The One Great Scholarly Search Engine has become something of the Holy Grail of indexing.. (Larson and Arret, 2001 in Willinksy, 2006).
Full-text Global
Search Globally, Retrieve Locally
If you can’t find
the Holy Grail, create it!
A White Paper proposing Full-Text
Global
·
The objective of this white paper is to propose a concept, business model and technology - Full-Text Global (FTG) - for further collaboration and development requiring expertise beyond that of the idea’s author. No special consideration is sought, other than inclusion in the project if reaches the next phase – seed funding for development. It is presented here to develop discussion around the concept in principle and how it could be implemented, inviting comment from those with varying expertise. Specifically, the goal is to work towards developing:
·
a single,
comprehensive search application customizeable for all varieties of access
privileges to online scholarly content encountered worldwide at libraries
·
ensure that
this search application indexes with a logic of ‘most direct route to highest
quality form of all available material according to the user’s preferences and
privileges’ – including all OA material and all material accessed through
library privileges.
·
Provide a
search application indexing all titles and providing a filtered search to all
full-text freely available materials (OA); available free for the ‘subscriptionless
researcher’; an individual anywhere in the world with no special privilege
beyond access to the internet.
·
Provide
access to high-quality index and search to all people, taking advantage of
increasing availability of gratis material through OA and other access programs
that vary geographically.
The
concept is geared towards peer-reviewed scholarly journal articles, however
depending on demand it could be expanded to include all types of library
material, with P/R research a filter option.
Rationale - Toward for a Commercial
Open Source Software enterprise for a Federated Global Search of Scholarly
Resources
v
Available to
all individuals worldwide with an internet connection
v
Customizeable
for Libraries worldwide that grant access privileges to members
v
Maximizing
access and uptake of research in the Majority World
The
world increasingly has access to full-text scholarly articles online, but as
yet has no comprehensive and centralized search engine that can deliver
full-text articles from the wide variety of routes to access that have emerged.
This is a particularly acute and
distressing problem for the developing world, where the benefits of open access
and concession access programs will largely be lost if scholarly material goes
unused due to the lack of a straightforward search engine that assists the user
to search-navigate-retrieve articles.
The rigour, quality and objectivity of any literature review aspect of
any research project is compromised if any article is selected in on the basis
of convenience sample, a distortion that is introduced when the search engine
used is either not comprehensive or not discriminating enough and buries
relevant research too far down to find.
Databases are limited by the contents in the database. Search engines can index metadata harvested
globally without technical limitations, but suffer from other limitations as
follows:
a)
They may not
be sensitive to the user’s access privileges by institution, country or region
(particularly non-library search engines).
b)
They may not
index open access full-text appropriately, particularly gold OA journals
(particularly library search engines).
c)
They may not
contain sufficient federated capabilities for filtering subjects, languages, peer-reviewed
vs. primary source materials etc.
In
essence, each provides a substantial degree of possibility of error and none
are sensitive to access inequalities in a way that minimizes their impact. The improvements provided by Serials
Solutions 360, in that regard, widen the information divide by serving to solve
only the problems of the largest research libraries globally. There are two types of error and two types of
access inequality that can be addressed by Full-Text
Global, a comprehensive and customizeable global federated search engine
for scholarship. Subscription-rich
patrons and subscription-poor researchers are prone to both errors.
Errors
1)
Type I error
– False Positive - Convenience Sample. Search is not sensitive enough, articles
are included just because they are available. This is common with Google
Scholar. Another possibility is that the refereed post-print is available to
the user by a less convenient route, but they use the more readily available
version that is not peer-reviewed, thus wasting the value of the peer-review
process.
2)
Type II error
– False Negative - Article Not Found. The search leaves out articles that are
relevant. This common to proprietary
databases that are limited by the database composition, one may be unable to
retrieve articles in searches such as Scholars Portal, PubMed etc. where articles
appear under same keywords in Google Scholar. With Google Scholar, many hits
point to pay-per-view (PPV) access.
There is no instruction on how to retrieve the article by other means
(library log-in, ILL, direct from author, document delivery service (INASP),
etc.), so the article is passed up. Subscription journals and OA journals may
not be retrievable within the same indexes and databases.
Access Inequalities
a)
Literature
Access – Researchers in developing countries belong to institutions whose
libraries typically carry few subscriptions.
There are within-country divides in the North and South as well, where
smaller libraries, rural and remote areas are at an access disadvantage. Open access creates a level-playing field for
all who have internet access. Concession
programs for developing countries such as HINARI make available thousands of
journals to research libraries, and eIFL and INASP programs also create routes
to access through bargaining. Much of
this material requires access management, a technical hurdle for the South,
both in set-up and in the context of low-bandwith internet access with frequent
interruptions that disrupt the user who has to log-in.
b)
Access to
indexing and search – Along with the cost of subscribing to journals that is
prohibitive, the cost of subscribing to indexing services provides a major
hurdle for research libraries in the poorest countries in the world. Despite the growth of OA and the concession
programs, uptake is hampered by the problems of indexing – search, navigation
and retrieval.
