe inter-organisational relationships between theMALMARC Consortium members is apparent fromwritings of their experiences in journals and inpapers presented at conferences.'

However, to date no single cataloguer of theConsortium has chronicled her experiences as acontributor to the pioneering effort of setting upa mechanised union database. In an attempt tofill this lacuna the cataloguers of the Universityof Malaya Library (UML) herein record how theycoped in switching from manual to computerisedcataloguing.

UML's Entry into the Consortium

The UML participated with all the localacademic libraries and the National Library Inthe pilot experimental project conducted by USMLin 1978 to test the viability and cost-effectivenessof the MALMARC project. However when thesystem became operational In 1979 UML did notcommit itself forthwith until a policy decision inlate 1981 was made to join the MALMARC Con-sortium and thus launched the Cataloguing Divi-sion Into the technological field in January 1982.

A short training stint was arranged at USMLfor the cataloguers to familiarise them with thecodes in the MALMARC manual and the opera-tional sequences. On their return the cataloguersdrew up a detailed workprocess manual and aflowchart to train the clerical and other supportstaff:' The staff had to re-orientate and adjusttheir traditional cataloguing practices drawn upto suit their clientele -- the doyen of the academicinstitutions - to accommodate procedural changesmooted by a sister Institution -- a younger sister atthat! UMSL's Chief Librarian Intuitively voicedthis dilemma thus: "One of the difficulties of evolv-ing a suitable co-operative structure stems fromthe fact that academic libraries have always pridedthemselves on their autonomy, and librarians havefor a long time sought to provide comprehenSiveservices to their clientele with minimal relianceon outside institutions. To design a new organisa-tional structure which removes some of the auto-nomy from individual institutions is very difficult". 5

However once the hand had been put to theplough the UML geared itself to the MALMARCoperation made up of the fallowing components:

I. Selection of records from MARC tapes.

II. Creation of original MARC records.

iiI. Production of various types of librarycatalogues In COM and special listings.

IV.Updating and amending records In theMALMARC database.

The ups and downs experienced by the cata-loguers in respect of each of these components arediscussed In turn.

2

I. Selection of Records

In the manual method of cataloguing at UML,the search clerks identified the entry for an item inhand from the Library of Congress Catalogs, Brit-Ish National Bibliography or other bibliographictools, copied the data on slips and passed theseto the cataloguers for editing. This process wastranslated in MALMARC to mechanical selectionof relevant entry from the UK or LC. MARC tapesby matching control numbers submitted by UMLto the processing centre at USML The controlnumbers were either the LC. card number, ISBN,Malaysian National Bibliography number (MNB),or for items not possessing any of these, a locallyassigned number Issued by USML and locatedfrom the union catalogue generated In respect ofthe MALMARC database holdings. Against eachcontrol number for a particular item its accessionnumber was also Included so that when a matchor "hit" was made the resultant printout (termeddiagnostic) came with the accession numberappended thereto for ease of matching.

Once the control numbers had been despatch-ed the books had to be stored strictly by accessionnumber order to facilitate matching with the diag-nostics when these arrived after a lapse, on anaverage, of three weeks. If an item was In themeantime on urgent request, it was manually cata-logued on a data input sheet and processed forthe reader. The data sheet was attached to adummy card and Inserted in the appropriateaccession number order to be used to edit therelevant diagnostic when it eventually arrived.

The batch processing procedure required eachbatch of diagnostics to be edited and returned enbloc. Colour coding of the batches prevented mix-ups In the event of two or more batches arrivingsimultaneously. In the editing of the diagnosticsUML cataloguers faced certain unique problemsdue to the lack of centralised cataloguing at theMain Library. The branch libraries located at theLaw Faculty, Medical Faculty, Institute of AdvancedStudies and the Academy of Islam In Kota Bharucatalogued their materials Independently and elect-ed to send control numbers directly to USML tocontinue the practice of separate processing.Classification schemes too differed. Whilst theMain Library adhered to the Library of Congressclassification, the Law Library adopted the Los

