Fedora Users Conference
June 19-20, 2006
University of Virginia Library
Charlottesville, Virginia
Abstracts
Monday, June 19
Session 1A:
Ryan Scherle and Jon Dunn, Indiana University
Title: A Fedora Architecture to Support Diverse Collections
Indiana University is in the process of migrating its digital collections from a mixture of ad-hoc systems and "digital library" systems into a single architecture based on Fedora. A primary challenge of this migration is converting collections with divergent features and implementations to fit into the common architecture without losing functionality. This talk will describe the Fedora content models we have designed to support the conversion, including problems and solutions in implementing these content models. We will describe two tools that interact with these content models, an ingest tool and the METS Navigator page-turning application. The ingest tool, which is still under development, expands on the features of existing ingest tools to support ingest of diverse collections. METS Navigator, which has recently been released to the public, can easily be combined with a Fedora system to provide a delivery interface for paged documents.
Rob Chavez, Tufts University
Title: An approach to modeling rich content and disseminators: promoting interoperability and reuse of content in the Tufts digital repository
This paper will highlight several collections in the Tufts Digital Repository (TDR) to demonstrate our approach to creating rich Fedora disseminators and content models. The collections, Tufts Oral History collection, Perseus Digital Library American History and Art and Archaeology collections, and the Tufts Fine Arts Department image catalog, illustrate a variety of digital object types to which we have applied our content modeling and dissemination principles to facilitate dissemination and access of digital repository content to multiple services and applications.
Our approach to content modeling is decomposing the disseminators into base groups or atomic units (behaviors). This allows content model designers to piece together unique content models from standard components. By utilizing common behaviors across content models, digital repositories can achieve high level of interoperability and predictability in their objects and to contribute to inter-repository reuse and sharing of objects.
The disseminators provided in the TDR are predominantly functional in nature. They perform carefully selected transformations over the object’s datastreams resulting in standard outputs such as XML, text, image, or aggregates. Rendering the results of the disseminators is solely the responsibility of the applications implemented over the repository. In essence, we provide a useful middle-ware interface for our digital objects, which could be used for a wide variety of applications.
We will illustrate these principles and demonstrate a variety of dissemination types available in the TDR including dissemination of streaming and static audio, text transcripts, RSS feeds, aggregation/collection browsing using the Fedora RDF relationship schema and triplestore. We will demonstrate a working audio collection through the Tufts Digital Library application, and a variety of working image (page and illustration) collections through the Perseus Digital Library, the Tufts Artifact web-based curriculum tool, and the Tufts Visual Understanding Environment.
Session 1B:
Carl Grant, VTLS, Inc.
TITLE: A VITAL business?
Commercial backing of open-source products is a challenging business model both for the vendor and the buyers. This paper looks at the VTLS history and business of supporting Fedora and developing open source and proprietary extensions to the software. It will discuss the challenges encountered ranging from buyer expectations, pricing issues, partnerships, vendor positioning within the open-source community and trying to find a sustainable business model when backing an open source product that is technically superior but that represents a subset of the total solution needed by a marketplace that is more accustomed to buying complete packaged solutions. Comparisons will be drawn with other companies that have provided backing of open source software. The business models of these companies will be examined to look at the paths they've chosen and how those decisions might be used to predict the future of this business model in this marketplace. The conclusions will also examine the how these offerings need to be marketed, priced and positioned now and into the future. Finally, customer perspectives will be discussed and final conclusions drawn about the overall viability and future of this kind of business model.
Jeffrey Barnett, Yale University
TITLE: Evaluating the role of vendor support for Open Source repository components in a Research Library Environment.
Authors: Jeffrey Barnett, Gretchen Gano, and Dave Gewirtz
In 2005, the Yale University Library conducted a research study that explored the use of the Fedora service framework to create collections to support digital teaching and research. Through structured use cases, scenarios were constructed and tested to determine how and where Fedora, and vendor functions built on top of it could meet the, often idiosyncratic, collection requirements of an academic community. As a result of that investigation, this paper summarizes the experience, with particular detail and emphasis on gaps and complementary roles of the framework, the library, and the vendor.
