Shlomo Sanders, CTO, Ex Libris
For years the library community has understood the value of interoperability and the sharing of bibliographic data. Sharing data saves money and can increase the quality of the data itself. Understanding this underlying value is inherent in the definition of a standard for describing bibliographic data, such as MARC 21 and associated encoding rules, and can be seen in other standards including z39.50 and ILL. MARC even goes so far as to describe what is core and what is “local” (9xx, x9x and xx9) in its attempt to make clear what can be shared and what is not shareable.
How nice it would be if there was one universal standard and if all institutions could easily share their bibliographic data. Zooming out from the institution-specific to the country level there are many variations of MARC. Just to name a few, these include UNIMARC, DanMARC, CNMARC, KORMARC and others further removed like MAB. We do see convergence in certain countries and regions, such as the replacement of MAB and of the Denmark-specific DanMARC by MARC21.
Another issue that complicates the metadata universe is the proliferation of authority files, one per country. These authorities do not just have differing data. MARC21 authorities and UNIMARC authorities have a different structure and prevent easy cross-format use.
At the institutional level, even if one institution uses the same “version” of MARC as another, the interpretation of how to implement the coding rules can be very different. As an example, identifiers, of crucial importance in the use of the metadata, can be managed in different fields in the MARC standard. 035 is the primary field for such identifiers, but some systems use 019 instead. Institutions that try to reuse metadata then have to make hard choices. They can have difficulty using the metadata as intended, may need to process every metadata record loaded, or in some cases may even need to fix them manually. In many cases crosswalks are created between metadata standards and institutions often implement complicated merge routines and match methods.
As more and more libraries need to manage digital collections, they have begun to use non-MARC standards such as Dublin Core, MODS and VRA. Since many digitized items describe a physical item, we can then end up with two separate descriptions: one metadata description in MARC and the other in a metadata format that supports digital items. The variations in the format and data make deduplication and grouping of such records a challenge.
Ex Libris has developed many workarounds to support the loading and sharing of differing metadata in its mature library solutions. This has eased some of the problem inherent in differing metadata standards but introduces complexity. The development time involved in creating these workarounds could have been used to develop other needed features and capabilities for the community.
The Alma development team has needed to take an even broader view of the multiplicity of metadata formats and has extended Alma’s reach to support a unified metadata management approach. This allows an institution to use the same catalog to manage MARC variants and Dublin Core metadata, with future plans to natively support emerging link data formats such as BIBFRAME.
And back to the dream. By moving forward with BIBFRAME, the community has the unique opportunity to reach consensus and create a universal standard. This is hard work. It will require collaboration, compromise and active participation and support by vendors. As in any compromise, the result may not be ideal in all cases, but it will have tremendous implications for the ability of libraries to share information. This is especially important since one of the stated objectives of BIBFRAME is sharing with non-library environments and systems. How can we expect such external systems to use a new standard if it turns out not to be truly standard at all, even within our own institutions? This is the promise of BIBFRAME.
To learn more about BIBFRAME in the Ex Libris development roadmap, read our recent press release.
May 18, 2017
