1 of 37

Welcome to the

2020 DLF Born-Digital Access Working Group Colloquium

August 20, 2020

#dlfbdawg

2 of 37

Get involved with BDAWG

The DLF Born-Digital Access Working Group, formed in 2017, produces research on and advances the practice of providing access to born-digital collections.

All are welcome to join!

Google Group:

https://www.diglib.org/groups/born-digital-access-group/

3 of 37

A few things to keep in mind...

Keep yourself muted unless you are presenting
Four presentations from BDAWG members
Q&A will be after the conclusion of all presentations
Add questions to the chat
Recording and slides will be made available following the Colloquium

4 of 37

A few things to keep in mind...

We are abiding by the DLF Code of Conduct:

“The Digital Library Federation (DLF) is committed to creating and supporting inclusive, diverse, and equitable communities of practice. We strive to be a welcoming organization and the focal point for a digital library culture that is anti-oppression, recognizes intersectionalities, and works compassionately across difference. Together, DLF members advance research, learning, social justice, and the public good through the creative design and wise application of digital library technologies. We know that the best problem-solving and critical thinking happens when people with a wide array of experiences and perspectives come together to work in comfort and safety as peers. We therefore expect participants in the DLF community to help create thoughtful and respectful environments where that interaction can take place.”

https://www.diglib.org/about/code-of-conduct/

5 of 37

Introduction to DiAGRAM: the Digital Archiving Graphical Risk Assessment Model�

David Underdown, Alex Green and Hannah Merwood

20 August 2020

Image copyright of the University of Warwick

6 of 37

7 of 37

Why do we need a risk model for digital preservation?

Our digital heritage is rich, complex and fragile. It's under threat from:

- Rapidly evolving technology

- Outdated policies and standards

- A skills gap across the archives sector

Photo by Marvin Esteve on Unsplash

8 of 37

What are we doing?

We are creating a new risk management framework for digital archivists on digital preservation risks.

This new model will:

Improve users' understanding of the complex digital archiving risk landscape and of the interplay between risk factors.
Empower archivists to compare and prioritise very different types of threats to the digital archive: from software obsolescence to natural disaster.
Aid in quantifying the impact of risk events and risk management strategies on archival outcomes to support decision making, communication with stakeholders and developing business cases for targeted action.

9 of 37

Who is involved?

Joint project between The National Archives and the Applied Statistics & Risk Unit, University of Warwick

UK archive partners:

- Dorset History Centre

- Gloucestershire Archives

- TfL Corporate Archives

- Special Collections, University of Leeds

- Design Archives, University of Brighton

The Digital Preservation Coalition for dissemination events and project evaluation

This project is receiving funding from the National Lottery Heritage Fund

10 of 37

How to build the model

Conditional probability graph for the node Identity in DiAGRAM

Copyright, University of Warwick

Can learn the structure and probabilities if you have a large and complete data set using ML techniques.

Structure

Probabilities

Drawing of the structure mid development

11 of 37

Digital Preservation Bayesian Network

Digital preservation outcomes

Renderability: The object is a sufficiently useful representation of the original file.

Intellectual control: Having full knowledge of the material content, provenance and conditions of use.

Copyright of the University of Warwick

This is our Bayesian network of digital preservation risks. Each ‘node’ represents an underlying threat or important variable in calculating the overall risk. The precise definitions are all contained within the application.

Ultimate aim is for renderability and full intellectual control. All other factors can affect these. Sometimes they impact both, sometimes only one.

As an example, lets look at the node called ‘Obsolescence’, highlighted now in yellow on the middle left hand side of the network. The definitions is as follows:

Essential equipment, hardware or software required to access the bit stream becoming out of date or unusable. For example, hardware no longer being produced or essential software to access the media no longer supported.

This depends a lot on the storage medium used. For instance, D-1, an early Digital Video format, is far more at risk of obsolescence than an LTO tape. But if you have good technical skills you may be able to recover the bit-stream. This is why ‘storage_medium’ and ‘technical_skills’ are its parents. An object can only be bit-preserved if you do have access to the bit-stream, therefore this feeds into the probability of bit-preservation, which then feeds into the probability of renderability.