The
movement for open access emerged out of the state of the publishing market in
an online environment in the context of the ‘serials crisis’ (Willinksy, 2006). What has emerged is a system that creates
wider access to research while introducing distortions into the process of
searching, and these occur whether one is at a subscription-rich institution or
a subscription-poor one, or if one is not an institutional member at all. This situation has created the following
routes to search-navigate-retrieval.
1)
Your
institution’s subscriptions – search by various means and subscriber databases
and indexes. Comprehensive search with
Google Scholar is possible, but not federated, not sensitive to access
privileges.
2)
Green OA –
articles deposited in institutional OA repositories (over 3000) or archived on
author’s sites. Can be retrieved over
Google Scholar (green arrow or ‘all versions’).
May not be refereed version.
3)
Gold OA –
articles published in gold OA journals (over 4000), OR, in mixed access
journals (either post-embargo or by author fee open access). Retrieveable in various
databases, but indexing is not necessarily consistent between gold, mixed and
non-OA resources.
4)
UN Programs –
HINARI/AGORA/OARE – articles available to low-income countries research
libraries, available through access management at participating institutions. Article versions may also be in OA
repositories.
5)
eIFL –
articles available to research libraries in developing countries by eIFL
library consortia bargaining.
6)
INASP –
articles available through INASP PERii program.
By document delivery. African
Journals Online as well – document delivery.
7)
Publisher’s
Concessions – A number of large to small publishers provide free to heavily
discounted access to resources, some simply by recognition of IP.
8)
Physical
library resources and ILL – Interlibrary loan.
Depending on the library’s network of agreements for ILL.
9)
Author
e-print or re-print by request. If
articles can’t be accessed by any of the other routes, they may be obtained
directly from the author by request.
10)
Public domain
digitized.
11)
‘Grey market’
– articles may be obtained in ways that circumvent digital rights management of
the publisher, but on the other hand may constitute fair use on the part of the
end user. For instance, it was reported
that physicians in Thailand resorted to purchasing unauthorized copies of
medical articles in order to gain the knowledge they needed to save lives,
while they were in no position to pay the purchase price for the article or
subscription and had no institutional subsidy like most doctors here would.
In
terms of the Northern libraries who purchase a number of indexing services to
manage their resources, John Willinksy (2006) has stated that “Despite this
considerable array of indexes and portals, earnest scholars and students still
have to wend their way through overlaps, gaps, and partiality in the coverage
of the research literature that the indexes provide…”(p. 173). When we add OA repositories, OA journals, UN
programs, publisher concessions by IP address etc., one wonders if the
available wealth of resources is used effectively, and how often specific
resources are passed up because of search-navigation issues. One can predict that the more barriers to
retrieval, the more complex the search, and the more specified prior knowledge
required of the user, the greater chance there will be an overall reduction in
the quality, objectivity and cost (in time) of literature search and
consequently the research endeavour itself.
Full-Text
Global Business Model
‘‘Having one place to search that would include relevant resources
would make research less fragmented’’ (Larson and Arret, 2001 in Willinksy,
2006).
Full-Text Global would be a commercial open source software (such
as Sugar CMS) that would serve to address indexing issues is a mixed
OA/subscription environment with particular attention to providing a truly
global service with regards to the opportunities and needs of developing
country institutions. It would also
provide a unique service accessible to anyone with an internet connection, with
its basic application for federated global article search and global full-text
search. In that sense, it should serve
people equally regardless of their location in the world, in the sense that it
maximizes everyone’s access privileges and reduces the impact of access
privilege inequality. While a pure OA
environment is ideal and may be inevitable, the need to manage knowledge
resources in a mixed environment provides the possibility of making this
product commercially viable.
Full-Text Global will be a company that believes in a vision for
Open Access, but provides a service for maximizing access in a world
complicated by access barriers. It addresses
the primary access inequality - access to literature - indirectly by addressing
the second – access to indexing and search.
This is because the privilege possibilities as they exist today and as
they will increase in the future risk under-utilization because of the
complications of search and retrieval.
In addressing the indexing issue, FTG also serves to solve problems
faced by the large research libraries of the North and provide a service that
treats all articles on the basis of the preferences of the user and the route
to retrieval, thus harnessing the full power of purchased and OA resources. It should serve to provide the most direct
route possible to free access to highest quality version of any article in the
world. It will do this by managing the
routes to access and deductively and logically providing the link to the
article, or information on the most direct alternative route listed above
(except grey market!).
The
non-profit business model would be as follows:
1)
Commercial open
source software (Creative Commons-type license) for a federated and layered
search - similar to Serials Solutions 360, but with attention to ‘searching all
resources available to the patron by any route’ as opposed to ‘searching the
library’s acquired resources’.
