Authoring, Editing and Visualizing Compound Objects for Literary Scholarship Anna Gerber, Jane Hunter School of Information Technology and Electrical Engineering The University of Queensland {agerber, jane}@itee.uq.edu.au This paper presents LORE (Literature Object Re‐use and Exchange), a light‐weight tool designed to enable scholars and teachers of literature to author, edit and publish OAI‐ORE‐compliant compound information objects that encapsulate related digital resources and bibliographic re‐ cords. LORE provides a graphical user interface for creating, labelling and visualizing typed rela‐ tionships between individual objects using terms from a bibliographic ontology based on the IFLA FRBR. After creating a compound object, users can attach metadata and publish it to a re‐ pository (as an RDF graph) where it can be searched, retrieved, edited and re‐used by others. LORE has been developed in the context of the Australian Literature Resource project (AustLit) and hence focuses on compound objects for teaching and research within the Australian literary studies community. However it can easily be tailored to support the creation of compound ob‐ jects for literary and bibliographic research more generally. 1 Introduction Within the discipline of literature research and teaching, the ability to relate disparate digital re‐ sources in a standardized, machine‐readable format has the potential to add significant value to distributed collections of literary resources. Such compound objects can be used to track the li‐ neage of derivative works which are based on a common concept, to relate objects around a common theme, or to encapsulate related digital resources for teaching purposes. For example, one might want to relate the original edition of Follow the Rabbit-Proof Fence to the illustrated edition, a radio recording and a digital version of the film – and to retrieve and present these re‐ sources, with their relationships visualized, regardless of their location. Our objective is to pro‐ vide a software tool to enable such encapsulation and subsequent re‐use and visualization, by building on the efforts of two previous digital library initiatives: • The IFLA Functional Requirements for Bibliographic Records (IFLA, 1998) • The OAI‐Object Reuse and Exchange (OAI, 2008) FRBR is a recommendation of the International Federation of Library Associations and Institu‐ tions (IFLA) to restructure catalogue databases to reflect the conceptual structure of information resources. It uses an entity‐relationship model of metadata for bibliographic resources that sup‐ ports four levels of representation: work, expression, manifestation and item. It also supports three groups of entities: products of intellectual or artistic endeavour (publications); entities re‐ sponsible for intellectual or artistic content (a person or organisation); and entities that serve as subjects of intellectual or artistic endeavour (concept, object, event, and place).
13
Embed
Authoring, Editing and Visualizing Compound Objects for … · Authoring, Editing and Visualizing Compound Objects for Literary Scholarship Anna Gerber, Jane Hunter School of Information
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
ThispaperpresentsLORE(LiteratureObjectRe‐useandExchange),alight‐weighttooldesignedto enable scholars and teachers of literature to author, edit and publish OAI‐ORE‐compliantcompound informationobjects that encapsulate relateddigital resources andbibliographic re‐cords.LOREprovidesagraphicaluserinterfaceforcreating,labellingandvisualizingtypedrela‐tionships between individual objects using terms from a bibliographic ontology based on theIFLAFRBR.Aftercreatingacompoundobject,userscanattachmetadataandpublishittoare‐pository (as anRDFgraph)where it canbe searched, retrieved, editedand re‐usedbyothers.LOREhasbeendevelopedinthecontextoftheAustralianLiteratureResourceproject(AustLit)andhencefocusesoncompoundobjectsforteachingandresearchwithintheAustralianliterarystudiescommunity.However itcaneasilybe tailoredtosupport thecreationofcompoundob‐jectsforliteraryandbibliographicresearchmoregenerally.
1Introduction
Withinthedisciplineofliteratureresearchandteaching,theabilitytorelatedisparatedigitalre‐sourcesinastandardized,machine‐readableformathasthepotentialtoaddsignificantvaluetodistributedcollectionsofliteraryresources.Suchcompoundobjectscanbeusedtotracktheli‐neage of derivativeworkswhich are based on a common concept, to relate objects around acommontheme,ortoencapsulaterelateddigitalresourcesforteachingpurposes.Forexample,onemightwanttorelatetheoriginaleditionofFollowtheRabbitProofFencetothe illustratededition,aradiorecordingandadigitalversionofthefilm–andtoretrieveandpresentthesere‐sources,withtheirrelationshipsvisualized,regardlessoftheirlocation.Ourobjectiveistopro‐vide a software tool to enable suchencapsulation and subsequent re‐use andvisualization, bybuildingontheeffortsoftwopreviousdigitallibraryinitiatives:
FRBR isarecommendationof the InternationalFederationofLibraryAssociationsandInstitu‐tions(IFLA)torestructurecataloguedatabasestoreflecttheconceptualstructureofinformationresources.Itusesanentity‐relationshipmodelofmetadataforbibliographicresourcesthatsup‐ports four levels of representation:work, expression,manifestation and item. It also supportsthreegroupsofentities:productsofintellectualorartisticendeavour(publications);entitiesre‐sponsibleforintellectualorartisticcontent(apersonororganisation);andentitiesthatserveassubjectsofintellectualorartisticendeavour(concept,object,event,andplace).