A large model can be treated as a collection of smaller local subsystems.

12 of 37

What is expert elicitation?

Investigate

Discuss

Estimate

Aggregate

Range graph from round 1 of the elicitation exercise

Copyright of the University of Warwick

“Expert elicitation refers to a systematic approach to synthesize subjective judgments of experts on a subject where there is uncertainty due to insufficient data, when such data is unattainable because of physical constraints or lack of resources. “ From http://www.nusap.net/downloads/reports/Expert_Elicitation.pdf

We followed the IDEA protocol with a trained facilitor. Process is designed to reduce bias and group think.

22 experts from 10 different organisations.

Investigate: Background information sent in advance. Full discussion of each question to get a common understanding of how it should be interpreted and answered. All experts then answer this individually.
Discuss: Follow on discussion of anonymised visualisation of responses. Sometimes everyone agreed, sometimes there was a big variation. Discussion on why some people may have answered high, some low. Aim is not to achieve a concensus, just open dicusssion of ideas.
Estimate: All participants make second and final estimate. This is the one that will be used.
Aggregate: Final probability for the model is a combination of all their estimates, weighted by their performance on the calibration questions. Calibration questions are used to determine individual accuracy and informativeness and were embedded amongst the questions we needed answers to so experts answered them all in the same way.

13 of 37

https://nationalarchives.shinyapps.io/DiAGRAM-dev/

Puts all this into practice, answer nine questions about your own archive, then investigate your risk levels, see how different actions will change your risk levels

https://www.nationalarchives.gov.uk/information-management/manage-information/preserving-digital-records/research-collaboration/safeguarding-the-nations-digital-memory/

DiAGRAM:The Digital Archiving Graphical Risk Assessment Model

14 of 37

Part and Parcel:

Weaving Born-Digital Into the UC Guidelines for Efficient Archival Processing

Kate Dundon�Supervisory Archivist�UC Santa Cruz Library

Shira Peltzman�Digital Archivist�UCLA Library

15 of 37

Background

�Guidelines for Efficient Archival Processing in the University of California Libraries

First published 2012
Revised 2018-2020
Defense of efficient processing holistic collection management
Version 4 published 2020

16 of 37

UC Guidelines Revision �Project Team

Elvia Arroyo-Ramirez

Jolene Beiser

Courtney Dean

Kate Dundon

Audra Eagle Yun

Jasmine Jones

Zachary Liebhaber

Charlie Macquarie

Laurel McPhee

Lara Michels

Shira Peltzman

Liz Phillips

17 of 37

‘Born-digital processing is archival processing’

Scant guidelines in the original 2012 version meant that we didn’t have much to go on�
After initially deciding to write the section from scratch, it became clear that a standalone section for born-digital didn’t make sense �
We proposed a new plan to interweave born-digital processing guidance throughout the document rather than addressing it separately

18 of 37

Overview of salient recommendations

Introduction and overall:

Acknowledge the need for a digital preservation policy
Ground the Guidelines in a more holistic approach to archival stewardship that emphasizes the interrelatedness of each stage in the archival lifecycle

Accessioning:

Address the reality that this means something very specific when it comes to born-digital archival materials
Provide baseline steps for born-digital accessioning

Appraisal/deaccessioning:

Emphasize the importance of appraisal prior to accessioning

19 of 37

Overview of salient recommendations, continued

Processing:

Encourage logical file transfer; discourage disk imaging except when appropriate
Push back on idea that it’s possible to easily quantify born-digital

Emphasize the challenges inherent to using metrics to guide/estimate born-digital processing

Make distinction between legacy media vs. contemporary files

Underscore that they should be treated differently with different expectations

Description:

Recommend noting the inclusion of born-digital material at the collection level whether or not it has been processed
Cross reference advice in UC Guidelines for Born-Digital Archival Description

20 of 37

bit.ly/UC-Guidelines

Guidelines for Efficient Archival Processing� in the University of California Libraries

Version 4�May 2020

21 of 37

Thanks!

