The SHSSERI Collaborative KnowledgeSpace : a New Approach to Resource Fragmentation and Information Overload

Humanities scholars today are faced with information overload created by the explosive growth of resources
and services on the web, by globalised instant
communication and by the spawning of new personal and intellectual networks outside the confines of discipline and geography. The pace and scale of developments is both exciting and challenging, necessitating the adoption
of new strategies which capitalize on the potential of
digital methods.
The issues
The explosion in information has not been matched by developments in the conceptual framework and
tools we use to manage information, by availability of
digital infrastructure or by the widespread adoption of new methods (other than the basic generic tools of email, wordprocessing and web browsing). Where we have adopted digital tools it is often within existing structures – faster communication, easier preparation of publications, easier access to library catalogues and content – rather
than in novel approaches to research, teaching or
collaboration. Uptake of digital methods in the Humanities
is hindered by lack of funding for infrastructure
development, by the richness and heterogeneity of our domain, by the lack of agreed methods and classificatory
systems and by our relatively slow adoption of
technology.
If we are to overcome the problems of information
overload we must develop – with very limited resources
– e-Research infrastructure and tools which support
and enable our tradition of individual scholarship and
interpretation, rather than the method- and data-driven tools of the sciences (or commerce). Our tools must
reflect the needs of Humanities scholars – low entry
barriers and intuitive structures which reflect the richness and complexity of domain practices. While technologically literate scholars will continue to build or adopt tools and find new ways to use them, how do we enable the rest of the discipline to understand and leverage their potential? The critical need is to develop and propagate authoritative information on Digital Humanities methods repackaged
in a form digestible by less technologically-oriented
Humanities scholars. We need to get the message out.
SHSSERI Collaborative Knowledge Space
The Sydney Humanities and Social Sciences
e-Research Initiative (SHSSERI) is a project to
develop e-Research tools and information which reflect
the needs of Humanities (and Social Sciences) scholars,
particularly tools and which move beyond the mere
collection and delivery of digital information into active
engagement with research methods and the structure of academic communities. Our aim is to provide a resource which becomes a ‘must-have’ by providing a single point of access to scattered resources, tools to manage those resources and authoritative information which shortcuts the process of adopting digital tools (and increases the chances of success). Building on existing work at the University of Sydney and using Open Source software,
we are developing a one-stop-shop – the SHSSERI
Collaborative KnowledgeSpace (CKS) - to support the information, communication, data management, analysis and archiving needs of Humanities and Social Sciences researchers.
In developing the SHSSERI CKS we have identified as our highest priority the management of basic research
information (references, bookmarks and notes) and access
to authoritative information about the practice and potential of the Digital Humanities. A key focus is the integration
of resources. Most researchers store bibliographic
references, internet bookmarks and research notes in
separate systems (eg. EndNote, Firefox and a Blogg –
or on paper). Their CV may be in Word, research
observations in Access, research expenditure and grad student details in Excel, email in Outlook, travel
arrangements in a corporate system, … the list goes on. Our resources are all too often tied to the desktop of a specific computer, scattered, fragmented, unfindable,
inconsistent, redundant, unlinkable and insecure. We work
inefficiently, waste time and effort, and stress out trying
to manage these disparate, inconsistent resources.
TMBookmarker
The TMBookmarker application (name undecided) tackles the management of bibliographic references,
internet bookmarks and note-taking, as well as access to authoritative sources of information on the Digital
Humanities. TMBookmarker is a web-accessible
knowledgebase which will handle conventional
bibliographic information, internet bookmarks and note-taking in a single integrated database. It aims to replace all these forms of machine-specific or special-purpose referencing with a single, integrated searchable database available anywhere one has access to the Internet – from University desk to Internet café in Khatmandu.
In addition to providing consistent anywhere-access and capture/annotation of notes, bookmarks and bibliographic references, the database will generate selective lists for inclusion in web sites, course readings or bibliographies,
and allow keywords and annotation to be attached
to any resource. It aims to reduce a multiplicity of
special-purpose programs, information folders, bloggs, bookmark lists and hand-built web pages to a single, easily understood, web-accessible resource. It will open new avenues of possibility for those who have not
ventured into the blogosphere, mastered EndNote or
managed to migrate their bookmarks from one machine to the next.
Social bookmarking
TMBookmarker also implements concepts of
social bookmarking using a structure we are calling
a databliki (database + blogg + wiki). We allow users to discover and share bookmarks/references through the database, while attaching their own personal notes and classifications to them. We provide wiki-based and blogg-based public editing and annotation of references,
allowing the community of scholars to participate in
expansion and refinement of the reference database, and the development of additional knowledge around the core resource. By mining the database we can identify patterns
of bookmarking and knowledge development which link people with similar interests and use the patterns to
provide relevancy-based searches of the database. Unlike generic social bookmarking systems such as del.icio.us or CiteULike, the SHSSERI CKS identifies users by name and institution. This provides a mechanism for identifying colleagues with similar interests and the
serendipitous discovery of relevant references through folksonomic tagging tied to specific communities of users. We aim to provide improved folksonomy-based searches by calculating co-occurrence of tags across the database, independent of their actual text values, and using these
correlations in combination with user ratings against
specific user-defined groups of colleagues and subject domains to generate targeted relevancy measures.
TMBookmarker has many smarts such as instant web bookmarking, automatic DOI identification and lookup, reference import/export from common systems, RSS feeds, saved custom searches, custom list generation and wiki-style change tracking. It has been in use by a test group since Nov 2005 and should be ready for general release by second quarter 2006. It is written in php and MySQL. We intend to release the code as Open Source.
In this presentation I will explore the basic concepts of managing information through a collaborative reference/bookmark/annotation system rather than a conventional desktop reference management system. I will focus on the benefits of integrated access to information in a single
database, and the advantages of social bookmarking in
an academic referencing system, compared with less structured systems such as del.icio.us or CiteULike. I will also compare the social bookmarking methods developed with those of other academic referencing systems, such as Connotea, and present preliminary results on defining communities of interest through database mining. I will conclude with observations on the relationship between
social bookmarking and peer-review methods in
establishing relevancy and value of published resources.