Brooking Ingesting Metadata - FINAL
2 likes621 views
The document discusses the challenges of ingesting metadata for library resources, particularly comparing ebooks and e-serials. It highlights issues such as the stability of content, catalog accessibility expectations, and the reliance on varying sources and standards. Additionally, it addresses the importance of high-quality metadata for discovery and matching, while considering the future implications of evolving library data practices.
Brooking Ingesting Metadata - FINAL
- 2. Outline • Background • The effect of a variety of sources and standards on library staff • Sharing metadata • Matching records • Some thoughts on the future
- 3. Background From microform sets to ebook packages • differences in stability of content (imagine a publisher comes into your library with scissors to remove titles from microfilm reels) • expectations for catalog access (public service librarians used to show tolerance for the years it took for catalog records to appear for individual titles in microform collections; now when a new ebook package is activated, catalog records are expected to be there the same day) • changes in library procedures • vendor services also went through same learning curve (it took years for the removal of titles or changes in urls to be reflected in metadata—still not always handled well)
- 4. Background Ebooks vs. e-serials metadata • linking vs. discovery • Full cataloging not as important for ejournals; the unit of discovery is mostly the article; metadata used for link resolution (ISSN, title, coverage data) • Full cataloging very important for ebooks; metadata used for discovery (but some KBs still only supply ISBN, title) • distribution of records • eJournals: the CONSER file serves as a primary full-level MARC record source for most commercially published journals and is widely distributed • no equivalent of CONSER for ebooks: more reliance on publisher metadata and local cataloging workflows
- 5. Variety of sources and standards • I don’t want to ingest your metadata (at least, not directly) • I want an integrated experience (librarians are no different than their patrons) • Too much DIY in-house: data wrangling, manual cleanup • different procedures for each package/file (not efficient, cognitive load) • size of file is large for UW (many titles, many packages) • lack of tools ? too many tools? (I use 7 now); or not the right tools? • escalating level of skills necessary to do the job
- 6. Sharing • Distribution issues (any barrier to free flow of full metadata is a problem) • ExLibris Community Zone (KB): commercial entity, share with other ExL customers only • OCLC WorldCat and WorldShare: cooperative, share with the world • But do publishers share everything with ExL? With OCLC? What may get lost in the process? • OCLC records vs. vendor records • As OCLC member, we prefer OCLC records; vendor record are more difficult for us to share and less likely to conform to library standards
- 7. Matching • High quality metadata for ebooks often exists • Is it shared? Can it be found? • Lessons from OCLC KB (aka WorldShare Collection Manager) • Inadequate matching algorithms (streaming audio matched to wax cylinders) • Early on libraries told OCLC “any record better than nothing” • Algorithm now much improved • Identifying collections/packages (extremely difficult, extremely important, not addressed) • Standard identifiers possible? • Consortial catalog (and beyond) • control numbers of various kinds, another problem, ISBNs not reliable match
- 8. Future • Adherence to a “traditional” library standards has not been easy in the past (e.g., MARC, Provider-Neutral, AACR2) • Libraries ingesting metadata from ever more sources with proliferating number of standards—will this help or hurt? • Linked data, a world with no records? • data with no context? • reconcilation, URIs from different sources • more matching algorithms than ever?