UC Davis Grant: Reinventing Cataloging: Models for the Future of Library Operations  Xiaoli Li Co-head of Content Support Services University of California Davis Library Canadian Linked Data Summit, Oct 24, 2016

Outline What Is BIBFRAME? What IS BIBFLOW? Roadmap (Data Flow) Discovery Information Flow Cataloging Data Flow If Authority Exists Cataloging Data Flow if No Authority Found

PART I: What Is BIBFRAME? BIBliographic FRAMEwork Initiative

What Is BIBFRAME? Library Congress’ project started in 2011 It’s the replacement for MARC can and will serve as an encoding standard for RDA and other content standards AND It leverages the current web technology (sematic web/Linked Data) and uses Resource Description Framework (RDF) modeling practice.

BIBFRAME 2.0 Model RDA Work RDA Expression RDA Manifesta-tion RDA Item

Goals for BIBFRAME Supply search engines with description in a form they can exploit Use/exploit linking traditional = textual, identifiers semantic technology = URIs URI – Uniform Resource Identifier MARC transition enable reuse of data from MARC provision of transformations to new models Extensibility to new and broader content Accommodate needs for different types of libraries New views of different types of metadata descriptive, authority, holdings coded data, classification data, subject data preservation, rights, technical, archival Reconsideration of the data related activities exchange, internal storage, input interfaces and technique Adapted from a presentation given by Sally McCallum, http://www.oclc.org/content/dam/oclc/events/2015/CI_SFPL_PPTS/BF-NCTPG-California.pdf

PART 2: What Is BIBFLOW? BIBframe + workFLOW

BIBFLOW Project Is a 2-year project of the UC Davis University Library and Zepheira, funded by Institute of Museum and Library Sciences (May 2014 – April 2016) Its official title is “Reinventing Cataloging: Models for the Future of Library Operations” Is a research project that will address questions like “What impact will adoption of BIBFRAME on technical services workflows in an academic library”? Its primary purpose is to understand ecosystem, test solutions, and provide a roadmap of how libraries can iteratively migrate to linked data without disrupting patron or business services.

Ruth Fischer and Rick Lugg, Study of the North American MARC Records Marketplace, http://www.loc.gov/bibliographic-future/news/MARC_Record_Marketplace_2009-10.pdf

This complexity leads to the inevitable conclusion that moving away from MARC into BIBFRAME represents an evolutionary leap for libraries and not a simple migration.

What Have We Done? Analyzed cataloging workflows https://www.library.ucdavis.edu/bibflow/category/workflow-2/ Tested pulling data from OCLC using BIBFRAME Scribe http://jarjar.lib.ucdavis.edu:8888/static/ Tested pulling data and URI from LC using BIBFRAME Scribe http://jarjar.lib.ucdavis.edu:8888/static/ Explored conversion of MARC records to BIBFRAME LC’s MARC to BIBFRAME Transformation Service http://bibframe.org/tools/transform/start Zepheira’s transformation tool https://linksmith.zepheira.com/training/ (password required) Developed a roadmap

PART III: Roadmap MARC to BIBFRAME

Roadmap: Primary Stakeholders Triplestore Human Discovery LD Editing Interface Machine Discovery Vendors ILs

Roadmap: Ecosystem (Hybrid: MARC + BIBFRAME DATA) Pull Authority Data Human Discovery Interface Push User Contributions SPARQL Endpoint Exchange Circulation Data Pull Graphs Triplestore Respond to Machine Queries Push Triples Push Triples Push Thin MARC Pull Schema OCLC Pull Triples ILS BIBFLOW:SCRIBE LD Cataloging DB Pull Authorities Catalog/ Ingest Marc with URI Authorities MARC URI Pull Push Native Operation or Ingest Script / Tool (All other transactions are via API / Endpoint) Pull MARC Records

1. Discovery Information Flow

Step 1: User Submits Search Query Discovery Interface User Submits Search Request SPARQL Endpoint Triplestore OCLC ILS BIBFLOW:SCRIBE LD Cataloging DB Authorities MARCURI Step 1: User Submits Search Query

