Why RDA?

On 18/11/2016 19:08, John Gordon Marr wrote:

Let me comment on the “Fors” you cite.

I was asked to summarize the responses I’d received regarding why some libraries have not implemented RDA, and why some have. I happily received numerous responses to my question posted on Autocat (thank you!), so I have compiled the responses here. The three most commonly cited reasons to NOT adopt RDA were cost, limited staff resources, and the loss of the GMD subfield h. Here is my summary: …

I agree with all of this but would like to make a couple of points:

1) there are some figures now, available at http://www.primaryresearch.com/AddCart.aspx?ReportID=349 (the report costs too much for me), however they do summarize some of their findings. I found those results highly revealing, e.g. “According to the survey participants 111.72 minutes is the mean extra time needed for every 10 library items cataloged using RDA vs. prior procedures. The median time was 50 minutes, and the range was from 0 to 600 minutes.”

This means that it takes almost *2 extra hours* to catalog every 10 items, median time was almost 1 hour. This represents a significant loss of productivity that must have serious consequences, even though I haven’t noticed anyone talking about these results. In some businesses, such a loss of productivity would be considered a disaster. It seems to me that this is the kind of finding that screams to be addressed, especially in an atmosphere of declining resources.

2) hybrid catalogs. Library catalogs have always been hybrid, containing early rules (i.e. pre-AACR rules), AACR1, AACR2 and now RDA. Once again, quoting from that same report, “A plurality of survey participants were not in favor of retro-conversion services for RDA cataloging as they do not think that it will result in saving of time and money, and high quality records. Out of all 56 responses received 26 were against retro-conversion, 12 favored it, and 18 responses contained mixed opinions.”

I shall add to this the popularity of library “discovery services” which search dozens, or hundreds, of sites to return a single result for the searcher. The overwhelming majority of those sites are not, and will never be, RDA. Apparently, the public prefers these kinds of results. From these considerations, it looks as if hybrid catalogs will be a fact of life for the foreseeable future, if not forever. Therefore, the FRBR/RDA dream to:

find/identify/select/obtain works/expressions/manifestation/items by their authors/titles/subjects

can never be realized and will remain only a dream, at least not until these fundamental issues are addressed and resolved.

I take a backseat to no one about the importance of consistency: consistency in description, and consistency in access, both name and subject. But the question we should be asking today is: consistency *with what?* Library catalogs are no longer the solitary tools they have always been. Now they are leaving their local libraries and becoming just one of many tools that searchers use. Our individual records are being jumbled-up with all kinds of other records, where they lose any trace of our vaunted “consistency.”

Nevertheless, I think consistency is important, but we must reconsider completely what “consistency” means today, and how it can be made useful for the different kinds of searchers.



One Comment

  1. Ron Murray said:

    Where are the Library Linked Data Metrics – Or: If You Can’t Measure It, You Can’t Manage It

    In our highly industrialized library environments – where our affection for mass production techniques goes at least as far back as Dewey – metrics have been developed and employed managerially to various ends. Among the most venerable are “widget counts” – the number of books, newspapers, catalog cards/records ingested, processed, circulated – along with measures of widget count change that can creep into Calculus if we let them.

    But, as so many of the benefits of LLD are claimed to follow from the linkages/relations among descriptive widgets of various sizes, one expects to see metrics that monitor the degrees and types of linkages created. A little time with the graph theory literature will surface quite a number of connectivity measures that make lots of managerial sense. (If you need a nudge, see: http://reference.wolfram.com/language/guide/GraphPropertiesAndMeasurements.html)

    Attention to link structures has been institutionalized at least as far back as Panizzi & Cutter – but LLD makes link-making and following central to the CH resource description process. So how do you measure (and manage) this linkiness in a managerially relevant way? I get the impression that the reports cited don’t pay too much attention to how authority and other types of linkages can shorten search paths (inside/outside of their institutions), or e.,g., how efficiently Expression-level links draw together multiple language versions of a Work (inside/outsise, etc.).

    Do you know of any LLD metrics that can tell one how a mandated or optional link will improve connectivity among linked/cross-referenced bibrecs?

    November 29, 2016

Comments are closed.