Supplementary memorandum from Hallward
Library, University of Nottingham, SHERPA Project
How much, approximately, would it cost to establish
a national network of repositories?
The easiest way to establish a national network
of repositories would be for each higher education institution
(HEI) to set up its own repository or set of repositories. The
cost of setting up an individual repository for each institution
is small. The software is freethere are several packages
available, all of which are freely downloadable.
The software can be installed on a standard server, costing about
£1,500. It takes a computer officer between 2 and 5 days
to get it all working (at a cost of say £600 for the time).
Experience from the SHERPA project shows that
further customisation work is required during the first 6 months
of the repository's life. This includes developing the "look
and feel" of the service, and ensuring it is fitted into
the support and maintenance procedures of the institution. We
estimate this adds up to 15 person days' work over the course
of 6 months. A cost of about £1,800. Thereafter, ongoing
technical work can normally be absorbed within an institution's
standard maintenance routines.
Populating the repository has its own costs.
At present, the major costs are in advocacypromoting the
service and persuading academics to deposit articles in the repository.
This cost would, of course, be removed if there was a mandate
(from institutions, research councils, or other agencies) that
publicly-funded research had to be made publicly available in
this way. Research shows that most academic authors would be happy
to deposit their materials in a repository if this kind of mandate
was introduced. The costs for depositing items are given below.
The overall cost of setting up a national network
of institutional repositories would therefore be the costs for
each institution multiplied by 131 (the number of HEIs in the
UK). See below for a summary of costs. It should be noted that
some institutions already have repositories or have plans to install
How many repositories would the UK need in order
to achieve maximum coverage?
Since the entry barriers (including costs) are
very low, the normal assumption is that every institution which
produces research publications should have a repository. Most
institutions would prefer to have their own repository which they
can set up in line with local requirements. The repository can
act as a research management tool and as a "shop window"
for the institution's output. Most institutions already have the
organisational and technical infrastructures to set up and maintain
a repository easily. However, some institutions might for various
reasons choose to join a consortium with a shared repository.
Both models (individual institutional repositories and consortial
repositories) are being tested in the SHERPA project442. In addition,
the British Library, as a SHERPA partner, is setting up a repository
which will hold content from non-affiliated researchers.
A network of institutional repositories would,
of course, operate as a single virtual archive. When repositories
are set up using the OAI-PMH (Open Archives Initiative Protocol
for Metadata Harvesting),
they can be searched in a seamless way. Users may not even be
aware that they are searching the contents of multiple repositories.
The actual location of the content is in this sense irrelevant,
as long as it is in an OAI-compliant repository. Once the material
is available online, then other services can use itto present
subject-based collections, for example, or as a source material
for teaching. Repositories run by institutions will provide a
stable foundation on top of which other services can be built.
Research results can be consulted and also re-used in new ways
to enhance the return on the original research investment.
Having a repository in each research-active
HEI plus the British Library repository would give coverage for
most UK research output.
What level of resourcing would be required to
maintain the repositories?
The technical maintenance costs of a repository
would be minimal. Most could be easily absorbed into an HEI's
standard web services maintenance. Every 3 years or so there would
be a need to replace the kit and possibly upgrade the software
but the costs of this (hardware plus 2 to 5 person days) would
be no more than initial installation.
Populating the repository would be an ongoing
cost. Two scenarios are possible here. The first is that authors
archive their own material as and when it is produced. This is
a simple process which takes at most about 15 minutes for each
paper using a specially designed web interface (available as standard
in repository software). In this case, the costs are effectively
hidden as they are distributed amongst a large number of researchers
who each spend a small amount of time (1 to 2 hours per year)
depositing their own papers.
The second scenario is that someone is employed
to deposit items on behalf of researchers. Most research-led HEIs
would require a full-time member of staff, able to work across
subject-disciplines and co-ordinate the collection of material.
The cost of such a post would be up to £30,000 per year.
An alternative might be that departmental administrative support
staff could carry out the role for their department. In this case,
the amount of work would be distributed and most of the costs
would again be absorbed by the institution.
In reality, a mixed economy may develop, with
a variety of academic authors, research assistants, departmental
officers and library personnel depositing papers.
In the longer term (20 to 50 years), the big
costs are likely to be the preservation of the digital files.
These costs are, of course, not unique to repositoriesthey
apply to all electronic resources intended for long-term use (including
those currently produced by commercial publishers). Although some
work has been done on digital preservation costs (by the British
Library amongst others), this is still a rather unknown quantity.
It is a problem that needs addressing for all digital objects
held in repositories and everywhere else. It need not stop repositories
being set up and populated in the short term, however, as the
issueand costswill need to be faced at some point
whether the material is held in institutional repositories or
How practicable would it be to establish a network
of UK repositories?
Very practicable. Establishing the network would
be low cost and the expertise is already there in most institutions
to do it without delay.
It is already happening. In a recent informal
survey of members of SCONUL (the Society of College, National
and University Libraries) we found that of the 51 respondents,
17 (a third) already had set up repositories (mostly research-led
universities). A further 13 were expecting to do so in the foreseeable
future. This means that about 59% of the respondents already had
a repository or expected to have one in the near future.
The main challenge at the moment is not setting
up the repositories per se but populating them. Academics do not
currently have many major incentives to archive their material
(or at least they are unaware of the benefits of repositories).
Mandating the depositing of research papers produced by HEIs in
open-access facilities would be a practical way to overcome this
problem. It would mean that academics or their institutions would
be required to retain sufficient rights in their work to deposit
it in repositories.
Costs are given per institution (there are 131
HEIs currently in the UK). It is assumed that many of the maintenance
costs of institutional repositories can be absorbed by institutions.
The costs given are minimal; higher costs would, of course, apply
if institutions wished to enhance the service in various ways.
|Installation (5 days)||600
|Customisation (15 days)||1,800
|Total per institution||3,900
|Ongoing maintenance costs:||
|Technical support||Absorbed by institutional IT services
|Supported archiving service||£30,000 per year
|Upgrades/migrations||£3,900 every 3 years
|Digital preservation||Significant costs (applies to all digital objects)
For example, EPrints.org software-see http://www.eprints.org/ Back
JISC funded project to co-ordinate the building of institutional
repositories in project partners' institutions-see http://www.sherpa.ac.uk Back
Open Archives Initiative Protocol for Metadata Harvesting (OAI-PMH)
defines common standards for storing and searching material in
repositories-see http://www.openarchives.org/OAI/openarchivesprotocol.html Back