Angeles County Law Library K classification andthe Medical Library followed the National Libraryof Medicine (NLM) scheme. USML managed toaccommodate these variations by creating repeatlevels at the call number tag (090) except for theNLM call number which was provided for in theoriginal input format at tag (096). Then therewas the need to distinguish the location of theUML holdings in the master institutional COMcatalogue which replaced the individual card ca-talogues maintained by the component libraries.The cataloguers solved this problem by ascribingto each component library a unique number to beused as the first digit of the nine digit accessionnumber. The distinguishing digits were:

o - Main Library1 - National Collection (a special collection

with closed accesswithin the MainLibrary)2 - Institute of Advanced Studies Library3 - Academy of Islam Library4 - Law Library5 - Medical Library

Thus each entry in the catalogue indicatedvia accession numbers the number of copiesavailable In the UML system and the location ofeach copy by the unique first digit.

For reader education purposes instructionalarticles were published describing the manner ofusing the fiche reader, the format of each entryin the COM catalogue, (including the locationalcodes), and changes In personal and corporateheadings effected In conformity with the MCR IIcataloguing rules."

Beforetheediteddiagnostics weredespatchedto the Processing Centre a card for each itemwith abbreviated data entry was typed and filedas an in-process catalogue. Each in-processcard contained the control number, call numberandmainentry data up to publication level informa-tion (Tag 260) plus the accession numbers. Theoriginal Intent was for the In-process catalogue toserve as a temporary record until receipt ofCOM catalogues from USML and the cards tobe weeded out after checking against the COMcatalogues to verify accuracy and comprehensiveinput. However, the unexpected time lapse of 4to 5 months from the despatch of records toUSML and the appearance thereof In the COM

catalogues made the In-process catalogue an in-valuable tool for checking of additional copies,dealing with reader enquiries on the availability ofrecert publicatbns ard other tlbliographc searches.The decision was then taken to maintain the In-process catalogue as a permanent feature untilUML goes on-line and it would then serve as themaster list to verify if the MALMARC databasecontains all the records input to date.

II. Creation of Original MARC Records

In every batch of control numbers submittedto USML for selection of MARC records theaverage hit rate was 50% and 80% for UK and LCrecords respectively. The itemsfor which no matchwas available were weeded out for manual cata-loguing on prescribed MALMARCdata Inputsheetsfor creation of original MARCrecords. In the earlystages these input sheets were despatched bypost to USML to be keypunched into the data-base. In March 1986 UML acquired micro-computers and thereafter all original cataloguingwas input into diskettes and sent to USML fordownloading. For Malaysian imprints withoutISBNs the Input sheets had first to be sent to theNational Library for allocation of control numbersresulting In delays in Inputting of Malaysianamaterials. At times Malaysiana Imprints of varyingtitles were found to have identical ISBNs. Thisarose when publishers who had requested forand been allocated ISBNs prior to publishing awork erroneously printed the numbers or causedmix-ups. UMSL returned these Input sheets forcorrection of control numbers and further delaysensued in seeking to verify and correct the error.

Ill. Production of Catalogues inCOM and Speciallistings

The end product of cataloguing is the librarycatalogues which serve as the key to the Institu-tion's collection. In the manual method of produc-ing card catalogues there was undue delay fromthe time an itemwas processed to the appearenceof the author/title/series and subject cards Inthe respective catalogue cabinets due to tedioussorting and filing of copious cards. When UMLcommitted itself to the MALMARC Project it wasenvisaged that the speedier production of cata-logues would enhance collection exploitation. Ini-tially there were monthly paper printouts of theInstitutions's Input superseded by cumulative

3

quarterly COM catalogues culminating in annualcumulations.

