18 · Survey Results: Executive Summary
the gift agreement or some other policy document is
highly recommended.
While 71% of respondents have policies regarding
whether files with PII should be retained with restric-
tions or destroyed, only 43% have policies indicating
whether born-digital materials can be made available
for research use before they are screened for PII. One
respondent’s comment that “all special collections ma-
terials have personally identifiable information (PII)”
is quite true. However, paper-based collections have
always benefited from security through obscurity.
There is no fast or easy way to uncover social security
and credit card numbers in paper-based collections.
With born-digital records, on the other hand, there are
many tools available that can search and locate PII,
even in deleted or hidden files. Such content, improp-
erly managed, not only puts the file creator at risk, but
also may be in violation of an institution’s security
and privacy policy for this type of information. Eighty
percent of respondents indicated that they do not
have a written PII policy. Greater security is needed
for unscreened born-digital records, especially if they
are stored on networked servers.
The responses to this survey indicate that many ARL
libraries and archives have begun working with born-
digital materials in their collections, despite the fact
that enterprise level systems and best practices for
managing these materials in an archival setting are
still in development, and despite concerns that they do
not have the resources to scale their work to meet cur-
rent and future demand. This willingness to experi-
ment, to learn new skills, and to seek to understand
the scope of the issues is building expertise within
the library and archives profession, and has insured
access to some born-digital holdings, at least in the
near term. It also signals a shift from a wait-and-see at-
titude to a more empowered something-is-better-than-
nothing approach to managing born-digital materials.
Respondents identified the following as critical for
transitioning their work with born-digital materials
from projects to programs:
Collaborative solutions for dealing with
hardware and software obsolescence.
More, and more appropriate, storage for
born-digital materials (long-term, authen-
ticated, secure, verified, backed-up, and
geographically distributed). As one re-
spondent noted, “Archives are guaranteed
preservation only if stored on enterprise
data storage.”
Automation of as much of the workflow as
Asset-level access control to enable tiered
access to restricted records.
Many institutions are working with digitized
content or licensed digital content and are only now
beginning to explore the ways in which born-digi-
tal, primary-source materials may be different. For
example, it is difficult to estimate storage needs for
born-digital primary sources stored on legacy media
prior to accessioning and processing them. Privacy
concerns are magnified when large bodies of easily
searchable digital material may contain personally
identifiable information. The workflows and infra-
structure built for digitized content are often insuf-
ficient for born-digital primary sources.
While some special collections rely on a single
staff member to manage all aspects of preserving
and providing access to born-digital materials, more
frequently staff from special collections, library IT,
digital repositories, digital curation, and other areas
work together to ingest, appraise, describe, preserve,
and provide access to this content. The distributed
nature of this model allows the library to leverage
existing expertise, but it may also mean that no one
has the big picture. These situtions make it difficult
to track the resources needed to manage the materi-
als—which then makes it difficult to estimate current
and future costs. Distributed responsibility can also
threaten the long-term survival of the materials, either
when no one feels empowered to make decisions or
when someone makes decisions without having all of
the relevant information. Staff need models of existing
Previous Page Next Page