I’m really glad you resuscitated this thread, Jarrett, because we’re talking a lot about this at Yale.
A lot of work is happening right now on integration of archival description in our database of record (ArchivesSpace) with our digital preservation system (Preservica) and our access system for digitized objects (Blacklight).
Here are some common questions that have come up and my answers to them:
1. Are we “managing” digital objects in ArchivesSpace?
This depends on what you mean by managing.
I cannot think of a situation where ArchivesSpace would be the only layer between metadata and a file system (other than, possibly, very basic digitization activities), so no, we are not doing that kind of management in ArchivesSpace.
But I think ArchivesSpace digital objects WILL be the glue between two different management systems — ArchivesSpace (for the description of functions like accessioning and description) and Preservica/FindIt/Quicksearch/Kaltura/HathiTrust/what-friggin-ever
where more robust information about complex objects, preservation actions, technical facts about the object, etc. are stored.
As I see it, the best thing that AS digital objects could do would be to be a place to keep URIs so that we can sync the systems together. If we think about it this way, this whole project becomes a lot less complicated, I think.
2. What is the good of the digital object record?
The digital object record lets us keep structured metadata in ArchivesSpace about digital objects that can serialize as ead//dao or METS. It can also be accessed through the API as structured, JSON objects. We had discussed the idea of using “location
of copies” and “location of originals” notes as a possible alternative to DOs, but there is an advantage to storing information about digital objects in a DO record rather than having a URL as part of a string in a note. Notes are difficult to query and manage;
digital objects are a bit easier.
There’s also a bit of extra metadata that can be created/stored in the digital object record that can help our public interfaces know what to do with these links to other system, which is pretty useful.
3. Should digital surrogates and born-digital records be treated differently in ArchivesSpace?
If the DO is just the glue between the description of the object and the system that gives you the object, then no. I think that they need to be described differently, because there’s a different facticity to them as records, but
I don’t think that they need to be managed differently. And since there are really pretty good attributes and elements on the digital object record to help us determine what kind of a digital object we’re dealing with and how it should load/display, I don’t
think it’s a problem to have many digital objects on an archival object that point to different manifestations in different systems.
So here’s what our digital objects look like:
Title: display title from archival object (title and date) — N.B., this is only because it’s required. I’d prefer not to have the duplicate data.
Publish: publish status from archival object
Digital object identifier: handle to object in Blacklight/Preservica/Whatever
Most of the creation of digital objects will be done through scripting or automatic integration between systems.
Since we’re not pointing to actual files in actual systems, we won’t be using FIleURIs.
We may include more metadata to indicate whether this is a digital object that takes the user to an access system or whether it takes a staff member to the place where she can do preservation actions (this will also affect the publish element).
What about everyone else? How are you using digital objects? By the way, we’re still in the middle of figuring this out, so the above only represents my thinking and current understanding of the direction at Yale.