However, within a year the monthly printoutswere discontinued to curb costs. Finally the COMoutputs for UMl consisted of:

a) COM divided catalogues of author/title andsubject plus the shelf list. The branch librariesof law, Institute of Advanced Studies andAcademy of Islam received in addition tothe foregoing an extra shelf-list of exclusivelytheir own holdings. These fiche catalogueswere quarterly issues with annual cumulationsleading in due course to quinquennialcumulations.

b) A monthly printout listing items processed bythe Centre for UMl for the current month.This list was for checking against and weed-ing out the in-process card catalogue butfor reasons explained above this step wasby-passed.

c) A monthly accession list printout from whichcopies were made for distribution to acade-mic and library staff. This list served in alimited manner to alleviate the lack of currencyof the quarterly COM catalogues. Howeverthe branch libraries continued to producetheir accessions lists manually from theirrespective in-process card catalogues.

d) A semi-annual printout of list of SoutheastAsian materials input by UMl. Copies of thisspecialised subject listing were distributed tointerested parties.

e) A union catalogue in COM of the entireholdings of the MAlMARC database embo-dying the total input by the participants of theConsortium. The frequency of issue followedthat of the institutional catalogues. The unioncatalogue is generated by main entry orderonly to curb expenditure but the lack ofsubject access is mitigated by USMl'swillingness to entertain requests for specialsubject searches as a fee-based service.This bibliographic output is the only toolcurrenty available for the dissemination ofsources of information in the nation's majoracademic and national libraries.

To some extent this union catalogue meetsthe shortfall highlighted by Massil that "a majorobstacle to the satisfactory sharing of resources

4

nationally is the lack of knowledge regardingthe availability of items rather than the inhibitingfactors of distance and delivery amongst thelibraries. The lack of an effective national unioncatalogue is thus a considerable drawback."

IV. Updating/Amending of Records

The need to update records in the MAlMARCdatabase arose when additional copies of a titlewere received or cataloguing data needed to beamended or deleted. In a card catalogue theseobjectives were immediately achieved by thesimple task of adding accession numbers of theadditional copies to the shelf-list or effecting therequisite data amendments to the relevant cards.However, with the holdings in a mechanisedMAlMARC database hundreds of kilometres fromUMl these operations proved quite tedious.

A prescribed form (Naskah tarnbahan/pinda-han) had to be filled in stipulating the controlnumber of the record to be updated or amended.Details of the author, title and call number had to becoded and the update/amendments had to bespecified at the appropriate tags and highlightedby circling in red. These forms could only besent to USMl for action after the record in questionhad appeared in the COM catalogue signifying itsexistence in UMl's institutional file.

Considerable time elapsed before the updated/amended record was sighted in the catalogue. Theintegrity of the COM catalogue suffered as userswould be referring to the unamended record in theinterim period. The in-process card cataloguehowever, reflected the changes made and proveda boon for quick reference by staff members.

The foregoing recounts the overall and generalexperience of the UMl cataloguers. The catalo-guers of the branch libraries and special collec-tions have expressed their special views inAppendix II.

MALMARC Database

I. Quality Control

In the managemeht of the MAlMARC databaseUSMldid not attempt to exercise quality control ofthe data input into the system. Every member of the

Consortium was expected to be responsible forthe integrity of its institutional input. Problemsarose in this arena as voiced by USML's ChiefLibrarian: "A final problem that we are facing atthe macro level is the difficulty of getting parti-cipants to be more actively committed to thesystem, and be more careful about the quality ofthe data input into the system. Within each of theinstitutions, the remoteness of the system makesit hard for the staff to feel committed to it. Thereis also the feeling that any errors in the data madewill be picked up during the processing stage inthe Universiti Sains Malaysla"."

As a result the contributed cataloguing byConsortium members created varying entry formatfor a single publication. Differing main entry andcataloguing practice resulted in multiple entries ofthe same item. For example, a publication ofseminar proceedings of the Urban Transport andEnvironment Seminar organized by OECD in 1979input by two institutions appears in the unioncatalogue thus:

a) One entry by title: Urban transport and theenvironment seminar, 1979. 10-12-July 1979/organised by OECD ...

b) Second entry as multi volume work catalo-gued with main entry by corporate name:

i) Urban Transport and the EnvironmentSeminar (1979 : Paris).Case studies.

ii) Urban Transport and the EnvironmentSeminar (1979 : Paris).Volume 1//: overview.

iii) Urban Transpot1and the Environment Seminar(1979 : Paris).Volume N: conclusions

