Hi Jeff,

I share your enthusiasm about bibliographic metadata.

Let me respond in brief (I could write books about the subject)

1) Librarians world-wide have organizations like IFLA. We know UNIMARC, which was a followup to MARC. I appreciate the energy and dedication of the Library of Congress to create a better solution than MARC after 50 years of stagnation on behalf of the community. There is no UNIMARC followup to my knowledge. Whether LoC is still a bibliographic format authority to librarians, the librarians themselves are up to decide. As a side note, librarians should remember Z39.50 or SRU for information retrieval, and why it failed since the age of the World Wide Web. It's not just MARC. You are right, cataloging as a job description is dead. In Germany, libraries have renamed the cataloging departments to metadata processing departments, while extending the job descriptions to research data archiving and repository maintenance, for a heavily reduced staff.

2) The economic argument is a wise argument. I agree it must be cheap to transform MARC to Bibframe data. And it must be cheaper to run Bibframe systems than MARC systems. There is also a "sweet spot" where Bibframe is becoming inevitable but we are far from it. Commercial vendors want to be paid and they wait for customer demands. There is not much demand for Bibframe (or other semantic web technologies) as a replacement for MARC for several reasons: unchartered territory, no first movers, no "killer applications", no best practices. So the price is expensive from a vendor perspective. I think it's good that initiatives like FOLIO have to start with MARC. They can keep costs low and could add features to transform MARC to Bibframe later, to build the right solution in the right situation.

3) Bibframe alone is not enough, it must be surrounded by vast efforts to get things going at the software level. I know a lot of effort is going into "Linked Data". Bibframe is just one part of "Linked Data" (which is a simplification of the Semantic Web). National libraries in US and Europe have started to express bibliographic authority data on the Web since many years. For example, https://id.kb.se/ There is a top-down strategy behind it. But you are right, the projects are rare, for many reasons. They seem to stall.

4) I disagree. Bibframe is simple, it simplifies many rough edges. It is so simple that it even does not follow FRBR and does not cover all the specifications of MARC at https://www.loc.gov/marc/bibliographic/ So, as a consequence, the MARC train is still on the road. Maybe I have a different perspective because I am familiar with the computer science behind it. I also like talking about Bibframe, I like to preach "Linked Data", but I also like creating catalog search solutions in a given timeframe which are imperfect but just do the job. Where I agree is probably the perspective from cataloging. For example, I do not understand why RDA cataloging examples and implementations have not picked up Bibframe as a prerequisite. They seem like not being made for each other, which is confusing and kind of bizarre. On the other hand, MARC was historically able to catch up with the ongoing requirements of AACR.

5) Less need for quality metadata is maybe correct for those who want it cheap and dirty. And Big Data sounds cheap and dirty in my ears, while the reality of Big Data is not. Big Data is a slogan made by large computer companies who want to bind customers to their data centers. But as always, there are interests in high quality data, it just depends on the long term goal. Imagine algorithms that combine many hundreds and thousands of library catalogs into a huge data pool, extracting bibliographic entities, removing duplicate entities, building new links to external open data knowledge base entities, enriching catalogs, and finally surprise users with a new experience. That is where quality metadata is required, at the engine level, for creating better search engine infrastructure and services. I have learned librarians do not focus on such search technology or "discovery" - they buy it or ask humbly for licensing fees. Maybe librarians just sit there and wait for buying Bibframe systems, too. If so, they are about to lose their traditional role of being highly skilled experts for defining the future of modern, global information retrieval systems, as a key for accessing knowledge.

Best regards,


On Wed, Feb 1, 2017 at 4:45 PM, Jeff Edmunds <[log in to unmask]> wrote:

Next month I'll be giving a presentation entitled "Life after MARC?: The Future of Discovery." One of the central theses of the talk is that, unlike MARC, BIBFRAME will never be widely adopted. I would be very curious to hear arguments countering my key points:

1) Unlike its stature in the Age of MARC, the Library of Congress lacks the authority to impose its standards or practices on the bibliographic metadata ecosystem. LC is no longer the chief source of bibliographic metadata, and the influence on the ecosystem of the bibliographic metadata it generates is minimal. The most likely allies of LC in moving BIBFRAME forward--large academic or public libraries with experienced cataloging staff--are shifting resources AWAY from cataloging and into other areas (digital humanities, assessment, student engagement, open access, etc.).

2) The technical infrastructure of current ILS, discovery, next-generation, etc. platforms does not support BIBFRAME and there is no market incentive to change. Open-source, collaborative ventures like FOILO must necessarily base much of their development around current and legacy MARC data, not largely hypothetical data models like BIBFRAME.

3) BIBFRAME in a production environment is wildly impractical: the BIBFLOW Project, although it officially ended in 2016, has issued no substantive final report of which I am aware. Does one exist?

4) BIBFRAME is highly conceptual, top-heavy, and too complicated to be understood and effectively implemented by most libraries (cf. the failure of many librarties to adopt RDA over AACR2). The people talking about BIBFRAME (various PCC committees, etc.) are not creating BIBFRAME-compliant metadata; they are merely talking about it.

5) Discovery of library (and archival, etc.) materials no longer runs primarily on bibliographic metadata; it runs on megadata, which I define as a complicated mess of metadata, full text, and big data (including personalization data). As a result, there is less need for, and appreciation of, quality metadata.

Counterarguments are welcome, either on- or off-list.

Thank you,