Speaking the Same Language Serials Standards and e-Resource Data Interactions Diane Hillmann Cornell University Standards Evolution • Normal development trajectory— where do we fit? • What current standards are relevant to our emerging needs? • What are the real problems we’re trying to solve? Development Learning Curve • • • • • Simple ---> Complex Monographs-----> Serials Human ----> Machine One-at-a-time ----> Batch Single use ----> Re-Use MARC 21 Holdings • Represents 20 years of experience with publications • Tested approach to encoding and exchanging data • Growing installed base of publication patterns • Solid infrastructure of sharing and collaboration So why hasn’t MARC Holdings been adopted outside libraries? • “It’s too complicated--we need something simpler” • Too specific to libraries to serve the needs of others Is it too complicated? • Everything in MFHD was developed in response to real publications (we couldn’t make all that up—but publishers did!) • MARC is a communications format—not a user interface • Too complicated for who? Not for machines! Why not start new with something simpler? • “A simpler solution should accommodate 80% of the titles easily” • The 80-20 rule (Pareto’s Principle): 20% of the titles will cause 80% of the problems • Which 80% will a simpler solution handle, and who will deal with the other 20%? • Likely answer: expensive humans (and who pays for them?) Shared Goals • Efficient and timely communication of transactions (updates, changes, new info) – Emphasis on machine rather than human communication • Unambiguous referencing to all levels of publication (titles, volumes, issues, articles, etc.) • Easy re-use of data from other sources Shared Information • CONSER records contain: – Bibliographic description: title, publisher, dates, related titles, etc. (MARC Bib) – Enumeration, chronology, captions, prediction patterns, etc. (MARC Holdings) • CONSER record is a collaboratively created and maintained “Publication History” for serials – Standardized for sharing, with an existing support infrastructure A Step up with “Super Records” • A possible solution to the FRBR “work” level for serials – Gathering: • Relationships (title changes, versions, etc.) • Publication History information (what was actually published) – Creating • A basis for better user display • A template for more efficient re-use of information on serials Down to cases • What library functions can shared “Publication History” support? – Interlibrary loan (better matching of user needs to holdings at the institution and partner institution level) – Remote storage (easier decision making and access to multiple physical and digital locations) – Reference linking (correlated citations from many sources) – E-Resources management (less ambiguous user displays) The Key is Information Flow • Determination of where Publication History content should be: – Created – Updated – Managed • Determine best methods of: – Distribution to interested parties – Notification of changes A Possible Shared InfoFlow? • Librarians continue to create and maintain Publication History – Using MARC Holdings and CONSER database as basis for sharing • Determine best distribution modes for publishers and vendors – Direct access via OCLC or other means? – OAI-PMH for XML transfer to internal databases? – Other? Maintaining the InfoFlow • Shared data but separate responsibilities – Agreements defining expectations – Investment in appropriate distribution pipelines – Development of a common infrastructure that supports efficient, machine-based interoperability New title publication (Publisher) Library acquires title & first issues Publication Pattern created & distributed Vendor distributes updates using pattern model Library uses updates to maintain subscription Library catalogs Publication history record captured by vendor Library posts pattern changes for redistribution Supporting Citation MARC 21 SICI 245 $a OpenURL jtitle 020 $a [ISSN] issn 853/863 $i-k (Chronology) date Enumeration volume/part/ issue 853/863 $a-c Conclusion • The world as we know it favors those operating furthest to the right on the Learning Curve, with solutions: – Complex enough to do the job – Emphasizing machine interactions as much as possible – Re-using information created and maintained by others when practicable