In fact USML only controlled duplicate inputsby an institution vide a special program to checkand eliminate duplicate control numbers within abatch submitted for selection of records. By thisprogram duplicates were eliminated and the re-spective accesssion numbers were added to therelevant record automatically. In this context UMLexperienced a fair amount of duplicate numbersdue to the aforementioned lack of centralisedcataloguing. Inevitably overlaps occurred in ac-quisitions by the Main Library and the speciallibraries. Publications on forensic medicine wouldbe taken by the Law and Medical libraries, Islamic

law by the Law Library and Academy of IslamLibrary and output on sociological aspects of dis-ciplines of law, medicine and economics mustevidently be duplicated in the holdings of theMain Library and the respective branches. Inde-pendent remission of control numbers to USML bythe branches led to duplication of input. In suchduplicate records, if discrepancies occurred inchoice of main entry, call number or subjecttracings, USML printed out these records forverification by UML as dropped records. Consi-derable professional time was expended inre-checking and re-editing these dropped records.Ultimately what was gained in the swings ofspeed in inputting of data seemed to be lost inthe roundabouts of re-editing dropped records.

II. Authority Control

To enhance exploitation of UML's premierholdings, the cataloguers made every effort tomaintain strict authority control of personal, corpo-rate and series entries to ensure collocation ofpublications emanating from a particular source. Inthe manual method the main author/title/seriescard catalogue served as the chief source forauthority checking. In MALMARC the correspond-ing COM catalogue could not asssume the samerole due to its lack of currency. Again the UMLcataloguers, in addition to despatching an author-ity input form to the USML Processing Centre,resorted to a temporary authority card cataloguewhich now runs into 120 catalogue drawers. Thiswill serve as a checklist against the authorityrecords input into MALMARC when UML even-tually sets up its own automated library system.

Conclusion

The MALMARC database is the product of anexperiment in co-operative cataloguing. Currentlyit is the only source available on the holdings ofthe National Library of Malaysia and the majoracademic libraries in the country. The unioncatalogue of the MALMARC database for theoperational years of 1979 to 1989 runs into 95fiches which embody bibliographic details of over500,000 records and could well form the nucleus ofthe ultimate goal of a national information networkbased upon the resources of the nation's uniqueand expanding collections.

5

UML's participation has produced a machinereadable database of over 100,000 items acquiredover the last decade ready for reader access whenthe computerised library network system is launch-ed this year.9 The readers, after some initialreluctance, have become familiar with and gainedconfidence in accessing the COM catalogues.Visions of closing the card catalogue totallywith the advent of MALMARC has not materialisedbut the experience gained Is immeasurable. Asproblems faced by a new undertaking can alwaysbe listed more readily after the fact than they canbe predicated In advance, participation in theMALMARC system has enlightened and preparedthe UML cataloguers for cataloguing online.

References

1. Tell, Bjorn. Pilot Project on the Developmentof a Library Network. Paris: Unesco, 1976.

2. Massil, S.W. Study of the Feasibility of UsingMARC Tapes for Co-operative Processing.Paris: Unesco, 1977.

3. (a) Lim, Huck Tee. The Malaysian MARC(MALMARC) project. Program 14 (1980) :101-120.

(b) .... 'Choosing the moment: a review of theorganisational problems and changes arisingout of conversion to computerised systems'in UnAtersityLibraries in Developing Courtries

6

IFLAjUNESCO Pre-Session Seminar forLibrarians from Developing Countries,Munchen August 10-19, 1983. Munchen :K.G. Saur, 1985. p.119-134.

(c) Lim, Chee Hong. MALMARC in relation toAACR2. Paper presented at LASjPPM JointCataloguing Seminar, Singapore 1980.

(d) ..... The experience In operation of theMALMARC system. Paper presented at the6th Congress of Southeast Asian Librarians.Singapore, 1983.

