Maureen Callahan & Don Thornbury Rare Books and Special Collections Princeton University Library ENUG 2011
1. We believe in the possibilities of single search.2. But how do we represent special collections materials?3. And how do we make this play nicely with Primo? (It won’t work out of the box).4. What are the implications of this change and what are our next steps?
“Users can search Internet resources through a single search engine query, yet often the resources of a single cultural institution or university campus are segregated into silos, each with its own dedicated search system.”Single Search: The Quest for the Holy Grail – OCLC Research Report by Leah Prescott and Ricky Erway, June 2011
Let’s say that you’re studying Orthodox Christians in Ethiopia…
So you’ll start with sources in the library collection.
Books, articles, art, archives, sound recordings, objects – the way we describe these are all records with many similarities. They all have some or all of the Dublin Core fields. So let’s put what we can into Primo, so that our researchers only have to learn one opaque research system, instead of many. We decided to start with special collections materials currently described in EAD.
MARC – Henriette Avram, late 1960s EAD (Encoded Archival Description) – UC Berkeley, 1993 Before this (and currently, too, depending on where you are): big, long essays explaining the collection and big, long paper lists of the contents of the collections. Archival description in five minutes…
We want each piece of valuable data in the contents list to go into Primo, so that users can find it in a institution-wide search. In fact, we’re not super nuts about the current system – users having to scroll through big-crazy-long finding aids (with the help of Ctrl-F). We want users to be able to go straight to the part of the collection that has the information they seek.
Our content standard (DACS) tells us the requirement for single-level records. DACS doesn’t care about the encoding standard – MARC, EAD, PNX, whatever. And, as we look at each component (that is, unit of description or line in the contents list) we realize that each one has the required parts to be its own record.
We’re going from one line in a finding aid to a big PNX record.
EAD Generic.dtd PNX You should see the massive stylesheet. Iterative process. It’s good to talk about theory. A lot. It’s good to push back against Primo.
1. And PNX is how we shall come to know our data…2. Re-thinking the folder list.3. Visibility for special collections.4. De-centering both organizational boundaries, boundaries of custody, and boundaries of creator/collector.5. Fidelity to FRBR model.
Getting those 650K archival records ingested… Working with other data sources, converting to generic then PNX (should be easier, since we don’t have to factor in hierarchical relationships, usually!) EAD site redesign Description revolution.