RDA-L

 View Only
last person joined: 4 days ago 

Open discussion of RDA, RDA Toolkit, and related topics
  • 1.  describing hybrid data

    Posted Aug 29, 2022 03:08 AM

    Dear RDA experts

     

    We are currently thinking about metadata which is termed «hybrid». By this we mean that one and the same dataset (in our case in MARC21) describes a resource which is available both in an analogue form and as a digitised version. An example would be a pencil drawing on paper which has been digitised and exists also as an image file in Jpg.

     

    The nuts and bolts of the hybrid character come about in MARC 007 (two of these: one for the paper object, one for the online resource), 024 (identifier for the Jpg), 337 (occurs twice for 2 different resource types), 338 (occurs twice for 2 different resource types), and to a lesser extent 856 (links to an external platform where the Jpg can be viewed in high resolution; we even have a regional application rule for RDA 2.1 explicitely allowing this hybrid attribute in context of mass digitisation).

     

    There are advantages of having all this information in one place. But we are concerned that this kind of hybrid metadata amounts to rule-breaking in terms of RDA: the integrity of the description is compromised, messed up by mixing attributes of different resource types.

     

    -> Has anyone taken the plunge and tried this out? How do discovery systems behave? Are there huge problems downstream when you dismantle the coherence of metadata?

     

    Thanks for any advice out there.

     

    Marcus

     

     

     

    --

    Marcus Zerbst

    head metadata

     

    Zentralbibliothek Zürich

    Zähringerplatz 6

    CH-8001 Zürich

    Tel +41 44 268 3221

    Fax +41 44 268 32 92

    marcus.zerbst@zb.uzh.ch

    www.zb.uzh.ch

    marcus.zerbst@zb.uzh.ch" target="_blank">Chat via Teams 

     

     



  • 2.  RE: describing hybrid data

    Posted Aug 29, 2022 07:30 AM

    For digitized holdings (that is, not for reprints issued electronically, but for reproductions we make of our holdings), we simply do not have the capacity to catalog everything again nor even to change the record that describes the original. As a result, we use holdings records in the local catalog (Alma) to describe the electronic reproduction, and, locally, only the holding record carries the information about the electronic resource. When reporting to OCLC, multiple bibliographic records are sent. The bibliographic record for the digital resource includes a subset of fields from the Alma holdings record that are converted to equivalent bibliographic fields, e.g., the 843 field is converted to the 533 and 539 fields. 

     

    Local catalog:

    https://id.lib.harvard.edu/alma/990013680150203941/catalog

     

    Worldcat record for digital: https://www.worldcat.org/title/609119762?oclcNum=609119762

    Worldcat record for original: https://www.worldcat.org/title/1172882?oclcNum=1172882

     

     






  • 3.  RE: describing hybrid data

    Posted Aug 30, 2022 07:49 AM
    There was a time when US libraries attempted such "multi-ver" (multiple version) records. The practice had its genesis in consolidating print and microform serial runs, particularly in a card catalog environment. The practice did not long survive in an online catalog environment when the modeling for treating different formats became more stringent and the labor overhead for managing multiple records became much lower than it had been for card catalogs. It could and did persist within some local catalogs since the consolidation of print and microform records was manageable and was often seen as a benefit to users who would find holdings consolidated on one record.  The death knell for my own institution came with the advent of packages of both e-serials and e-books. The quantity of records involved under these packages were such that it was an insurmountable task to manually reconcile them against the corresponding print records. There was also the issue of ongoing maintenance, since titles would be added to or dropped from such packages as contracts with the package aggregator shifted. With separate records, these could just be incorporated or dropped from the database where "multi-ver" records again would have required manual manipulation (and because neither set of records -- tangible and electronic -- would be a proper subset of the other, and entire update file would have to be managed manually). Meanwhile, in the current catalog modeling, and in RDA built on such modeling, different formats are treated as distinct manifestations warranting their own description.

    The above paragraph lays out the theoretical and practical considerations against "multi-ver" or "hybrid" cataloging practices. Having said that, the scope of local digitization efforts may resurface the older dynamics that drove the earlier embrace of such treatments -- the scale of the digitization work may be sufficiently limited to afford manual matching as a more labor efficient effort than crafting distinct records (although cloning the records of the original tangible resource could be fairly straightforward). Even with distinct records for electronic versions of resources, there are still many tangible version records with URLs appended that point to the electronic version, although typically with minimal further modification as was seen in "multi=ver" records (for example, no longer adding another 300 field,  006 field, and 007 field). If such "hybrid" treatment is confined to the local catalog, one is largely free to do as one feels is best for the institution; there won't be cataloging police knocking at one's door. My own institution has such hybrid records for our students' theses from the brief period when we transitioned from strictly tangible to strictly digital formatting -- there were several years when both formats were produced. But short- and long-term costs and benefits should be carefully weighed. In particular, as the scope and scale of digitization efforts expands (and harvesting of data from a digital repository becomes the reality), the "hybrid" treatment may become less tenable, however appealing it may be in the short-term.

    John Myers, Catalog & Metadata Librarian
    Schaffer Library
    Union College
    Schenectady, N.Y.

    ------------------------------
    John Myers
    Catalog Librarian
    Union College
    He/Him/His
    ------------------------------



  • 4.  RE: describing hybrid data

    Posted Aug 30, 2022 09:15 AM
    There is a theoretical aspect missing here: we are talking about copies of individual *items*. It may not matter for commercially published works in a circulating library, but for special collections, which might have artifactual characteristics or provenance, disassociating the copy from the original then somehow re-associating it is much more work than "cloning" a record implies.




    Kate Bowers

    Collections Services Archivist for Metadata, Systems, and Standards

    Harvard University Archives

    kate_bowers@harvard.edu

    voice: (617) 998-5238

    fax: (617) 495-8011

    web: http://nrs.harvard.edu/urn-3:hul.eresource:archives

    Twitter: @k8_bowers








  • 5.  RE: describing hybrid data

    Posted Aug 31, 2022 04:59 AM

    Dear Kate

     

    Thank you for that valuable contribution. In our case (Prints and Drawings), those item-specific attributes and descriptions are indeed frequent and critical. Mixing these goes against the grain of entity-based cataloguing by blending attributes of two different manifestations in one description. A lot can be achieved with 776 or 856 fields without mixing the intrinsic description.  

     

     

    Alice Robinson-Baker

    Stv. Leiterin Graphische Sammlung und Fotoarchiv

     

    Zentralbibliothek Zürich

    Zähringerplatz 6

    CH-Zürich 8001

     

    Tel.      +41 44 268 31 00

    Direkt: +41 44 268 31 69

     

    alice.robinson@zb.uzh.ch

    www.zb.uzh.ch

     

    Abwesend: Montag