2)
Federated
Search Service for Institutional Members - Customization service to integrate
access privileges of institution, so that members can retrieve full-text available
to them by privilege or by OA, from the comprehensive global search while logged
in (without having to search Google Scholar and then return). This puts the customized privilege-sensitive
search engine and results page on your library website, giving your patrons the
best possible search capacity. The
federated search would have three levels, two of which are common to the open
version and one which is an option for libraries.
a.
A single
search bar, all-in-one keyword search.
Simple, and effective (depending on purpose on strategy and keyword
skill of user).
b.
An advanced
search filtering for advanced keyword, subject, author, title, discipline,
years etc.
c.
A specified
search allowing the user to focus in on databases, and apply specific knowledge
of them to their search.
Finally, the search results page, customized to
the library patron’s privileges will provide information on the best way known
to retrieve any article (except grey market!).
3)
Support
Services to these libraries
4)
Service for
Access Management and indexing/search (2+3 above) to consolidate privileges
through HINARI/AGORA/OARE, eIFL, INASP, OA + publisher concession access in the
South. Payment scale model based an ability of
institution to contribute - with contribution from UN, eIFL, INASP
organizations. Sponsorship and shared programs
for North-South-South partners and networks (to outfit entire network of
partners with net contribution from them.)
5)
Federated
Open Access Search. For general internet
users who have no additional access privileges. Will accept donations and generate a community
of developers. This is essentially the basic software and one that can be
skinned to any website. One can turn on
and off a filter to go from comprehensive abstracts/article listings to only
results that provide a full-text. Full-text
search means that anyone in the world who can get on the internet will have a
global federated search that indexes every journal article available to them
(only hits that lead to full-text gratis articles are included, in essence
creating the widest library possible for any general user at any given time.
But there are several reasons for filtering back in the comprehensive index of
titles.
a.
Your library
does not yet have Full-Text Global,
and you want to use the open version to find the article and then return to
log-in to retrieve it (as people often do with Google Scholar).
b.
You want to
ensure that your literature search is not biased by availability and
convenience, to know what exists and/or to try alternative routes to
access.
6)
Provide both
paid and open search services in as many languages as possible.
7)
Provide a
scholarly and principled guide to literature-based research, focused
on a logic for routes of retrieval for online materials but also utilizing your
closest physical library, ILL, author re-print, document delivery etc. etc. in
as many languages as possible.
8)
Provide a
data-based map and accounting of the world’s scholarly resources.
9)
Phase
II. Work on web 3.0 Semantic search
capacities, and research and
development into higher sophistication in computer-automated translation
services, towards providing a service of automated computer translation plus
human translator peer-review to accelerate high-quality translation of the
world’s scholarly resources into as many languages as possible.
The
advantage for endowed libraries in the North in choosing Full-Text Global’s business model to provide the comprehensive
search for online resources available to your patrons:
a)
the
commercial open source commercial not-for-profit model builds sustainability
into your purchases:
a.
no
proprietary rent-seeking and no legal hassles with regards to patents
b.
freedom to
build internal capacity to maintain and modify the software
c.
any
programmer globally can work with standardized open source code, and
communities develop around open source platforms - you are not reliant on any
one company for trouble-shooting, maintenance and improvements.
d.
you pay for
services, not for software and intellectual property.
e.
payments
cover costs for operating Full-Text Global, and ‘profits’ are invested in the
social mission.
b)
Full-text Global capitalizes on open access resources by integrating them fully into
your indexing services used by your patrons. Even though your library did not
have to acquire OA materials, this does not change their value to your patrons. However, material that is green and gold OA and
non O may not be searcheable simultaneously and seamlessly with information in
the result page providing the most direct route to the highest quality version
available to the user and all access options.
Your patrons’ can log-in one time, and be assured of a single,
comprehensive and federated search for all materials available to them, and
though privileges differ between individuals, the service that maximizes the
use of those privileges for anyone in the world would be Full-Text Global.
c)
Full-text Global has a Social Mission that improves the global environment for
librarianship, with long-term benefits to your research library.
a.
Improved
access to literature and to search and retrieval through indexing worldwide
directly translates into increased production of knowledge resources by all
countries in all regions, a capacity currently well below potential in the vast
majority of the world’s societies. We
may be described globally as operating at marginal capacity today due to the
information divide.
b.
Produced in a
research and archival context, these resources will become available to your
library’s patrons, likely at no cost through OA or at an affordable price. This means more unique, diverse, authentic
global sources for your patrons.
c.
By purchasing
Full-Text Global services, you will
be contributing to indexing access and Open Access’ uptake for everyone,
regardless of subscription, through supporting the open version of Full-Text Global.
Reference:
Willinksy, John. 2006. The Access
Principle. MIT, Cambridge Press.
