Modify digital object search and retrieval functionality, to also search collection titles, and return the results of any digital objects linked to collections that include the search term


We have recently come to realize that the search and retrieval function for digital objects in ArchiveSpace only searches across digital object titles, and returns those results. We would like to have the search and retrieval function also search across collection titles, and return the results of any digital objects linked to collections that include the search term. This is a problem both in the staff and public interface. As we have allocated many resources to increase our capability to digitize and provide remote access to our collections over time, we believe this issue provides a fundamental barrier for our users to discover and access this material.

As an example, our repository (Harvard Law School Library) recently digitized a large amount of material for one of our collections (the Manley Ottmer Hudson papers), resulting in 673 digital objects being ingested into ASpace and linked to the collection's resource record. The collection was digitized at the file level, and the folder titles (the overwhelming majority of which do no include the term "Hudson") were used as both the file-level archival object and digital object titles in ASpace. Having the title of ASpace digital objects be identical to (or perhaps slightly different and more specific than) their corresponding archival objects has been a standard practice for most, but not all, of our collections (as well as a large amount of collections across all Harvard repositories).

Once the Hudson resource record was complete, we performed test searches for digital objects in both our staff and user interface using the term "Hudson." Only nine items were returned, and only one of which is actually in the Hudson papers. We assumed this is because the term “Hudson” happened to be in the title of that particular ASpace digital object. We confirmed our theory by changing the title of another digital object to include “Hudson,” after which, that digital object was also returned in our search results.

We recently brought the issue before Harvard's ArchiveSpace Working Group (ASWG), who did some testing in other institutions' PUI instances, and confirmed that they believe this is a general/core code issue within ArchiveSpace.


Maureen Cresci Callahan
September 23, 2020, 2:11 PM

I wanted to comment to say that this is an important user need that we’re also encountering. During quarantine, our staff are doing amazing, heroic things to get materials digitized and link them to archival description via digital objects. But if a user asks “what do you have digitized from x collection?” it’s not possible to return a list of results.

On the Princeton finding aids site, for instance, I can do a search for “Dulles” and click a box to see online content only and even then filter by which collection the content belongs to:

I’m not sure that I agree that I would like for only the collection title to be indexed – instead, I think I would prefer for the associated archival object record to be indexed with the digital object record so that I can find all related archival description that explains what my digital object is.




chris spraker