SPEC Kit 340: Open Source Software  · 13
The most frequently adopted OSS systems
include institutional repositories (52 respon-
dents), blogging (50 respondents), digital
preservation (47 respondents), and publish-
ing (40 respondents).
OSS Adoption
Seventy-four of the 76 responding libraries (97%) re-
port having adopted open source software. Of these,
only five (plus one parent institution) have a formal
written policy related to adoption of OSS. Twenty-
five libraries (34%) have an informal policy, but the
other 43 (59%) have no OSS adoption policy. Several
respondents reported that policies were currently be-
ing created, but could not be shared at the time of their
Most respondents indicated their institution
had neither a sustainability strategy (50 of of 71 re-
spondents, or 70%) nor an exit strategy (53, or 75%).
Reported strategies include minimizing customiza-
tions, providing sufficient staffing with needed exper-
tise, and only adopting systems with good documen-
tation and an active community. More than half of the
respondents who commented on their exit strategy
emphasized the criticality of data migration (8 of 15
relevant comments).
Survey respondents were then asked to identify
the system they had most recently adopted and to
provide the number of staff and hours required to im-
plement that system. A wide variety of projects were
identified, the most common being Drupal, Blacklight,
Omeka, and DSpace. Respondents reported from one
to eight staff members dedicated to implementation,
with a mean and median of three staff. The number
of hours required for initial implementation varied
dramatically, ranging from 0.75 hours to 9,000 hours
with a mean of 573 hours and a median of 160 hours.
Respondents were asked to identify the open
source system they most recently adopted that is still
in production and to describe the resources needed
to support that system. For most respondents, the
system referred to in this question was the same sys-
tem described in the implementation question above.
The number of staff required to maintain this system
ranges from 0 (for a digital exhibit) to 10 (for a CMS)
with a mean of 2.1 and a median of 2. The number of
hours required to support the same system ranged
from 0 (for the exhibit) to 512 (for a digital repository)
per month, with a mean of 68 hours and a median of
20 hours.
Only 16 of 72 respondents (22%) were able to track
the costs of either adopting or contributing to an OSS
system. Ten respondents who could track the cost
of their most recently adopted OSS system reported
that expenses ranged from $400 to over $600,000 and,
in some cases, represented a multi-year investment.
These funds covered a variety of expenses including
staff time, hosting, travel, and consulting. The nearly
universal source of funding for adopting or contrib-
uting to an OSS system was the library’s operating
budget (69 of 70 respondents, or 99%). A few had ad-
ditional funding from grants, their university, or a
consortium. One ArchivesSpace project received only
consortium and grant funding.
The survey asked respondents to describe three
benefits and three challenges associated with adopt-
ing OSS. The most common benefit is the ability to
customize the software (50 responses). Other common
themes include low cost or time to implement (27 re-
sponses) and the association with an active commu-
nity (27 responses). The most common challenge was
the need for highly skilled staff who could provide
support for the OSS system (40 responses). Other com-
monly cited challenges include poor documentation
(19 respondents), a need for additional training or
expertise (16 respondents), and substandard develop-
ment practices (12 respondents).
OSS Contribution
Fifty-six of the responding libraries (78%) have contrib-
uted resources to an open source project. The number
of projects contributed to by each library ranges from 1
to 20, with an average of 4.6 and a median of 3. Thirty-
two libraries report being the primary code contribu-
tor for at least one project; a different set of 32 libraries
(with significant overlap) identified themselves as the
original developer of an open source project.
Commonly reported examples of projects in-
clude DSpace (12 respondents), Fedora (11 respon-
dents), Hydra (9 respondents), Kuali (6 respondents),
Previous Page Next Page