Design and Development of a Network-Based Electron

DesignandDevelopmentofaNetwork-BasedElectronicLibraryRayR.LarsonAssociateProfessorSchoolofLibraryandInformationStudiesUniversityofCalifornia,BerkeleyBerkeley,[e-mail]sedinnovationsintheClintonAdministration’splanstodevelopaNationalInformationInfrastructureisthecreationof,andsupportfor,digitalorelectroniclibrariestostoreandprovideaccesstothevastamountsofinformationexpectedtomadeavailableoverthe\informationsuperhighway.Althoughtheexactnatureandfuturearchitectureofsuchlibrariesisstillamatterforexperimentation(anddebate),thereareseveralpioneeringeortsunderwaytoestablishelectroniclibrariesandtoprovideaccesstothem.ThispaperdescribesonesucheortunderwayattheUniversityofCaliforniaatBerkeley.IncollaborationwithfourotheruniversitieswearedevelopinginteroperableelectroniclibraryserverscontainingtheComputerSciencetechnicalreportsforeachparticipantandmakingthemavailableovertheinternetusingstandardprotocols.1IntroductionThispaperdescribesthedesignanddevelopmentofanElectronicLibraryforComputerSciencetechnicalreportliteraturecurrentlyunderdevelopmentattheUniversityofCaliforniaatBerke-ley.TheplannedlibraryistheresultofacollaborativeresearchprojectinvolvingUCBerkeley,Stanford,Cornell,MIT,andCarnegie-MellonUniversitysponsoredbyCNRI(CorporationforNa-tionalResearchInitiatives)andARPA.ThisprojecthastheinitialgoalofputtingthefulltextandscannedpageimagesofallComputerSciencetechnicalreportsforeachparticipantintolo-callymountednetwork-basedserverswithacommonretrievalprotocol.ThispaperconcentratesonthecontributionsanddevelopmenteortsatUCBerkeleyinthiseorttodevelopanelectroniclibraryanddoesnotnecessarilyreecttheviewsofCNRIortheotherparticipants.TwoareasofthedesignanddevelopmentfortheComputerScienceTechnicalReport(CSTR)electroniclibraryTheworkdescribedinthispaperwassponsoredbytheAdvancedResearchProjectsAgency(ARPA)oftheDepartmentofDefenseunderGrantNo.MDA972-92-J-1029withtheCorporationforNationalResearchInitiatives(CNRI).Theviewsandconclusionscontainedinthispaperarethoseoftheauthorandshouldnotbeinterpretedasnecessarilyrepresentingtheocialpoliciesorendorsement,eitherexpressedorimplied,ofARPA,theU.S.Govern-mentorCNRI.AversionofthistechnicalreportwaspublishedinNavigatingtheNetworks:ProceedingsoftheASISMid-YearMeetingPortland,Oregon,May21-25,19941willbediscussed.Therstistheconceptualmodeloftheelectroniclibrarythatisdrivingthedevelopmenteorts,andhowthatconceptionisbeingturnedintoaworkingelectroniclibrary.Thesecondareaisadiscussionofthelongertermbasicresearchissuesthatwillneedtobeaddressedaselectroniclibrariesproliferateinaglobalnetworkenvironment.TheBerkeleygroupworkingontheCSTRprojectincludesfacultyandgraduatestudentsintheSchoolofLibraryandInformationStudiesandtheComputerScienceDepartment1.ThefollowingsectionsofthispaperwillthedescribetheviewofelectroniclibrariesthathasbeendrivingtheworkatBerkeleyontheCSTRprojectandhowthatviewtranslatesintoanarchitectureforfutureelectroniclibraries.ThiswillbefollowedbyabriefdiscussionofthecurrentandplannedarchitectureoftheBerkeleyserver.TheindexingandretrievalmethodsandthedatabasemanagementsystemtechnologyunderlyingtheBerkeleyserverwillbeexamined.Finally,thepaperwilldescribesomeoftheresearchissuesinthedesignanddevelopmentofelectroniclibrariesandfutureresearchagendaoftheCSTRprojectandotherelectroniclibraryprojectsatBerkeley.2DeningtheElectronicLibraryThedesignanddevelopmentoftheBerkeleyCSTRelectroniclibraryhasbeenbasedonaparticularvisionofelectroniclibrariesandhowtheywilldevelopandproliferateinthefuture.Thisvisionassumesthattherewillbemanysitesthatwillwanttocontributematerialtoandaccessmaterialin\TheLibrary.Thislibrarywillbeaglobalvirtuallibrary,thetraditionalcollectionmodelofgatheringallinformationpossibleintooneplaceisnolongertenable,ordesirable.Weenvisionavastpopulationofusersscatteredaroundtheglobewhoareabletoaccess,easilyandconveniently,thecompletecontentsofthousandsoflargeandsmallrepositoriescontainingtexts,images,soundrecordings,videos,maps,scienticandbusinessdata,aswellashypermediacombinationsoftheseelements.Thelibrarymust,therefore,beanetwork-baseddistributedsystemwithlocalserversresponsibleformaintainingindividualcollectionsofdigitaldocumentsrangingfromsetsofelectronictextstovideo-on-demandservices.Ineect,thisvirtuallibrarywillactuallyconsistofasetofpublishersofelectronicinformationandasetofconsumersdistributedacrossthenetwork.The\gluethatholdstogetherthisdistributedlibrarywillbeconformancetoasetofstandardsfordocumentdescriptionandrepresentation,andasetofcommunicationprotocols.Webelievethattheuseofmultiplestandardsforbothdocumentdescriptionandrepresentationandforcom-municationareinevitableinthenearterm.Foreventualstandardizationofdocumentdescriptionandrepresentation,webelievethatSGMLwillprovidethebasisfortextandcompounddocumentarchitectures(supportedbyadditionalstandardsforimage,video,andcompounddocuments).Forcommunicationswebelievethatthestandardprotocol,forlow-levelquery/responseanddocumentdelivery,willbesomeextendedversionoftheANSIZ39.50inf