4. See Appendix I

5. Lim, Huck Tee. Choosing the moment (supra 3(b)) at p.128-129

6. (a) Chuah, Molly. Bentuk tajuk nama :peruba-han-perubahan di dalam katalog-katalogPerpustakaan Universiti Malaya Kekal Abadi2(1) (Mac 1983) : 3-10

(b) Selvaratnam, M. Entri katalog COM: setengahciri baru. Kekal Abadi 2(1) (Mac 1983) :11-13

7. Massil, S.W. Study of the Feasibility of UsingMARC Tapes for ·Co-operative Processing(supra 2) at p.9.

8. Lim Huck Tee. Choosing the moment (supra3(b)) at p.129.

9. See Appendix III

Appendix I

ClLRK

CLERK

CAIALOGUER

A'I"nlIIlA!II

NEW TIILES

FLOW CHART FOR COMPUTERISED CATALOGUING

Collect alternativecontrol numbers

No alternativecout.r o l numbers d'Rush' itemsCheck NUC

Original cataloguingon to input .heeta

1

CLERKCollect Control

Ilumbers

nc ep at.ch ControlnU01ucrs tiy Telex

r------- ----.,, :Processing

, Centre .

i. .. _... _. __. _. j

Receipt & sorting ofd iaguos t icpr inlouts

Eo it d iagnos t ieprintoulS

1CheclLi.ng of edited diagnostic .h(!:.ts

& original input sheets

Filin& 0,(

colorcJ. vorkalipain Ln-pr oce •• Ll Le

Weeding outDC workll ip.

(wr .Dnth. laler

DC"l'lIlch oC,lia&u08lic nudinpul shccl!J

rccce •• .iug Ce ut r e

L__.....__._ .... __...__.. J

1OUT1'UT

1. Monthly l'npc r l'rinloul

2. QuarleLoly talccaliche.copie,

4. ACcc,sLon list

ADDITIONAL COPIES

CLERK

Monographs; t it Ie snot on HARC

Honogral.hs & Scrialsi i t Le s a Lr.e ady on MARC

(-

I) Check shelf list2) Obtain conLrol no.J) Note ac c , no. or

previous cop ie,

IYl'lSI

ObLain Controlno . [rom HARC

calalo&uc

Fill N:t.&ko.hto.rob3hau Iorsn

CL~){K

CATALOCUER

curer CATALOClJER

TYl'lSI

.... Typing I) bookcarda,-

2) croaa-re!erence.

-l--1

Ir i1. rerer.nee

CLERKin authodtr'tll ..

Check t yp Iug o( Look.c"rua

"T

nouk Ie t t.e r iugArTENIWI1'

1Check Le t t e r in& o! I) ap Lue

ClLlUC. 2) book. poc:k.et

.J...

D 0 0 k S t A C k • ArTENDANl'

~ -4

.-,

Ca la Iogu LngC'Lr cu Lat.Lon

~ ~. Cataloguing D~v.Circulation

Dilllriuucion

Serial title.not on MARC

L

Retc i.eve SheUlilt & ent.er new..cce s s Lcn no.,

as done formerly

7

Appendix II

SPECIAL VIEWS

National Collection Division

The National Collection Division has been cata-loguing and classifying all materials in the Malayand Indonesian languages acquired at the MainLibrary. Since September 1988, all Malaysianamaterials with the exception of those in Chineseand Tamil, have been processed in the Division.

The Division has noticed typographical errorswhen data sheets were sent to USML for inputthere. These errors may have arisen from illegiblehandwriting of cataloguers. However the situationwas rectified ever since cataloguing data has beeninput in-house, computer printouts proof-read andcorrected, and the diskettes despatched to USMfor downloading.

Computer assisted cataloguing gives catalogu-ers quite a free hand to provide as many addedentries and subject headings in the tracings with-out the nagging worry of too many cards to beprinted and overtyped for the traditional dictionarycard catalogue.

Cataloguers have taken advantage of thissituation and have practised in-depth catalogu-ing. Multiple subject tracings, added author entriesfor the fourth author or editor of a multi-authorwork, iUustrators, photographers, translators of non-fiction; foreword writers; etc. have been input.

Analytical added entries for composite workshave been made under the contributors of chap-ters.