SUMMARY: At release 2.1, Fedora could be said to be entering late adolescence as a software product. While the Fedora service framework approaches maturity satellite services (advanced workflows, preservation, storage management and cross repository communication) are nascent but have future promise to significantly extend the value of the system to support enterprise digital scholarship and communication. To gain experience with the Fedora architecture, while retaining the advice and assistance of a professional staff, the library elected to collaborate with VTLS in an examination of the VITAL package, built on Fedora in meeting some immediate needs and as a foundation for emerging long term goals. As with any software package on the cusp of maturity Yale found both encouraging early successes, and some surprising limitations. The environment of mixed open and vendor development based on open standards promises to be well suited to exploiting the advantages of both.
Yale University Library views the collaborative process used to evaluate VITAL as a valuable model for learning new technology, and plans more research projects based upon this assessment model.
Session 2A:
Dr Andrew Treloar, Information Technology Services, Monash University
Title: Dataset Acquisition, Accessibility, and Annotation e-Research Technologies Project: Fedora and the new collaborative e-research infrastructure
This paper provides an overview of the DART project. It commences by describing the context within which DART (and similar projects) were funded by the Australian Commonwealth Govermment. It then discusses the underlying theoretical basis of DART before proceeding to describe the areas of activity in which DART is engaged. The paper describes the various work packages in the DART project, including the core role played by Fedora-based repositories. It concludes by reviewing the progress so far of the project and its anticipated outcomes.
Matthias Razum, FIZ Karlsruhe
Title: The eSciDoc Project - an Overview
eSciDoc is as a joint project of the Max Planck Society and FIZ Karlsruhe, funded by the Federal Ministry of Education and Research (BMBF), with the aim to realize a platform for communication and publication in scientific research organizations. The result of the entire eSciDoc project is intended to:
- Ensure permanent access to the research results and research materials of the Max-Planck Society and seamless integration within eSciDoc as well as integration into an emerging, global, electronic knowledge space.
- Provide effective opportunities for access to information for scientists of the Max-Planck Society and their work groups.
- Support scientific collaboration in future eScience scenarios.
Especially the third goal requires a shift from traditional digital library systems to a more interactive environment in which information consumers become as well information producers. Repositories have to open up and move towards a federated world of distributed systems.
eSciDoc consists of applications built on top of a shared framework. Within the scope of the project, four applications (Publication Management, Scholarly Workbench, eLib, and e-Laboratory Journal) will be implemented, highlighting the potentials of the overall concept. Other applications may be added later. The framework offers basic operations like object storage, security, ingestion and export of objects, search, and workflow management. It is designed as a Service Oriented Architecture to allow for easy reuse in other applications within and outside of the Max Planck Society.
Session 2B:
Carol Minton Morris, National Science Digital Library
Title: Publishing in the NSDL: Fundamental Concepts for Creating and Reusing Content
The National Science Digital Library (NSDL) dramatically broadens the information about STEM resources that it can accept and make available to its users with the introduction of the NSDL Data Repository (NDR) architecture. [1]
In this paper we present concepts for On Ramp (ONR), a content and communications system that supports handling of multiple types of content collected together as a single entity; that facilitates reusability and distribution of content to multiple locations; and that leverages existing and emerging technologies to enable far-flung networks of content contributors to create and widely share context and insights about Library resources that can transform information into warranted knowledge. [2]
David Ruddy, Center for Innovative Publishing, Cornell University Library
Title: Using DPubS to Publish Fedora Content
In July 2004, Cornell University Library, in partnership with the Pennsylvania State University Libraries and the Pennsylvania State University Press, received support from The Andrew W. Mellon Foundation to generalize and enhance electronic publishing software originally developed at Cornell. The resulting application, called DPubS (Digital Publishing System--http://dpubs.org), will be released under an Open Source license in mid-2006. One goal of the current development project has been to facilitate interoperability with institutional repositories such as Fedora and DSpace, allowing DPubS to operate as a publishing application layer on top of existing repository service software.
DPubS is a full-featured, extensible publishing application designed on an open services model. The software enables publishers to organize, present, and deliver both open access and subscription controlled scholarly communications. Responding to publishers' needs for content branding and identity creation, DPubS has implemented a flexible XSLT front-end architecture. This allows a service provider to support a diverse range of publications, each with their own look-and-feel, while maintaining content within a single instance of DPubS. Application wide services include full-text searching, OAI compliance, flexible access controls, e-commerce capabilities, etc.
