Thursday, September 9, 2010

Defining and Exploring Networks and Circles with LORE

The amount of literature on the study of social networks is massive, having a long history that predates the analyses that are being pursued in this age of Facebook and Twitter. See the Wikipedia entry on Social Network. Students of book history and print culture might have come across studies by Charles Kadushin and others that have explored the social networks and circles of publishing in the United States of America. Books: the Culture and Commerce of Publishing (1981) draws deeply on work done in the 1960s and 1970s to identify and define the American intellectual elite. The 1974 book The American Intellectual Elite was reprinted in 2005 with a new introduction by Kadushin, demonstrating the continued relevance of his studies to an understanding of the ways in which books are written, published, sold and read. To see how widely Kadushin's research extends, see his bibliography.

The primary social networks of Australian literature have been well defined in research conducted during the last fifty years, but the validity of these networks and their
hidden complexity have not been explored. AustLit data can be used to visualise relationships between people and organisations, but the data model does not extend beyond fundamental bibliographical descriptions such as "also writes as", "influenced by", "co-author with".

Visualisation of these relationships has been achieved by drawing on such connections within AustLit, but LORE promises to extend this by enabling researchers to enhance and add relationships with additional descriptions that can better define the social structure of publishing. For instance, the structure of a publisher or periodical can be enhanced by using terms such as "employed by", "literary editor of", "ac
quaintance of", "father of", "mother of" "literary agent of", "mentor of" etc. Settling on the appropriate terms for the study of publishing is not easy and probably would never stop as unique circumstances arise that better describe certain relationships. Nevertheless, creating an "ontology" that lists all possible relationships within a certain social structure is important. The gradual accumulation of such information will ulimately produce a rich relational database that can be searched and analysed in order to better understand the social networks involved in the production and consumption of Australian literature.

As LORE is refined over the next six months, the possibilities of creating and exploring social networks will be more closely considered. Automatic mining of links between people and organisations has created thousands of compound objects that show undefined relationships between two or more people and organisations. For instance, the diagram at left visualises the many relationships that radiate from the performer, writer and producer, Nat Phillips. Such visualisations suggest the potential for collaborative refinement of these inferred relationships if individual researchers and groups can add or correct any link in this complex network.

The visualisation need not be as complex as the Nat Phillips example. Simply defining the relationship between an
author and their literary editor adds a link to the network of Australian print culture, such as this relationship between Thea Astley and Beatrice Davis. Automatically inferring relationships from AustLit data and manually creating relationships using LORE will provide a foundation for future research into the social networks of Australian literature.

To see how relationships can be created using LORE, please visit the Aus-e-Lit Project Site or AustLit. Links to updated video demonstrations will soon be added, providing more detailed information about all Aus-e-Lit tools. Until then, please visit the current demo page.

And anyone interested in helping to better define the definitions of relationships between people and organisations in literary or print culture studies .... We'd love to hear from you.

Tuesday, March 30, 2010

Compound Objects

With an increasing amount of images and full-text being digitised and delivered by local, national and international institutions, the need to be able to bookmark, relate and describe these resources is becoming more and more necessary. The Aus-e-Lit Project aims to assist researchers by developing tools that extend the specifications and protocols delivered by the Open Archives Initiative, specifically specifications relating to Object Re-use and Exchange. For researchers in the field of Australian literary and print culture studies, Aus-e-Lit tools developed from these principles will allow them to gather text, image, video and audio files in their Firefox browser window and realate them to specific fields in an AustLit Record. I'll run through a few examples to demonstrate what I mean.

If you're unfamiliar with the Aus-e-Lit Project, follow the link provided for a full description of the project's aims, a number of technical articles, and a few video demonstrations of the tools in action. As the interface is constantly changing, new videos will be produced later in 2010, but, for now, these give a good idea of the basic functions of compound objects and the associated annotation tool.

LORE (Literature Object Re-use and Exchange) is the name given to the software that has been developed as an extension to the Firefox browser. A small file can be easily installed in the Firefox "Add-ons" found in the Tools menu of the browser. Once this is done, you are free to bookmark and describe internet resources in as little detail or as much detail as you want. The screen-shot at the top of this posting is a collection of resources (AustLit records, text files and images) available on the internet. These resources have been bookmarked and the user has begun to create relationships between the reources and add metadata to describe each resource in order to gather as much information as they can for future reflection or for sharing with peers.

By adding metadata derived from Dublin Core standards to specified fields, the collection of resources and the relationships between them become more meaningful. Ultimately, users will have the ability to choose whether they want to share this information with others, but the potential for collaborative research on a large scale is very significant.

One of the most obvious examples is the creation of social networks around particular periods or cultural groups. At the moment AustLit records limited relationships between people (primarily "is influenced by"), but networks of influence could be created by relating the AustLit records of individuals associated, say, with the Bulletin in the 1890s. The variety of relationships that A.G.Stephens had with writers, publishers, artists could be clearly defined to visualise the networks of influence that determined who was published or reviewed and why. Is it just literary, or are their other factors determined by the various relationships between people. Such an activity could be done collaboratively by small, large and mass groups of contributors, building networks of influence that might begin with A. G. Stephens but merge with other editors, publishers, critics and writers. In collaboration, a better understanding of the print culture networks would emerge quickly, potentially revealing clusters of influence that have escaped notice.

