Demonstrator of Enhanced Publications

Introduction

This demonstrator is part of Milestone 4.1 'Demonstrator of Enhanced Publications', from Work Package 4 'Discovery', from the 'DRIVER II' project, from the 'Seventh Framework Programme'

Related tasks that were used as input for this demonstrator:

• Deliverable 4.1: Report on Publication Models
• Deliverable 4.2: Report on Object Models
• Deliverable 4.4: Selection of datasets for Enhanced Publications

The goal of this demonstrator is to trigger and help researchers and repository managers think about:

• possibilities of enhanced publications
• constraints / difficulties of enhanced publications

Requirements

The most important requirements for this demonstrator are:

• