[Archivesspace_Users_Group] Digital Objects – examples of ASpace records + repository objects?

Tue Dec 5 15:05:28 EST 2023

It's my favorite ArchivesSpace question!

I've managed this a couple of different ways. When I worked at the University of Denver we had (well, presumably they still have) a home-grown digital asset management system that relied entirely on ArchivesSpace as the metadata system of record. For that we assumed that every object in the digital repository also had a metadata record in ArchivesSpace, and we used the update-feed API endpoint to communicate when new digital objects had been created or existing ones had been updated so that the repository would reindex. I'm sure somewhere I still have my documentation of how that data model worked and can forward it to you if it's of interest, with the caveat that I left DU four years ago and I don't know how much it has changed in the meantime.

Here at Penn State, our digital object guidelines state that every item in our repository should have a unique identifier, which for bibliographic materials is its OCLC number and for archival materials is based on the ArchivesSpace reference ID. We've found, as Sarah described, that this poses problems with older digital collections that were published before we had guidelines, where their source collections are processed at the folder level but the digital collection is published at the item level. How we handle that varies based on the collection. Sometimes we will re-publish the digital collection to reflect the arrangement of the archival collection; more rarely, we will maintain the digital collection as it is and re-arrange the archival collection to reflect CONTENTdm. Ideally we want to be at the point where we have a strict 1:1 relationship between archival object and DAO records in ArchivesSpace. (Technically 1:2, because we manage access and preservation digital objects separately.)

We considered and abandoned the One Item, Many DAOs approach Sarah describes, although if you know where to look in our PUI you can definitely still find some examples of it...

Happy to talk more about this and/or share documentation, although I feel like our documentation is probably inadequate (isn't it always, though?).

cheers
-kevin

________________________________
From: archivesspace_users_group-bounces at lyralists.lyrasis.org <archivesspace_users_group-bounces at lyralists.lyrasis.org> on behalf of Newhouse, Sarah <snewhouse at sciencehistory.org>
Sent: Tuesday, December 5, 2023 1:09 PM
To: Archivesspace Users Group <archivesspace_users_group at lyralists.lyrasis.org>
Subject: Re: [Archivesspace_Users_Group] Digital Objects – examples of ASpace records + repository objects?

Morgan,

We've been putting off a similar discussion for a while now, because of all the reasons you've already outlined.

We initially tried using the DAO component for access, to link our ASpace PUI with our digital collections, but ran into issues with differing levels of description between the PUI and our DC. Specifically, a few large collections that had grant-funded digitization to scan and create item-level records in the DC for everything. So a folder or volume/item-level component in ASpace would have 100+ DAOs linked to it, which made for a really unhelpful PUI experience for users looking at that component with just an endless vertical scroll of DAOs.

Our longer-term solution is to look not at ASpace, but at the data in our DC and see if there's a way we can create or simulate folder or volume-level works for relevant items in the DC and link to those using the ASpace DAOs, so there's one DAO for one ASpace component (folder, volume, box, what have you). Luckily, when the DC was built (before ASpace adoption) the developers built in a field for physical location that records box, folder, and volume info (scroll down, under "Institutional location"): https://digital.sciencehistory.org/works/k0lxixk

Our interim solution is to use the Existence and Location of Copies note and link to collection-level works in our DC. Example here: https://archives.sciencehistory.org/repositories/3/resources/635

Happy to talk more over Zoom or email if more details would be helpful! We're a small library just trying to make access as easy as possible with the tools we've got.

Science History Institute
Chemistry • Engineering • Life Sciences
315 Chestnut Street • Philadelphia, PA 19106 • U.S.A.
We tell the stories behind the science.
sciencehistory.org<https://www.sciencehistory.org/>

__________________________________

Sarah Newhouse   (she, her, hers)
Digital Preservation Archivist
Othmer Library of Chemical History
t. +1.215.873.8249

________________________________
From: archivesspace_users_group-bounces at lyralists.lyrasis.org <archivesspace_users_group-bounces at lyralists.lyrasis.org> on behalf of McKeehan, Morgan <morganem at live.unc.edu>
Sent: Tuesday, December 5, 2023 10:42 AM
To: archivesspace_users_group at lyralists.lyrasis.org <archivesspace_users_group at lyralists.lyrasis.org>
Subject: [Archivesspace_Users_Group] Digital Objects – examples of ASpace records + repository objects?

Hi everyone,

I’m looking for examples of how other institutions have handled relationships between ASpace records and digital repository content, in cases when the physical arrangement vs. intellectual arrangement poses object-modeling challenges for digitized representations.

For example, in your systems do you have scenarios when an ASpace archival object record may have many ASpace digital object (DO) records linked to it, with each DO record corresponding to an individual object in your digital repository system? Conversely, perhaps there are also  situations where a single digital repository object contains files/digitized pages that are described by separate archival object records – ie, a folder of stuff that was digitized as one grouping, but actually corresponds to more than one individual descriptive component?

At my institution, we’re working through these issues as we’re migrating our digitized materials to a new digital repository system at the same time as we’re also migrating our finding aids/archival data to ASpace. We would welcome any examples that others can share, in case you may have approaches from what has worked well, or lessons learned about what to avoid.

Please also feel free to message me off-list if more context or clarification about what I’m asking would be helpful. I’m happy to discuss more about our workflows and object modeling plans so far.

I’d also be happy to set up a zoom call if anyone has examples to share that it would be easier to walk through via zoom. I totally understand there may be examples that just seem too complicated to try to explain in an email!

Here's a little more explanation about the kinds of linking situations I’m asking about:

In our new digital repository, we plan to model digitized materials as repository objects based on the physical containers. For example, all digitized pages from a folder of correspondence would equal 1 repository object (a “Work”). The URL of the repository object will provide the File Version value for an ASpace Digital Object record, and that DO record will be linked to the relevant archival object record that contains the description for the original materials.

In cases such as a grouping of documents that is housed in a range of physical folders, this will mean many DO records linked to an archival object record. A common scenario is something like: a descriptive component with Title: “Correspondence, 1800”, housed in “Folders 65-75”. This arrangement will give us 11 DO records linked to the “Correspondence, 1800” archival object record. In many cases, folder ranges will be even larger, so there could easily be 20-30 or more DO records linked to an AO record. However, if we instead modeled all scans from Folders 65-75 as 1 object in our digital repository, since each folder contains 100+ pages, that arrangement would be a large and cumbersome repository object for users to navigate. Neither option seems great. We’re interested in learning about how others handle these types of situations!

Thanks for any examples or insights you can share,

Morgan

------------------------
Morgan McKeehan (she/her/hers)
Digital Collections Specialist
University Libraries
The University of North Carolina at Chapel Hill
morganem at email.unc.edu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lyralists.lyrasis.org/pipermail/archivesspace_users_group/attachments/20231205/a9987333/attachment.html>