Several important contributors to the development of Aus-e-Lit tools will sit on a panel at the annual conference of ASAL, 8-10 July 2010. Over the next few months their experiments will be featured here to demonstrate what is possible as far as the discovery, organisation, description and communication of related internet resources is concerned. Textual transmission, adapatation, artist-poet collaborations, and collaborative interpretation will be some of the themes covered. Please leave a comment if you'd like to join the conversation or visit us at ASAL.

Tuesday, February 23, 2010

Towards Open Access and Collaborative Research

WE use the word collaboration quite freely these days and point to collaborative research as a promising direction for humanities scholars to take. But in a field where lone scholars continue to dominate research output with single-authored articles and monographs, there is understandable resistance to growing institutional pressure to collaborate.

As those of us who work in the academy know all too well, measurement of research output is closely connected to the trajectory of our careers and so points scored are jealously guarded as tokens of value to be collected and redeemed at times of promotion or institutional movement. But as the potential of large-scale collaborative projects become more possible with the development of tools designed to enable contributions from a widely dispersed collection of scholars, this model of reward has become problematic.

The authoring tools being developed by the Aus-e-Lit Project fall right in the middle of this problematic area because of the support they give to multiple contributors from various levels of the academy, the community of independent scholars or, potentially, the general public.

The ANNOTATION TOOL will allow users to enhance AustLit records by adding information about a work or an author that would not normally be indexed by AustLit. For instance, information about the role of a literary agent in the publication of a particular edition could be added, providing a foundation for generating networks of influence in the publishing industry. If we can imagine scores of independent researchers contributing such information, cultural fields could emerge from the contributed data that tell us more about the publishing industry than we currently know.

The COMPOUND OBJECT AUTHORING TOOL allows users to define relationships between AustLit records and the growing number of disparate resources delivered via the internet by local, national and international archives and libraries. Using this tool a researcher could connect all available digital images and transcriptions of a work to their respective AustLit records and generate a digital archive to which the researcher's colleagues can contribute explanatory and textual notes that help to explain the work's textual transmission, its reception at various times, or its possible meanings in literary and cultural terms.

All of this connecting and describing is supported by a format known as RDF (Resource Description Framework) that delivers (currently) the internet resources through a Firefox browser. In this research environment, no images or text are saved, relieving the user of maintaining an archive or worrying about copyright. Ultimately, engagement with digitised archival material through the Aus-e-Lit interface will enrich the archival material by adding enhanced descriptions in stand-off markup that is stored in an Aus-e-Lit repository as RDF. Click on the Video Demo link to see the tools in action.

In the Aus-e-Lit interface all contributions can be identified by author with an associated date and time stamp. Such an environment supports peer-reviewing at a micro-level, allowing a researcher's work to be validated, quantified and assessed. At this early stage of development there is no established reviewing process, but such contributions to knowledge could provide an alternative or a supplement to the conventional article or monograph.

A fertile research environment emerges with tools that support the aggregation of disparate resources and annotations that describe these resources and the relationships between them. Such research environments have already emerged in the sciences, but an application in the data-rich environment already provided by AustLit is obvious. Models such as 'open notebook science' or 'enhanced publications' or 'mesotext' are available for AustLit researchers to consider, all of which favour open-source data with crowdsourcing contributions that make the most of publicly funded data and the publicly funded expertise that can contribute analyses.

Such models show the promise of Aus-e-Lit tools, but this promise pushes the boundaries of academic reward, publishing, copyright and the potential for collaboration in the near future. As we continue to test those boundaries, the annotation and compound-object authoring tools developed by the Aus-e-Lit Project will make a valuable contribution to the practice and theory of humanities research.

Thursday, January 28, 2010

Testing LORE with an Eye on ASAL 2010.

Aus-e-Lit development is approaching a very busy time. Members of our testing group are planning small projects that can be supported by annotation and 'compound object' authoring tools (known as LORE: Literature Object Re-Use and Exchange) that our software engineers, Anna Gerber and Andrew Hyland, are busily 'refactoring' and 'debugging'. A new version of the tool is imminent, but video demos of older versions can be seen at the Aus-e-Lit Project page.

Our Aus-e-Lit testing group will propose a panel session to the organisers of ASAL 2010, UNSW, 7-10 July 2010. The projects will include explorations of textual scholarship and scholarly editing, adaptation, literary criticism, film studies and the relationship between poets and artists in the production of limited edition artist's books. Issues such as authorship, textual integrity, aesthetics and reception will be discussed with particular attention paid to the way we might talk about these issues within the emerging field of digital humanities. The major themes and issues will emerge through testing and I will report on them here.

Over the next few weeks I'll contribute posts that look more closely at the way LORE supports individual or collaborative annotation and guide you through the world of compound objects and the potential of the semantic web for those working with Australian writers, artists, publishers and readers. I'll talk more fully about the foundation of this project in the Open Archives Initiative and speculate where a project such as this might lead in the years to come. Discussion might also lead to proposals for new names for the thing that is currently known as a compound object! Your recommendations on this will be enthusiastically received, but 'Fred' will not be accepted.