This presentation will introduce DPubS, the methods used to connect to Fedora, and opportunities and challenges for further interaction between the two systems.
Session 3A:
Richard Green, University of Hull
Title: The RepoMMan Project: automating metadata and workflow for Fedora
The RepoMMan project at the University of Hull is funded under the JISC Digital Repositories Programme and is closely aligned with the University commitment to deploy an institutional repository.
The project has two main technical objectives.
The first is to develop a standards-based, flexible, workflow tool through which users can interact with Fedora. This will be surfaced within the institutional portal and, as well as the basic repository functions, we intend that it should provide assistance both with versioning and automatic population of metadata. The project will carry out further work to surface the tool within the Sakai C&LE.
This second objective, the automatic population of metadata, is seen as a major problem in many repository projects. Because the workflow tool will be surfaced within the University portal, we expect to be able to support relatively sophisticated population of contextual metadata based on user profile and current role. This process will be informed by a number of use case studies. We also intend to investigate ways of extracting metadata from the repository objects themselves. The range of metadata will be presented to the user through the workflow tool so that it can be edited should they wish.
Gert Schmeltz Pedersen, Technical University of Denmark
Title: Development of the Fedora Generic Search Service
The Fedora Generic Search Service is an ongoing development aimed at inclusion in the Service Framework in version 2.2 of Fedora. It contains a generic interface that allows plugin of your favorite search engine. Initially, plugin modules for Lucene and Zebra are developed.
The generic interface has operations for index updating, for searching, for index browsing, and for getting engine specific information. The set of index fields is tailorable, and they may have contents both from the FOXML records with their inline XML datastreams and from the datastreams, whether managed, referenced or external. The query language for searching is determined by the search engine, and correct queries are generated either by the engine specific plugin module or by the user interface.
Xslt stylesheets for index document creation and for result page generation, extensive use of configuration properties, and mimetype specific transformers to text all contribute to the flexibility of the service.A number of concerns are considered, including how to deal with security, how to deal with different document types, and how to deal with multilingual content.
Prototypes have been available at http://defxws2006.cvt.dk/fedoragsearch since February 2006, and members of the Fedora community have contributed to the development in various ways.
Session 3B:
Dean B. Krafft, National Science Digital Library
Title: Building a National Science Digital Library on Fedora
For the past year, the National Science Digital Library (NSDL) project at Cornell University has been implementing a new version of the library in Fedora, creating the NSDL Data Repository (NDR). In the first half of the talk, I will present some background on the overall project, and then give our results and experiences in taking an existing metadata repository database of over a million science resources and porting it to a new data model in Fedora. The model makes significant use of explicit relationships, which has resulted in some interesting scaling challenges, particularly for the Kowari-based triplestore. In the second half of the talk, I will discuss work now underway to integrate the NDR with blogs and wikis. In addition to creating new science resources, these collaborative tools are used to create direct and implied context for NSDL resources, building on Fedora's ability to flexibly manage and query content, metadata and relationships. I will close by looking at some of the challenges in using context to aid in discovering and selecting resources from the library.
Daniel Davis, Harris Corp.
Title: Enabling the Integrated Information Network with Fedora
The majority of the information produced or utilized by software applications exists as unstructured data on networks and file systems. Programmers constantly have to reinvent methods to capture, store, manage, preserve, and deliver this information especially where it outlives the application. Web technologies, Web services, and Service Oriented Architectures using Fedora as a component can provide an evolvable infrastructure for handling and exploiting this information and are key elements for enabling incremental integration of new applications in implementing large information systems. Fedora may be incorporated as the repository for enterprise content, document, and record management systems enabling a uniquely flexible approach to handling unstructured data in highly networked systems and hinting at the future of the World Wide Web. This presentation will examine architectural elements of a Fedora-centric integrated information network.
Tuesday, June 20
Session 1A:
Beth Kirschner, University of Michigan
Title: Sakai Fedora Repository Tool
Many research projects require both the collaboration tools of a CLE (Collaboration Learning Environment), such as Sakai, and a powerful repository framework such as Fedora. In a well-integrated solution, the data and metadata in the repository must be accessible to the Sakai tools, and the Sakai user attributes must be accessible to the Fedora authorization component. Many existing repositories suffer from inflexible designs, making it difficult to modify or extend the underlying data model. A Fedora-based repository interface, integrated with Sakai authorization, would be capable of supporting a wide variety of data models. Different data models can be configured using XML Schema definitions, and presented using XSL Template files. Fedora can support a strictly hierarchical data model, or more complex data models incorporating a Kowari triple-store ontology. Data stored within the repository can be transformed using custom dissemination methods as needed by Sakai collaboration tools. This paper describes the current Sakai Fedora Tool Prototype, as well as future development work.
