Institutional Repositories · 17
it is based on) is used on all the vendor-hosted
platforms.
There is a greater variety of hardware in use.
Implementers are about evenly divided between
Intel-based servers (Dell models in particular) with
either Linux or Windows operating systems and
Sun servers with Solaris. Only one uses an Apple
Xserve running OS X. All but two of the planners
use Intel-based servers, again primarily Dell
models, with Linux or Windows operating systems.
One uses an IBM RISC server and the other a Sun
system.
Memory on the implementers’ Sun systems
ranges from 2 GB to 64 GB and disk storage rang-
es from 10 GB to 6 TB. (One institution reported
a 12 TB storage unit, but it was not dedicated to
the IR.) The Intel-based servers have memory rang-
ing from 512 MB to 4 GB and disk storage ranging
from 28 GB to 3 TB. The Macintosh server has 1 TB
of storage space. The planners’ Intel-based servers
have memory ranging from 512 MB to 12 GB and
disk storage ranging from 108 GB to 3 TB. The IBM
RISC-based server has 4 GB of memory and 275 GB
of storage space.
Roughly a third of respondents have made no
modifications to the IR software and another third
have made minor modifications. About 20% of
both implementers and planners have made ma-
jor modifications to IR software. Implementers are
much more likely than planners to have made fre-
quent changes, 22% vs 6%.
Policies and Procedures
Seventy-five percent of implementers and 71% of
planners indicated they have or will have writ-
ten policies and procedures for their IRs. For both
groups, 54% have submitted their policies and pro-
cedures to an institutional authority for review, or
are planning to do so. Most of those who identified
the reviewing authority indicated that their policy
documents went to the University Counsel.
Respondents place a wide variety of materials
in their repositories. Electronic theses and disserta-
tions are the most common type of deposit (imple-
menters, 67% planners, 79%). Articles, including
preprints and postprints follow closely. The major-
ity of respondents include official published ver-
sions of articles, conference presentations, technical
reports, and working papers. Only a few include
university catalogs, yearbooks, or alumni publica-
tions.
Only a handful of respondents are actively ne-
gotiating with publishers to secure permanent de-
posit of e-prints from published serials, but 46% of
implementers and 63% of planners are considering
doing so in the future.
The widespread inclusion of traditionally un-
published material in IRs may reflect the relative
ease of recruiting this type of content as well as the
fact that these materials in print format do not have
robust publishing avenues. Data sets, learning ob-
jects, and multimedia materials are the most preva-
lent non-traditional materials deposited, with over
a third of all respondents indicating they include
or will include these materials in the IR. Several
respondents mentioned using the IR to house ret-
rospectively digitized images and other archival
material.
Seventy-four percent of implementers and 83%
of planners indicated that they accept any digital
file type into the IR, but relatively few (26% and
39%) are committed to functional preservation of
every file type. Eighteen percent of implementers
and 17% of planners will only accept and preserve
specified file types. A few accept certain file types
but do not preserve them. Several respondents
mentioned following the support levels outlined
in MIT’s DSpace guidelines (http://www.dspace.
org/implement/policy-issues.html#digformats),
which include full support and preservation for
common file types such as PDF, XML, AIFF for au-
dio, and GIF, JPEG, and TIFF for images, among
others.
Most deposits to the IR are or will be made by
authorized depositors (implementers, 89% plan-
ners, 79%). A significant number of IR staff also
it is based on) is used on all the vendor-hosted
platforms.
There is a greater variety of hardware in use.
Implementers are about evenly divided between
Intel-based servers (Dell models in particular) with
either Linux or Windows operating systems and
Sun servers with Solaris. Only one uses an Apple
Xserve running OS X. All but two of the planners
use Intel-based servers, again primarily Dell
models, with Linux or Windows operating systems.
One uses an IBM RISC server and the other a Sun
system.
Memory on the implementers’ Sun systems
ranges from 2 GB to 64 GB and disk storage rang-
es from 10 GB to 6 TB. (One institution reported
a 12 TB storage unit, but it was not dedicated to
the IR.) The Intel-based servers have memory rang-
ing from 512 MB to 4 GB and disk storage ranging
from 28 GB to 3 TB. The Macintosh server has 1 TB
of storage space. The planners’ Intel-based servers
have memory ranging from 512 MB to 12 GB and
disk storage ranging from 108 GB to 3 TB. The IBM
RISC-based server has 4 GB of memory and 275 GB
of storage space.
Roughly a third of respondents have made no
modifications to the IR software and another third
have made minor modifications. About 20% of
both implementers and planners have made ma-
jor modifications to IR software. Implementers are
much more likely than planners to have made fre-
quent changes, 22% vs 6%.
Policies and Procedures
Seventy-five percent of implementers and 71% of
planners indicated they have or will have writ-
ten policies and procedures for their IRs. For both
groups, 54% have submitted their policies and pro-
cedures to an institutional authority for review, or
are planning to do so. Most of those who identified
the reviewing authority indicated that their policy
documents went to the University Counsel.
Respondents place a wide variety of materials
in their repositories. Electronic theses and disserta-
tions are the most common type of deposit (imple-
menters, 67% planners, 79%). Articles, including
preprints and postprints follow closely. The major-
ity of respondents include official published ver-
sions of articles, conference presentations, technical
reports, and working papers. Only a few include
university catalogs, yearbooks, or alumni publica-
tions.
Only a handful of respondents are actively ne-
gotiating with publishers to secure permanent de-
posit of e-prints from published serials, but 46% of
implementers and 63% of planners are considering
doing so in the future.
The widespread inclusion of traditionally un-
published material in IRs may reflect the relative
ease of recruiting this type of content as well as the
fact that these materials in print format do not have
robust publishing avenues. Data sets, learning ob-
jects, and multimedia materials are the most preva-
lent non-traditional materials deposited, with over
a third of all respondents indicating they include
or will include these materials in the IR. Several
respondents mentioned using the IR to house ret-
rospectively digitized images and other archival
material.
Seventy-four percent of implementers and 83%
of planners indicated that they accept any digital
file type into the IR, but relatively few (26% and
39%) are committed to functional preservation of
every file type. Eighteen percent of implementers
and 17% of planners will only accept and preserve
specified file types. A few accept certain file types
but do not preserve them. Several respondents
mentioned following the support levels outlined
in MIT’s DSpace guidelines (http://www.dspace.
org/implement/policy-issues.html#digformats),
which include full support and preservation for
common file types such as PDF, XML, AIFF for au-
dio, and GIF, JPEG, and TIFF for images, among
others.
Most deposits to the IR are or will be made by
authorized depositors (implementers, 89% plan-
ners, 79%). A significant number of IR staff also