All in all the catalogue as it is and in the futurewill be a rich source of information on Malaysiana,Malaysian writers and scholars and their output.

MEDICAL LIBRARY

In 1982, when the Medical Library first partici-pated in the MALMARC Project, the percentage ofhits (43.7%) was less than half of the titles sentand of the hits obtained, almost all were for titleswith LC control numbers. Hits for ISBN numbers

8

were negligible. To ensure therefore that medicaltitles would have a higher hit rate, it was decidedthat only LC control numbers would be sent.Titles with ISBN numbers would be processed asoriginal cataloguing. The wisdom of this movewas borne out by higher percentages of hits(62%-84%) for 1983 and subsequent years.

A survey of the MALMARC operations in theyears 1982-88 showed that the total number oftitles processed by original cataloguing (2994)was almost equal tothe nurnber otritles processedthrough diagnostics. Seen in this light, participa-tion in MALMARC did not seem to be of greatbenefit or advantage to the Medical Library.

The main problems faced by the MedicalLibrary with regard to MALMARC were connectedwith the delay in the receipt of diagnostics andCOM catalogues. On the average, it took 1-2months for the receipt cI diagnostics and 8-9 monthsfor data to appear in the COM catalogues. This inturn resulted in a delay in providing up-to-date in-formation about the collection.

To overcome this problem the Medical Libraryhad to resort to cards typed for each title inputin MALMARC as well as to adopt a differentcataloguing procedure from that practised in theMain Library. While In the Main Library, controlnumbers were sent before books were catalogued,in the Medical Library, books were cataloguedand sent out to the shelves at the same time thecontrol numbers were sent to USML.

This meant that by the time the diagnosticswere received, the books would already havebeen in circulation for a few weeks. This waspossible because most of the medical books withLC control numbers had CIP data which werereliable. Very little editing and amendments had tobe made to the CIP data. As such, sending controlnumbers to MALMARC served as a means ofverifying data already obtained from CIP entries.

In this way, the Medical Libarary was able tomake available its collection in the shortest timepossible and indirectly, overcome the problem ofdelay.

Speedy circulation of recent acquisitions Isimportant to the Medical Library because Itscollection is much smaller than the Main Libraryand the currency of medical information Is ofutmost importance to the medical fraternity.

Periodicals library

1. The inflow of serial titles is not a continuousprocess. It is a slow sporadic one, and ittakes a few months before the requisite 50titles could be cumulated for inputting Incompliance with UMSL's indication thatdiskettes of data sent for processsing shouldpreferably have a minimum of 50 records.This coupled with the 4 to 5 month delay Inthe appearence of the titles In the COMcatalogues Impaired the dissemination ofserial holdings in UML. The maintenance ofa manual visible index of current acquisitionsand the production of an in-house computerprintout of holdings provided Interim findingtools for reference purposes.

2. The above mentioned delay In the sightingof serial titles in the COM catalogues alsohindered the updating of records. On receiptof subsequent volumes of a serial title therequlslte data coded on the prescribed form(Naskah Tambahan) had to be withheld untilthe existence of the relevant record in UML'sinstitutional file was assured. Information onavailability of latest issues had to be verifiedfor readers from the visible index.

3. The imposition of 15 character limitation Inthe recording of accession numbers at Tag999 In the MALMARC Input codes posedgreat problems in serial cataloguing. Unlikemonographs, for serials the nine digit accessionnumber had to be qualified by the volumenumber, year of issue, plus, where pertinent,notation that the issue represented supple-ment, Index, special jubilee production orbiennial or triennial output. Inclusion of a"these details Invariably exceeded the 15character limitation. Hence essential holdingsdata had to be curtailed to contain thisproblem. Thus important information washidden and at times quite misleading.

4. Lack of centralised cataloguing in UML andthe production of a master Institutional COMcatalogue of its holdings in the MALMARCdatabase highlighted conflicts In the cata-loguing of certain serial titles. Serially Issuedworks on a specific topic with distinctive titlesinput by the Main Library as monographicseries appeared at variance with the serialcataloguer's treatment thereof as a journaland cataiogulng by title with open entry.Quick remedial action was taken by preparing

serials/series authority cards indicating thecataloguing format edopted in respect cI fNerysuch title Input into the MALMARC database.