Gert Schmeltz Pedersen, Technical University of Denmark
Title: Considerations about a Peer-to-Peer Service for Fedora
The Technical University of Denmark is a partner in the EU project "ALVIS - Superpeer Semantic Search Engine". Superpeers are nodes in the ALVIS peer-to-peer network with a protocol that allows them to exchange information about their repository contents and to distribute queries and assemble hit lists from responding other superpeers.
Currently, it is under consideration, how Fedora could be wrapped up as a superpeer. Seen from the standpoint of the Fedora Service Framework, it would look like adding a Peer-to-Peer Service. This would facilitate cooperation not only among Fedora repositories, but also between a Fedora repository and a large number of superpeers that are based on other types of repository or database systems, if these are wrapped up as well. The difference from simple web service-based cooperation between two systems is that the peer-to-peer network maintains a global index and directs queries to relevant superpeers.
The talk will explain and illustrate functionalities and possibilities of such a service.
Session 1B:
Chris Awre, University of Hull
Title: An institutional repository for the University of Hull: supporting user needs
The University of Hull has a somewhat unconventional view of the services that an institutional repository might provide. Far from being just a collecting point for finished work, the repository is seen as supporting staff (and potentially students) in the development of materials from idea, through development, to publication.
The University has adopted Fedora as the underlying repository software because of its flexibility and scalability. The University's repository will eventually support thousands of individual users storing myriad file types, some as works-in-progress, some as finished, published, objects. These files need to be accessed, stored and managed simply and in the knowledge that they are safe and secure - now and in the future.
This presentation will describe how the security and content-model features of Fedora are being harnessed to provide such a wide-ranging implementation.
Dr Andrew Treloar, Information Technology Services, Monash University
Title: The ARROW project: a Fedora-based Institutional Repository
The ARROW consortium consists of Monash University as the lead institution, Swinburne University of Technology, the University of New South Wales and the National Library of Australia. ARROW decided to have the main modules built on top of Fedora (http://www.fedora.info/) and written by VTLS (http://www.vtls.com). ARROW licensed VITAL and has been working with VTLS to extend the functionality of Fedora by commissioning a series of Open-Source Web Services. This has been a true partnership, with significant transfers of intellectual property in both directions.
We propose to offer an overview of the progress of the ARROW project, which has been developed within the following three stage framework:
- conceptualising and developing an ARROW institutional repository solution comprising software, policy frameworks and implementation strategies, particularly related to the effect of using Fedora as an underlying architecture
- implementing the ARROW repository within the project's partner institutions, and
- offering the ARROW repository solution to other Australian universities.
We will discuss the decisions the ARROW project made at the outset and provide a review of those decisions after two years of operation. We will also look at the way that the use of Fedora has impacted on the way that the repository works.
Session 2A:
Christian Tønsberg, Technical University of Denmark
Title: IRIaB - Institutional Repository In a Box
The Digital Repositories at The Technical University of Denmark (DTU) is gearing up to meet the demands of the foreseeable future (and present), and in this process, depositing digital objects (fulltexts, primary datasets, software, audio/visual materieal, etc) will play a more and more prominent role.
Manual dataproduction to those repositories is done through a taylormade
web-based input management system (MetaToo), also used in related contexts on a national level.Joining MetaToo and Fedora in our repository solutions is deemed a critical milestone for factoring Fedora into these solutions, which is the strategic goal at DTU.
The primary requirements on this join are:
- seamless cooperation between MetaToo and Fedora wrt. ensuring a desirable workflow for data treatment (inputing, proofreading, releasing, depositing, displaying, correcting, republishing, preserving, etc)
- easy installation and configuration of Fedora-enriched repository solutions in the DTU operational environment, in order to pave the way for rapid development and robust production, both of which are important to DTU.
Thus, IRIaB aims to become a starting point for a generic academic institution wishing to experiment with Fedora equiped with an input management system designed specifically for dataproduction for Institutional Repositories.