TheOpenArchivesInitiativeObjectReuseandExchange(OAI‐ORE)isaninternationalcollabora‐tiveinitiative,focusingonaframeworkfortheexchangeofinformationaboutDigitalObjectsbe‐tween cooperating repositories, registries and services.OAI‐OREaims to support the creation,managementanddisseminationofthenewformsofcompositedigitalresourcesbeingproducedbyeResearchandtomaketheinformationwithintheseobjectsdiscoverable,machine‐readable,interoperableandreusable.ResourceMapsandtheircomponentresourcesareallwebresourceswhichcanbe identifiedandunambiguouslyreferencedbyHTTPURIhandles, thusprovidingabasisforreuseandexchange.NamedGraphs(Jeremyetal,2005)areendorsedbytheOAI‐OREinitiative as a means of publishing compound digital objects that clearly states their logicalboundaries (Lagoze,2007). Theydo this inaway that isdiscipline‐independent,but thatalsoprovideshookstoincluderichsemantics,metadata,ontologiesandrules.InthetermsoftheOAI‐ORE,compoundobjectscorrespond toOREAggregations, and theNamedGraphs thatdescribethemtoOREResourceMaps.
AustLitisacollaborationledbytheUniversityofQueensland,betweentwelveAustralianuniver‐sities,theNationalLibraryofAustraliaandtheAustralianResearchCouncil.ItprovidesthepeakresourceofbibliographicdataforscholarsundertakingresearchintoAustralianliteraryheritageand print culture. Within AustLit, research activities are undertaken by communities of re‐searchersfocusingonparticulartopics,regions,genresorofotherspecialinterests.AustLitcur‐rently supports sixteen different research communities, includingBlackWords, Children’s Lit‐erature,ExpatriateandPopularFiction.
TheAustLitdatamodelisbasedontheIFLAFRBR(Kilner,2005),makingitidealforevaluatingLORE. Figure 1 shows anAustLitwork recordwith correspondences to FRBR entities.WithinAustLit,eachworkrecordpresentsinformationrelatingtoaFRBRWork.ThetermsversionandpublicationareusedinterchangeablywithFRBR’sExpressionandManifestation.AustLitdoesnotrecordanydetailsaboutItems.
AuthoringandeditingofAustLitrecordsisrestrictedtoAustLitstaffandafewkeymembersoftheresearchsub‐communitieswhohavebeentrainedtousethecomplexdataentryinterface.Research communities cannot create their own additions or extensions to the datamodel torecord specialized research data – theymust request changes to bemade to the underlyingAustLitdatabaseontheirbehalf.AstheamountofspecializedresearchactivitywithinAustLithasincreased,theproliferationofadditionstothesharedunderlyingdatamodelhasincreased.ThishasincreasedthecomplexityoftheAustLituserinterface,makingitevenlessaccessibletoscholarswhohavenotbeentrainedinitsuse.
AnumberofpreviouseffortshaveappliedOAI‐OREtospecificscientificdisciplinestoencapsu‐late experimental data and results. These include: Foresite (2008), eChemistry (VanNoorden,2008),UIUC(Cole,2008)andSCOPE(Cheungetal,2007).AlthoughCULTOS(2003)usesRDFtorepresent multimedia and hypertext presentations for e‐Humanities applications, it does notcombineOAI‐OREandIFLA‐FRBRtocaptureorlabelthepreciserelationshipsbetweenentities.Also relevant isanoverviewofprevious implementationsandapplicationsof IFLAFRBR,pro‐videdbyBabeu(2008).
AlthoughasignificantpastfocusofeHumanitiestoolsdevelopmenthasbeenonscholarlymark‐upandannotation tools toattach interpretations to individualobjectsorpartsofobjects (e.g.,paragraphswithinanarticle)(Schreibmanetal,2004) .LOREtakestheannotationparadigmastepfurther,enablingauthorstoannotatelinksbetweenmultipleresourcesofmixedmediatype,withtagsfromanontology.
Previous efforts focusing on the visualization of OAI‐ORE compound objects have developednode and arc diagrams for visualizing Resource Maps and their evolution over time. Van DeSompel&Lagoze(2007)conductedanexperimentthatusedtheInternetArchive’sWaybackMa‐chinetoarchiveanddisplaychangestoResourceMapsovertime.ThisworkusesWebdotandGraphViztovisualizeResourceMapsontheWeb.
Figure 3 illustrates the display of an OAI‐ORE resource map within LORE’s graphical editor.Nodesrepresenttheindividualresourcesthatareaggregatedwithintheresourcemapandarcsrepresentthetypedrelationshipsbetweenthem.
conservescreenspace,as shown for the Jerilderienode in thecentreofFigure3,or resized todisplaymore content, as shown for the imageof the Jerilderie Letter on the right‐hand sideofFigure3.ThisallowsuserstoviewandinteractwithaggregatedresourcesdirectlyfromwithinLOREratherthanhavingtoloadthemindividuallyviathemainbrowserwindow.
