DPubS

Transforming Academic Publishing

Availability of a robust and versatile open source system for electronic publishing will enable transformative change.

The DPubS Development Project

Enhancing and Extending Cornell University Library’s Digital Publishing System

Annual Report, July 2004 – June 2005

In July 2004, the Cornell University Library in partnership with the Pennsylvania State University Libraries and Press initiated a project to generalize and extend the functionality of DPubS (Digital Publishing System). This software was developed in the Cornell Library to support Project Euclid, a new initiative supporting electronic publishing for scholarly journals in theoretical and applied mathematics and statistics. The capacity of this software to support new capabilities and publishing models encouraged interest in the further enhancement of DPubS and its dissemination as an open source system enabling new opportunities for education and research publishing. With the support of a $670,000 grant from The Andrew W. Mellon Foundation, participants have made significant progress during Year One, and Year Two activities are underway. This report provides a brief summary of initial achievements and present plans.

Thomas Hickerson

Co-Principal Investigator

Cornell University Library

Sarah Thomas

Co-Principal Investigator

Cornell University Library

Project Agenda and Goals

Although the focus of this endeavor is on the design and creation of improved electronic publishing software, project goals include enabling new models for scholarly communication and innovative partnerships for academic publishing.

Principal elements of the DPubS Project Agenda are:

Cooperative Development and Testing
Technical Design and Programming
Design of a DPubS Business Plan Supporting Open Source Distribution
Promotion of DPubS and Its Application in Enabling New Publishing Initiatives

DPubS Planning Conference

In an initial effort to solicit advice and interest from university libraries and publishers, a DPubS Planning Conference was convened at Cornell University on October 18-19. Invited attendees included forty-three participants from five countries representing 16 academic libraries and university presses. http://dpubs.org/meeting.htm The meeting highlighted the new features being proposed for development in DPubS and reviewed the functionality necessary and most appropriate to each new feature. The meeting also solicited suggestions regarding features that were not originally envisioned. There was additional discussion focused on strategic planning for the rollout and testing of the open source software.

Discussions were lively and enthusiastic, as participants broadly agreed on the need for affordable publishing tools in academia, and offered constructive suggestions not only on the technical development of DPubS but on potential business models for sustaining such development. An Executive Summary was compiled by October Ivins and Judy Luther of Informed Strategies. http://dpubs.org/summary.htm

Cooperative Development: Pennsylvania State University Libraries and Press and Cornell University Library

Collaborative work thus has been both technical and strategic in nature. With the goal of publishing Pennsylvania History, Cornell worked at a technical level with metadata specialists at Penn State familiar with the publication. The Libraries in turn were in close contact with the Penn State University Press, current publisher of the journal. Pennsylvania History represented a serial publication quite different from the mathematics journals the Cornell team was familiar with, and these differences were invaluable in considering how DPubS should handled serial literature metadata. In the process of applying proposed metadata schemes to Pennsylvania History, PSU librarians critiqued these schemes and assisted in the creation of a more generalized metadata scheme for serial literature. Working with the PSU Press to design the interface, they have also alerted Cornell to unanticipated functional requirements.

Penn State has also taken the lead in developing an alternate metadata scheme for handling conference proceedings. Both partners agree that the ability to handle proceedings and other types of gray literature will be an attractive and useful capability in DPubS. All universities have such content, and much of it is now largely inaccessible.

In the coming year, the technical collaboration between Penn State and Cornell will deepen further. PSU Libraries’ Digital Library Technologies group has now begun working with the DPubS code. As they work to install and configure DPubS, they will assist in clarifying and augmenting the installation and administration documentation. They will also provide advice to Cornell staff on the code packaging. With this information, the development team expects to prepare an alpha release of the code by the end of the summer. At that time, PSU will be working with a fully functional publication system, and expects to install and maintain Pennsylvania History at PSU, soon to be followed by additional publications.

At the strategic level, PSU Libraries and Press and Cornell are actively involved in delineating methods to encourage, build, and maintain a community for DPubS open source development. PSU Libraries is already involved in the elaboration of their campus partnership with the Press and their university administration. With involvement of Cornell staff, they plan to conduct an Editors Forum at PSU, similar to an event held at Cornell in November 2003. This event brought together faculty and staff from throughout the university who served as editors and managing editors of scholarly journals. They will articulate current trends in scholarly communications and encourage faculty to consider alternative publication options. With PSU participation, Cornell plans to repeat their 2003 Forum. While the availability of robust open source publishing software is an essential enabler of change, building active campus involvement is critical to substantive change.

Center for Innovative Publishing: Expanding Partnerships at Cornell

As the Library’s Center for Innovative Publishing enhances existing partnerships and develops new ones, DPubS use is continuing to grow and diversify as the software evolves. Four new Euclid journals were added, Kodai Mathematical Journal, Tohoku Mathematical Journal, Communications in Mathematical Physics, and Pacific Journal of Mathematics. The latter two included extensive back files, as well as current issues. Indonesia , http://e-publishing.library.cornell.edu/Indonesia, a semi-annual journal published by the Cornell Southeast Asia Program newly available via DPubS, offers more than 700 articles and reviews devoted to the timely study of Indonesia 's culture, history, government, economy, and society from 1966 to the present.