Kate Dundon�Supervisory Archivist�UC Santa Cruz Library

Shira Peltzman�Digital Archivist�UCLA Library

22 of 37

Virtual Reading Rooms and Leveraging Shifting Priorities

Annalise Berdini, Digital Archivist

Kelly Bolding, Project Archivist for Americana Manuscript Collections

Princeton University Library

23 of 37

How we got here

Firestone Library

Reading room laptops
Very little online content

Mudd Library

Links from finding aid to online content in Webspace -- about 1500 components
Manual intervention for non-online content

24 of 37

Obstacles

Reading room computers difficult to manage efficiently

Process for adding new content
Capabilities lacking for local network
Public services workflow issues

Webspace was retired

Not very user-friendly to begin with
1000 digital objects migrated to Figgy, about 200 remaining complex objects that can’t be ingested at this stage

Digitized materials prioritized over born-digital

Lower use / no usage data

Low institutional risk tolerance
Library IT staff busy working on many other simultaneous projects

ASpace
New FA site
Remote support

Technical

Organizational

25 of 37

Pandemic’s Here,

All Bets Are Off,

Might As Well Put Stuff Online

26 of 37

Virtual Reading Room Task Force

Born-digital access project underway beginning Dec. 2019 -- mostly led by Digital Archivist
Pandemic shifted priorities; timeline for project shifted by department leaders in May 2020 and expanded project scope
Task Force formed June 2020 with goal of creating virtual reading room by Sept. 2020

Members:

Two archivists from tech services
Two archivists from public services
Curatorial rep
Head of Technical Services in SC
Library IT Project Manager

27 of 37

Figgy

Homegrown DAMS originally developed to provide access and long-term preservation to digitized materials
Expanded to include born-digital after Webspace sunset
Allows for tiers of mediated access, custom metadata, automated DAO insert to finding aids, viewer for all PDFs, TIFFs, and MP3s
Other features include OCR and automated metadata extraction from finding aids and catalog

28 of 37

In Development: Controlled Digital Lending

29 of 37

Where we’re headed

30 of 37

Thanks! We’d love to hear from others working on similar efforts and we’re happy to answer any questions.

aberdini@princeton.edu

kbolding@princeton.edu

31 of 37

Ethical Online Internships: Considerations and Methodology

Sheridan Sayles, MA, MLS, DAS

Brian Real, PhD�

32 of 37

Southern Connecticut State U: Context

Online MLIS Program

Recently (re)accredited by ALA
Only MLIS program in Connecticut, with none in Vermont, New Hampshire, or Maine

Internship Was Not Online Before COVID-19

Sudden need to find online placements
Most prior host organizations (i.e., public and school libraries) could not accommodate our students
Geography is no longer an issue

Basic Requirements

Three credit hours and (approximately) one semester
120 hours of site time plus 30 hours designated for academic work
“Unpaid,” but that will change

33 of 37

Southern Connecticut State U: Placement Sites

Seton Hall University Archives
Tribesourcing Southwest Film
Institute for American Indian Studies
National Baseball Hall of Fame
Connecticut State Library

Middletown Library Support Center
State Archives

Tech4Teens Summer Camp

Run by nonprofit group Concepts for Adaptive Learning

34 of 37

Seton Hall University: A Context on Interns

Ethics:

Start with looking for “gaps in resume”
Interns must learn

Connect work experience to theory from classroom

Supplement work experience with mentorship

Weekly check ins over Teams

Major decisions come from us, show intern tools on how we got there

Project:

Web crawls (digital collecting and accessioning), connecting captures to finding aids (arrangement and description), add tags to digital project (linked data)

35 of 37

36 of 37

Questions?

Sheridan: sheridan.sayles@shu.edu

Brian: realb1@southernct.edu

37 of 37

Thank you for joining us today!

Questions?