5. With the anticipated merging of records InPERPUNET (a union serials cataloguingproject) with the MALMARC database,tracing bibliographic Information on serialstitles, especially local publications which donot appear In New Serial Titles or CONSER,will be greatly facilitated. Cataloguing InMALMARC has Impressed upon UMLcataloguers the need to liaise with fellowcataloguers In the system to obviate conflictsIn processing of serials and contribute to amore refined database.

Institute of Advanced Studies library

The Institute of Advanced Studies Library(IASL), although a component of the Main Librarydecided to participate in the MALMARC Projectonly in 1983, a full year after the Main Library'sentry Into the Consortium. Since its acquisitionswere processed independently of the Main Library,control numbers were despatched and resultantdiagnostics edited In-house.

Initia"y when a large stock of books awaitedcataloguing control numbers submitted resultedin high hit rates (82% - 90%). In contrast,original cataloguing constituted only 32% of thetotal records added in 1986 and 27% In 1987.

However, in 1988 due to drastic budgetarycutback and resultant severe drop In new acqui-sitions it proved more expeditious to processthe books via original input than to call for hitrecords from USML due to the delay of almost amonth in the receipt of diagnostics. Hence in 1988the cataloguing input consisted of 72% originalcataloguing and 28% of diagnostic records. In1989 no diagnostics were called for and a" 590titles added to the database were original input ondiskettes despatched to USML for downloading.

By participating in MALMARC there was somesaving in clerical effort, particularly as the produc-tion of the microfiche catalogues did away withthe manual filing of cards. However, the intellec-tual effort required in determining form of entries,subject headings and classification numbers wasstl" an essential task required of every cataloguer.

9

Contrary to themisconception of many,MALMARCdid not do away with basic descriptive and analy-tical cataloguing tasks. MALMARCmerely set astandard machine-readable format for recordingbibliographic details which have first to be deter-.mined by cataloguers.

A shortcoming of the MALMARC systemdetracting from its usefulness is the batch modeof obtaining and processing hit records. For abranch library like the IASL where the stock issmall, to resort to calling for hit records in batchescaused undesirable delay in processing. It wasmuch faster to check the NUC or transcribeacceptable CIP data and keypunch these records.Moreover the hits for UK Marc records oftenrequired heavy editing and this overrode anysaving in time and effort that came withobtaining diagnostics.

IASL, like the Main Library, continued toproduce a main card and shelf list for immediatereference of users and library staff. In addition it'sholdings were reflected in the in-process cardcatalogue at the Main Library vide distinctivegreen cards with abbreviated data entry toobviate duplicate cataloguing. From the shelf-listcards quarterly accessions lists were producedmanually, keeping readers abreast of current ac-quisitions until the receipt of the COM catalogues.

In the long term the IASL's participation inthe MALMARC has enabled it to join in themainstreamof automated bibliographicand house-keeping activities. The titles added by IASLwill bedownloaded together with the bulk of the titlesfrom the Main Library and other branch librariesinto UML's soon to be acquired ATLAS system,thus enabling IASL to have access to as well as

10

contribute to the entire University's biblloqraphlcdatabase.

Academy of Islam Library

Academy of IslamLibrary (AIL) located in KotaSharu participated in the MALMARCProject as acomponent of UML, processing its acquisitionsindependently. It experienced a 69% hit rate forcontrol numbers submitted but the greater portionof AIL's input was via original cataloguing.

The major drawback of MALMARC for AILwas the inability of the System to accept input inthe Arabic or Jawi script. All such titles had tobe transliterated for input, while a card cataloguein the original script was maintained in situ forusers. At times records input never appeared inthe COM catalogue necessitating re-input ofdata. A fair amount of the diagnostics receivedneeded heavy editing thus belying the muchtouted speed of input theory.