Thefeaturesthattheresearchersrankedmosthighlyweretheinteractivenodepreviewsandthedirect integration of the editorwith the browser. Theywere also very enthusiastic about thegeneratedslideshowsandmultimediapresentations;howevertheywouldliketobeabletocus‐tomizetherulesusedtogeneratethevisualizations,andwouldalso liketobeabletogeneratemoreconventionalscholarlyanddocumentformats,suchasreferencelists.Ourtrialusersfoundaddingresourcestocompoundobjectstobeintuitive;howevertheyrequestedtheabilitytoaddresourcesinbatches,forexamplefromAustLitsearchresults,bookmarksorbrowserhistory.Whenusing theAustLitontologydirectly,users foundapplying theontology terms tobemorecomplexthanusingthesimplifiedrelationshipontology.However,thenumberofrelationshipsshowninthearccontextmenuforbothontologies isoverwhelmingformostusers.Presentingtherelationshipsinamenuthatusesthesub‐propertyhierarchyfromtheontologyratherthanas a flat list in alphabetical ordermay bemore intuitive. Alternatively, allowing users to typeahead,and tohave theUIonlyshowmatching termsmaybeeffective.Otherstrategies forad‐dressingthisissuecouldincludeaddingmoresemanticcheckstotheUItoassistusersinapply‐ing theontology terms,or tailoring thedomainontologiesbasedoncommunityneedsandun‐derstanding. Enabling discovery of additional ontologies via a metadata schema registry mayalsoassistindividualuserstolocateexistingontologiesthatbettersuittheirneeds.
Manyuserswereconcernedaboutcopyrightissues,astheyassumedthatwewerestoringcopiesoftheresourcesthatwereaddedtocompoundobjects.Theyrequestedmoreobviousattributionofthesourceofeachresourceinthepresentationformatsandtomakeitclearerthatweonlyre‐corda linktoanduser‐enteredmetadataforeachresource.Therequirementthatresourcestobeaddedtocompoundobjectsmustbeaccessibleonlinealsopresentedanissueduringthetri‐als.Toaddanewresource,usersmustfirstpublishit, forexamplebymanuallyuploadingittotheirinstitutionalrepository.Usershaverequestedthatthisprocessbemadesimpler,withtheability to upload new content directly fromwithin LORE. In addition, some objects that existwithininstitutionalrepositoriesonlyhavelocalidentifiers,andresourcesfoundonthewebthathaveURIsmaynothavepersistentURIs,whichwillresultinerrorswithresourcesfailingtoloadiftheyaremovedordeleted.WemayneedtoincorporateaservicetoassignpersistentURIstoobjectsthatdon’talreadyhavethem,andusearchivessuchastheInternetArchiveorPANDORA.
OREResourceMapsmayaggregateconceptsornon‐informationentities(asopposedtoinforma‐tionresources)providedtheyfollowtheprinciplesoflinkeddata,i.e.thattheyareaccessibleonthewebviadereferenceableURIs.WhenusingLOREtorelatethingslikeWorks,Agentsorotherabstract concepts from the AustLit thesaurus, the AustLit record and thesaurus pages can beconsideredtobeproxiesfortheunderlyingobjects.However,ExpressionsandManifestationsaredisplayedontheWebonlywithinthecontextofaWorkrecordpage,makingitdifficultforuserstoattachmetadataorrelationshipstothoseobjects.Also,becauseLOREusestheURIdisplayedinthebrowseraddressbar, it isalsopossibletoaggregateresourceswithessentiallythesamecontent via different URIs. Examples include the address of a work record with and without“www”atthestart,ortheURIofasearchthatreturnsthatworkasasingleresult.Thishasbeena sourceof confusion for someofourusersandmakesquerying the repository for relatedre‐sourcesmoredifficult.WearecurrentlyexperimentingwiththeuseofRDFaembeddedinAus‐tLitrecordsandsearchresultpagestoensurethatAustLitresourcesareidentifiedusingconsis‐tentURIs,andtomakeiteasierforouruserstoattachmetadataandrelationshipstoembeddedobjects.
7FutureWork
Tomakeiteasierforuserstoprovidenewcontenttoincludeincompoundobjects,weintendtoadduploadsupportviaSWORD(SWORD,2009)orbyusingtheFedoraAPIdirectly.WearealsoplanningtoprovidebettersupportforqueryingandexportingtoFedorausingtheRESTAPI,andare investigatingmappinga compoundobject toa singleFedoraobjectwithadata stream foreachcontainedresourceandtherelationshipsstoredintheRELS‐INTdatastream,ratherthanmultipleobjectswiththerelationshipsstoredinRELS‐EXT.