Step 2: Query Results Returned as Thin Local Graph Discovery Interface SPARQL Endpoint Search Results Returned (Thin local Graph) Triplestore OCLC ILS BIBFLOW:SCRIBE LD Cataloging DB Authorities MARC URI Step 2: Query Results Returned as Thin Local Graph

Step 3: User Requests Detailed Graph for Item Discovery Interface User Submits Request for Item Detail SPARQL Endpoint Triplestore OCLC ILS BIBFLOW:SCRIBE LD Cataloging DB Authorities MARC URI Step 3: User Requests Detailed Graph for Item

Step 4: Complete Local Graph Returned Discovery Interface SPARQL Endpoint Full Local Graph Returned Triplestore OCLC ILS BIBFLOW:SCRIBE LD Cataloging DB Authorities MARCURI Step 4: Complete Local Graph Returned

Supplemented by OCLC Schema.org graph Prior to display, returned graph is supplement with OCLC Schema.org graph data using linked OCLC Work URI Discovery Interface SPARQL Endpoint Triplestore OCLC ILS BIBFLOW:SCRIBE LD Cataloging DB Authorities MARC URI Step 5: Graph Supplemented by OCLC Schema.org graph

Step 6: Graph Supplemented by Authority Graphs Prior to display, graph is supplemented with authority data linked via local and OCLC graph nodes Discovery Interface SPARQL Endpoint Triplestore OCLC ILS BIBFLOW:SCRIBE LD Cataloging DB Authorities MARCURI Step 6: Graph Supplemented by Authority Graphs

Graph Building (steps 4, 5 & 6) Triples in the rectangle forms a record view which is displayed to a user via discovery interface

Step 7: Holding and Availability Info Added Discovery Interface SPARQL Endpoint Graph data supplemented with holding and availability data from ILS Triplestore OCLC ILS BIBFLOW:SCRIBE LD Cataloging DB Authorities MARCURI Step 7: Holding and Availability Info Added

Step 8: Completed Graph Displayed to User Discovery Interface SPARQL Endpoint The detailed view is displayed to the user after the complete graph has been assembled from various sources. Note that the computer is capable of handling each of the steps involved in the process in fractions of a second, so the user does not experience any delay more than now. Triplestore OCLC ILS BIBFLOW:SCRIBE LD Cataloging DB Authorities MARC URI Step 8: Completed Graph Displayed to User

2. Cataloging Data Flow If Authority Exists

Step 1: Check to See if Record exists. New if No Discovery Interface SPARQL Endpoint Triplestore OCLC ILS BIBFLOW:SCRIBE LD Cataloging DB Authorities MARC URI Step 1: Check to See if Record exists. New if No

Step 2: Pull Schema Data from OCLC Discovery Interface SPARQL Endpoint Triplestore Pull Schema OCLC ILS BIBFLOW:SCRIBE LD Cataloging DB Authorities MARC URI Step 2: Pull Schema Data from OCLC

Step 3: Pull Data (both literal and URI) from Authorities Discovery Interface SPARQL Endpoint Triplestore OCLC ILS BIBFLOW:SCRIBE LD Cataloging DB Pull URI and literal Authorities MARC URI Step 3: Pull Data (both literal and URI) from Authorities

Step 4: Push New Graph to Triplestore Discovery Interface SPARQL Endpoint Triplestore Push Triples OCLC ILS BIBFLOW:SCRIBE LD Cataloging DB Authorities MARC URI Step 4: Push New Graph to Triplestore

Step 5: Push Thin MARC to ILS Discovery Interface SPARQL Endpoint Triplestore Push Thin MARC Pull Schema OCLC ILS BIBFLOW:SCRIBE LD Cataloging DB Authorities MARC URI Step 5: Push Thin MARC to ILS

3. Cataloging Data Flow if No BIB or Authority Found

Current Original Cataloging Workflow for Print Monographs Melvyl (UC OPAC) Aleph Auth Bib Harvest (UCD OPAC) Search Aleph and OCLC BIB not exists OCLC WorldCat Vendor Service Auth Automated process (weekly done by Systems Dept) New name auth needed? YES Auth Control Create Bib Export Create holdings record Add/complete item record Physical Processing Unit Current Original Cataloging Workflow for Print Monographs

