Information Technology Reference
In-Depth Information
medium term.
A variation of this scenario, suggested by some commentators, is that
journals would no longer fulfil the dissemination function at all, and all
access to journal articles would be via repositories (which presumably
would need to be fully interoperable). Journals would still provide
certification through the combination of peer review and association
with their brands. In this case the outcome would be the same, except
that dissemination, preservation and curation costs would not be
duplicated.
Probability of dominance: High for some subject disciplines, low for others
Benefits of OA: Fully achieved
Preservation: Journals and repositories
Costs: Possibly lower than in Scenario 1, but higher than at present, assuming
author-side payments amount to roughly the same as subscription outlay and
there is duplication of access provision and preservation by journals and
through a repository infrastructure
Result: Order
Scenario 3: substitution
In this scenario widespread manuscript archiving leads over time to
large-scale cancellation of subscriptions to journals. No funding is made
available for author-side payments and so journals go out of business,
and with them disappears the certification provided by the combination
of peer review and journal brands. However, given that 'necessity is the
mother of invention', a new mechanism of certification quickly emerges
to substitute for that provided by journals. It is difficult to speculate
what this mechanism might be. Jensen (2007) suggests that 'heavily
computed reputation-and-authority metrics' will form part of what he
terms 'authority 3.0', but the details of any replacement system are, of
course, unknown. It is therefore impossible to guess how widespread a
new system could become and what it might cost. However, the costs
of preservation and curation would fall only on the repository, so there
would be at least a potential for the costs to be lower than in the current
system, though once again each article would be likely to appear in
multiple repositories. This scenario would offer the potential for the
benefits of OA to be fully realized and, because a suitable alternative for