The Library’s growing cooperation with the Cornell University Press will soon see the inaugural volume of a series of Comstock Press electronic editions. The first title in this highly regarded imprint will be Dukes’ Physiology of Domestic Animals, a standard text for veterinary practitioners and other professionals.

Each of its 55 chapters will be available for purchase as a fully searchable PDF and can be either downloaded or printed. Technical work completed; the Press awaiting permissions.

Other campus partnerships are in development. The first will result in the publication of the publication of the journal, Medieval Philosophy and Theology. New initiatives with the East Asia Book Series and expanded cooperation with the Southeast Asia Program are presently being defined.

Technical Development Agenda

The DPubS development agenda proposed four areas of work:

Generalization of software
Creation of administrative interfaces
Editorial management support
Interoperability with two institutional repositories

In the first year, we moved forward on design work in all four areas. The bulk of our design and programming effort was directed toward the first task, the generalization of the software. As proposed, the generalization task had two major areas of development. One was the re-engineering of the system’s User Interface Service (the module of DPubS responsible for all interaction with users). This service has now been completely re-designed and re-built, making use of XML/XSLT technologies, which allow flexible customization of all user pages. Doing this, however, required more alterations throughout the entire system that we had anticipated.

The second area of generalization work involved re-designing the way underlying publication configuration metadata and document structure metadata were handled by DPubS. To accomplish this, we have had to confront how inflexible the existing methods of the prior system were, since they were put in place to handle a single document structure with a fixed metadata type. This was true of Dienst (the precursor to DPubS), since Dienst was principally designed to deliver computer science technical reports. In building Euclid , we exchanged that document type for serial literature. But a truly flexible publishing system must handle a range of document formats and metadata types. This was one of the principal findings of the DPubS Planning Meeting. To allow for multiple metadata types and varying document structures has been a complex challenge, but one that we have adequately addressed.

At the same time, we recognized other ways in which the system needed to be modified for more general use. One example is the system’s method of service-to-service communication. Dienst assumed that the various components, or services, making up the complete system could well be geographically distributed. To ensure this possibility, communication between services was conducted via HTTP. In Euclid , however, all services were running on the same machine, and it was inefficient for them to communicate via an external protocol. We have now redesigned service-to-service communications so that the system can be configured to know which services are local and which are remote. This allows for much more efficient communications in a single, local DPubS installation, while preserving the potential for a distributed system. We also re-wrote all service communications to comply with current web services design, standardizing the service response packages. DPubS is now compliant with the REST (Representational State Transfer) protocol, and can easily be brought into SOAP (Simple Object Access Protocol) compliance, two common Web Services request/response protocols for distributed service communication.

Generalizing and updating the system for a broader range of potential environments and needs has proven to be a complex undertaking, but these improvements have widened the utility and potential of the core publishing functionality of DPubS and created a stronger platform from which to add expanded functionality.

Technical Development and System Testing, 2005/2006

We have now turned our attention to implementing the other new functionality proposed in our development agenda. In the fall, the administrative interfaces and the editorial service components will be incorporated into the core publishing system. Also in the fall, design work on interoperability with Fedora and DSpace will begin. Discussions have been initiated with representatives of both development projects (Sandy Payette/Fedora and Richard Rogers/DSpace) to outline our thinking and goals.

Plans are currently being made to distribute the software to a limited number of alpha testers. In addition to Penn State , several other institutions have expressed strong interest in similar participation. Decisions must be made soon regarding the number of alpha testers and the degree of testing support that is optimal. Based on the level of interest in early access to the software, we expect that broader distribution will enhance the growth of the DPubS development community.

With the completion of alpha testing early in 2006, we expect beta testing to begin by March. This phase will be critical in insuring that the DPubS software is ready for installation and use worldwide. It will also initiate the transition to open source development.

Managing and Promoting the Distribution and Use of DPubS

The design and implementation of an effective business model for managing and supporting the distribution and use of DPubS is critical to the success of this effort to advance the transformation of academic publishing. The planning process will address three critical elements:

Adoption of open source licensing terms and distribution processes conducive to easy acquisition and innovative implementations.
Design of an effective operational and financial model for supporting services valuable to ongoing open source development and effective use of DPubS by a broadly distributed constituency.
Delineation of a combination of hosting and management services enabling institutions and organizations to quickly and efficiently implement new models for scholarly communications.

Addressing these issues in the design of effective managerial and technical means for administering DPubS open source distribution and supporting its further development and use is a principal focus in the upcoming months. A survey of potential users of the system will identify planned uses and the open source licensing agreement most suitable to those uses. User needs for ongoing support and their interest in purchasing those services from Cornell will be explored. In delineating a framework for conducting such services, managerial and financial practices of MIT’s support for DSpace and the University of Michigan ’s management of DLXS will be examined. Additionally, practices of commercial distributors of open source software services, such as Red Hat’s support for Linux, will be reviewed. In the process of examining the interests of potential system implementers, information will also be gathered regarding the nature and extent of interest in hosted services. We will be assisted in this critical work by the experience and effort of selected consultants and by the knowledge and advice of colleagues worldwide.

Availability of a robust and versatile open source system for electronic publishing will enable transformative change.