Step 1: Check to See If Record Exists. New if No Discovery Interface SPARQL Endpoint Triplestore OCLC ILS BIBFLOW:SCRIBE LD Cataloging DB Authorities MARC URI Step 1: Check to See If Record Exists. New if No

X Step 2: Attempt to Pull BIB Data, but no Match Found Discovery Interface SPARQL Endpoint Triplestore OCLC X ILS BIBFLOW:SCRIBE LD Cataloging DB Authorities Marc URI Step 2: Attempt to Pull BIB Data, but no Match Found

Step 3: Cataloger Enters new Descriptive Data Discovery Interface SPARQL Endpoint Triplestore OCLC ILS BIBFLOW:SCRIBE LD Cataloging DB When a bib cannot be found, the cataloger enters descriptive data data (title, publisher, etc.) directly into the cataloging interface. Authorities MARC URI Step 3: Cataloger Enters new Descriptive Data

X Step 4: Attempt to Pull Authority, but no Match Found Discovery Interface SPARQL Endpoint Triplestore OCLC ILS BIBFLOW:SCRIBE LD Cataloging DB X Authorities MARC URI Step 4: Attempt to Pull Authority, but no Match Found

Step 5: Cataloger Enters new Entity Data Discovery Interface SPARQL Endpoint Triplestore OCLC ILS BIBFLOW:SCRIBE LD Cataloging DB When a bib and authority cannot be found, the cataloger enters data (name, dates, etc.) directly into the cataloging interface. Authorities MARC URI Step 5: Cataloger Enters new Entity Data

Step 6: Push New Descriptive and Entity Data to Triplestore Discovery Interface SPARQL Endpoint Triplestore On submission, a new item graph and a new Entity graph are created in Triplestore with a unique URI for each. Cataloger submits the graph, including information about the item and the new Entity OCLC ILS BIBFLOW:SCRIBE LD Cataloging Authorities MARC URI Step 6: Push New Descriptive and Entity Data to Triplestore

Step 7: Item Graph and Entity Graph Pushed to OCLC for Reconciliation Discovery Interface SPARQL Endpoint New entity graph and item graph are pushed to OCLC for reconciliation. OCLC service either connects to existing, overlooked bib and authority or creates a new bib and Authority and links entities to the new Authority and work. The reconciliation service provides a publicly accessible (LOD) and machine actionable map of “Same As” relationships between entities and Authorities. Triplestore OCLC ILS BIBFLOW:SCRIBE LD Cataloguing DB Authorities MARC URI Step 7: Item Graph and Entity Graph Pushed to OCLC for Reconciliation

Step 8: Push Thin MARC to ILS Discovery Interface SPARQL Endpoint Triplestore Push Thin MARC OCLC ILS BIBFLOW:SCRIBE LD Cataloguing DB Authorities MARCURI Step 8: Push Thin MARC to ILS

Staff Inventory Control Interface Future Ecosystem (BIBFRAME/Linked Data Only) Pull Authority Data User Discovery Interface Push User Contributions SPARQL Endpoint Pull Graphs User Discovery Interface Push User Contributions ??? Respond to Machine Queries Pull Graphs Push Triples Push staff contributions Pull Schema OCLC Pull Triples Staff Inventory Control Interface Pull Graphs BIBFLOW:SCRIBE LD Cataloging Pull Authorities Authorities Pull Push Native Operation or Ingest Script / Tool (All other transactions are via API / Endpoint) Pull MARC Records

Acknowledgements Carl Stahmer, BIBFLOW Project Manager, who has provided many of the PowerPoint slides used in this presentation.

Thank You! For more information about BIBFLOW: Email: Bibflow@lib.ucdavis.edu OR Xiaoli Li (xlli@ucdavis.edu)

Reconciliation Service Other Data Sources owl:sameAs Mint Local URI