Delay in receipt of the COM catalogueadversely affected use of the collection. AILrequested the issue of a separate shelf-list ofexclusively its holdings but when a stock-takewas done in 1989 from a computer printout ofthis shelf-list it was found to be corrupted byinclusion of call numbers and accession numbers.of other institutions.

On the whole however, the MALMARCsystemas a pioneering attempt at setting up a uniondatabase has afforded an avenue to gainexperience in automated cataloguing. Catalo-guers of participating institutions have realised theneed for standardizing bibliographic data inco-operative cataloguing.

co~

I N "'it C")

~ ~~

..- 0I,.._ ..-..- ~ 10.... C")

II

8 ~ N I <0 I ,.._ I~

,.._ I c:; I ~ I ..-~g} ..-

~ :isco 0 N co 0)~ C\i ,.._- I oi I ~ "'it I 10N ........ N ~I~ I8 ~ ~

~ :;l C")0 "'itco C") co

~ R N C")co 0 C") ..- "'it co ,.._ 10~~ ,.._- c? ~ :;l 0 ..-N ..- N

N....~

....N~

8 N ~ co~

<0 ,.._"'it

~ ,.._~ N 10 ,...

~ C") C")

~co co 0 ..- "'it ~ co 0

~10~ ~ N- eD ,.._- ,.._ ~ ~

,.._ ,.._.... ....~

.... 0) ..-N"0 ~C..N

~,.._ <0 a;CO 8

~10 C")~ co ..- N co ~ g} ..-.... co ~ 10 C\I_ 0) co ,.._ ,.._~ as "'it- ,_: ~ ,.._

~ co 0)c .... N .... "'itCD N ~CD ~~a;Jl 8 ..- co 0)

~..-

~i It)

~ ~ "'it It) ,.._

~~

~co 0 ..- <0 co

~~ C\i "'it- as ~ ,.._«I) ..-

«I) .-N ..-

~.... ..-CD ~U- 0.~ ;s,

-g «I) 0 m ~ ~..- ..-

~'E ~ 0~

co ~ ..- ,.._ ..- ..-CD co 0 C") 10 co 10~

"'it C") "'itQ, 0 ~ C\i C\i ..- ,_: ~ ..- N C") ,.._Q, u ....N

~.... "'it« e ~

"0C.. 0 <0 0) ..-

~~

~e M 0

,.._ ,.._~ M 0) 10 10

~co 0 N <0 "'it co ~ N "'it:::J ~ C\i C\i ,_: ~ ,.._

N 0 ,.._:: ....~

.... 10 <0"0 NC ~CDQ,)C.CD 0 10 ..- <0

~N- N 0 "'it 0 ,.._<0 N N ,.._ 0) <0 0) '*

0 co 0 "'it ..- N ,.._ co 10 C") co I/) ,.._"'it- ~ C\i C\i C\i ,_: ~ ..- ,.._ :is ... I/) C")

c ....N N .... <0 NCD

E ~ ~CD-Sfn

>-J:: -a;(,)c~I/) ~"E ......... s_I/) .D I/)0 CD

i Q)

~~ 1§ ~...~

g I/)

~ ~ "E~ 1~

0 I/)I: I: "0~ "E0 C) C) 0) I/)

0a I: I: C -g "E J:: ~0;: o~ 0c;; o~ >- 0 0) (,)~ I/) tU i ~ I: C U

Q) Q) Q)Qj~~ "3 I/) ~.D e g giO ~

... i "E a:u g .D c(~ 8:1 Q) ~

~

I/) ... ... ... a: '0 ~a. a. a.s_ e~

8'~ ... ... .2~ ... I/) c(C) 'i0 0 o I: a. ~

~CI: - - -0 Q) -- - - I/) co- l/) i - o~ ~

I: I: I: 1:"0 Q) ~ "E ~ o~tU Q) Q) Q) ...

~ Qj(,) E E E 0 E 8 0 :~ ~~ g CI)

0c;; >- >- >-~ >-(,) ~ -a; ~ ...tU tU tU tU ... ~c( a: fn 0 .... 0 CL '*CD CL CL CL

11