Christiaan Kortekaas, The University of Queensland
Title: Don't keep it under your hat
In response to demand for a robust and scalable system to host and manage access to a range of electronic content, including theses, book chapters, articles and other research output and teaching materials, the University of Queensland has developed "Fez" - a new, open source, Fedora-based digital repository management and workflow system. Fez is based on PHP and MySQL and works as a front-end and administration and content mangement tool using Fedora 2.1.1 (http://www.fedora.info/). It is a highly flexible and configurable system for repositories.
In 2006, the University of Queensland is undertaking a Research Assessment Exercise (RAE) in which academic staff from different schools and centres must nominate the three research works from the last five years that they consider their best. These works will then be assessed by different international panels of experts in each of the research fields. This case study explores the development and architecture of a Fez based repositories and looks at ways in which such a system could support the Research Assessment Exercise.
Specific content models were developed to meet the reporting requirements of the RAE. These included a 'citation view', with links to either the object's Digital Object Identifier or to a locally housed version of the file to grant reviewers full access to materials.
To simplify content creation Fez was integrated with data feeds from UQ central human resources systems to provide rich form controls based on cutting-edge AJAX technology. For example an 'Author Suggest' control which acts like Google Suggest.
Fez is also being prepared for federated authentication and authorization based on the international standard eduPerson attributes, implemented with Shibboleth technology. This will assist future RAE processes by allowing external reviewers to authenticate using their own institutions central identify provider and Fez will be able to base access control rules for these external reviewers based on their individual eduPerson attributes.
Session 2B:
Ron Jantz, Rutgers University
Title: Report from the Fedora Preservation Services Working Group
The working group is specifying generic capabilities designed to address some of the critical components of a trusted repository architecture. In our meetings, we have worked to clarify philosophy and key concepts including the meaning of terms such as curation, preserved, managed, and archived. In the specification process, we are focused on the underlying capabilities to support digital object persistence, life cycle management, multidisciplinary collections, and management of the repository environment (e.g. storage, memory, operating system, etc). Our work is informed by the RLG/NARA report on certification of trusted repositories (http://www.rlg.org/en/pdfs/rlgnara-repositorieschecklist.pdf), the Tufts/Yale grant project from the National Historical Publications and Records Commission (NHPRC), the Harris/NARA Electronic Records project, Rutgers University Libraries repository development, and the Fedora Development Group. This report will cover the concept architecture, core features, an early development plan, and the next steps for the working group. The working group is also hoping to get input from conference attendees on our general direction and the relative importance of the core features.
Members of the Working Group: Grace Agnew - Rutgers, Dan Davis - Harris Corp., Kevin Glick - Yale, Ron Jantz (chair) - Rutgers , Karen Miller - Northwestern, Sandy Payette - Cornell, Eliot Wilczek - Tufts.
Eric Jansson, National Institute for Technology and
Liberal Education (NITLE) and Stacy
Pennington, Rhodes College
Title: Years of Hope, Days of Rage: Integrating Fedora into the Small College
This presentation will focus on our ongoing work on strategies and tool development for integrating the Fedora Repository System into the small liberal arts college environment. We will describe the unique perspective of this type of institution on a technology like Fedora, which promises to unite diverse functions of campus digital archive and provision efforts under a single structure. This promise has been difficult to realize, however, as tool development and technology dissemination continue to leave gaps in the "stack" of technologies needed by these institutions to commit and integrate Fedora.
We will investigate many of the strategies employed to allow a small institution to take advantage of Fedora with a limited budget and even less staff to dedicate to the process. We will look at ways the Fedora Project could specifically help smaller schools to begin to use Fedora for digital archiving, and we will also explore methods for smaller schools to give back to the project.
Looking specifically at the experiences of Rhodes College and its "Crossroads to Freedom" project's use of Fedora, we will examine how a merged information technology and library team is combining its skills, focusing primarily on content but simultaneously developing metadata standards that not only provide classification but can be used to drive the user interface of the digital repository. We will show how the eventual consolidation of digital objects into Fedora and standardization on Dublin Core is driving application selection throughout the campus.
We will also describe our open-source tool development efforts - including Elated and Manu (a TEI manuscript publication system) - and detail the lessons learned from these experiences and the work planned for the future as we try to fill in the technology pieces necessary to make Fedora succeed in small colleges.