Implementing ArchivesSpace

The man who moves a mountain begins by carrying away small stones.

It takes time to do great things. Mike has already outlined our lofty goal to employ ArchivesSpace, Archivematica and DSpace in a single, end-to-end digital archiving workflow. That won't happen overnight. Before we do anything else, we have to get each of these systems up and running individually.

DSpace, the repository software package underlying Deep Blue, the University of Michigan's institutional repository, has been in place for quite some time now, so I won't dwell on it here. An upcoming post will detail our work to implement Archivematica, to customize workflows and pipelines and to determine how it will replace and/or extend our existing systems and procedures.

Today's post is about ArchivesSpace:

Implementing ArchivesSpace

We've had an ArchivesSpace test environment running on a re-purposed Windows machine in the back of what would become my cubical for just about a year now. At that time, Bentley Historical Library staff and a number of graduate students from the University of Michigan and Wayne State University did some preliminary testing of the features and functionality of ArchivesSpace, including its import and export functionality.

In January of this year, a couple of folks from the ArchivesSpace team at Lyrasis came out to do a three-day workshop on ArchivesSapce. The first two days of the workshop covered the basics:

• creating Accession records;
• creating Resource records;
• creating and managing Agent and Subject records, and linking them to Accession and Resource records;
• recording and managing physical locations within a repository;
• producing description output files in standardized data structures such as EAD and MARCXML; and
• importing legacy data and performing data cleanup tasks.

The final day covered Digital Objects:

• the functional scope of the ArchivesSpace Digital Objects module;
• how ArchivesSpace might be used in tandem with external digital asset management systems;
• modeling, creating and updating simple and complex Digital Object records;
• linking and relating Digital Object records to Resource and Accession records; and
• generating Digital Object metadata exports in standardized data formats such as METS, MODS and Dublin Core. 

By February, therefore, we had implemented the software and reviewed its functional and technical requirements for use and development. That is, we had our very own ArchivesSpace instance and we knew [basically] how to use it. We could check that preliminary item off the list. 

In order to fulfill the requirements of the grant, however, our test environment of ArchivesSpace would need to become a production environment of ArchivesSpace.

Our Strategy: Call the A-Team

If you have a problem, if no one else can help, and if you can find them, maybe you can hire the A-Team...

Not this one [1].

...that is, the ArchiveSpace Implementation Team at the Bentley Historical Library, or the "ASpace" Implementation Team (for short), or, simply, the A-Team (if you're trying to be pop culture savvy).

The four of us (Mike, Dallas and I, and our colleague, Lead Archivist for Description and Workflow Management Olga Virakhovskaya) sat down for our first meeting. After deciding which A-Team member we wanted to be (first things first, of course), we came up with a charter of sorts, which begins like this:

The Bentley Historical Library (BHL) will implement the ArchivesSpace (AS) archival management system to replace current resources and procedures, increase efficiency, and join peer institutions in establishing archival best practices for the 21st century. BHL seeks to have a full implementation of AS by March 2016, with the capacity to create new accession and resource records and manage legacy records (primarily accessions and EAD) imported from current resources. 

We also broke our tasks down into three main areas, and divided up the work.

     1.  Leading the Way (and Liaising with MLibrary)

Mike is the project leader. Using lessons learned from a recent Society of American Archivists offering here on "Project Management for Archivists," he has laid out project outcomes (see above), goals and objectives, roles, constraints (e.g., the [im]maturity of some features of ArchivesSpace, the timeline imposed by the Mellon grant and the current [in]compatibility of ArchivesSpace with existing software platforms in use here, such as Aeon and Digital Library eXtension Service, or DLXS), costs and timeline.

We're experimenting with Teamwork, a web-based project-management tool, to track our progress.

Mike is also responsible for liaising with the University of Michigan Library. A production environment of ArchivesSpace will not be able to run off of the computer we have sitting in the back of my cubical. Instead, the current plan is that once we are "live," ArchivesSpace will be hosted, maintained and supported by the Library Information Technology division. Charged with "the design, development, management, and maintenance of a flexible and reliable technology environment," they have human and technological infrastructure in place to support library management systems like ArchivesSpace in a way that we simply cannot.

     2.  Developing New Accessioning Conventions and Descriptive Practice

At some point, we will have to begin creating accession records for new donations and transfers, as well a generating resource records for archival collections. Since both of these are likely to change after implementing ArchivesSpace, and since finer granularity has overhead for data input, Olga will take the lead for developing new accessioning conventions and descriptive practices.

     3.  Import of Legacy Accession Data and EAD Finding Aids

Finally, Dallas and I are working on importing legacy accession data (all 19,000+ records) and Encoded Archival Description (EAD) finding aids (all 2,800+ of them). We have decided to start with the latter, and hope to begin work on legacy accession data by the fall of this year.

Accession Data

The accession data will be coming from a CSV export of a homegrown FileMaker database, the Bentley Electronic Accessioning and Locating System, or BEAL, which also happens to be the name of the street on which we are located:

This has been described as the "lifeblood" of the Bentley Historical Library, and one of the many challenges we foresee with migrating legacy accession data is figuring out a way to ensure that any tasks and processes (both internal like mailing lists, reports, &c., as well as external) reliant on accession data in BEAL will be able to be done once we make the move to ArchivesSpace.

EAD Finding Aids

Finding aids start life here as Microsoft Word documents and get converted to valid EAD XML through a series of Word macros. Next, they are made searchable and displayed using DLXS. You can read more about that process here.

Our work to import legacy finding aids created like this into ArchivesSpace builds off of work Dallas did during a practicum he completed here before becoming a Project Archivist.

It hasn't exactly gone smoothly.

We have an error rate of about 30%, and even when we don't get an error, finding aids don't always import the way we expect, or in a way that takes the advantage of ArchiveSpace's functionality (or potential functionality, once we have a Hydra-based implementation of Deep Blue and, possibly, DLXS). Part of this is because the Microsoft Word method lends itself well to human-understandable finding aids, while data in ArchivesSpace needs to be both human- and computer-understandable.

Even if it doesn't affect import, we have discovered through this process that some of our data is fairly messy. We feel that we might as well clean it since we are spending so much time with it now and because it may be harder to manipulate in batch once it's in ArchivesSpace.

Next Steps

Longer term goals and objectives that the A-Team will have to address include:

• importing legacy MARC records;
• exporting EAD finding aids;
• automating the transfer and upload of EADs to the DLXS platform;
• exporting MARCXML records for import into the catalog; 
• developing training materials and documentation; and
• determining how ArchivesSpace will communicate with other Bentley Historical Library and university systems.

Be on the lookout for the aforementioned post on Archivematica and a post on some of the errors Dallas found during his practicum while importing a sample of our finding aids into ArchivesSpace. In the near future we'll also launch an ongoing series highlighting an ArchivesSpace plug-in we created, as well as some basic programs we have developed and tools we have used to address those errors.

[1] "Ateam" by Source. Licensed under Fair use via Wikipedia - http://en.wikipedia.org/wiki/File:Ateam.jpg#/media/File:Ateam.jpg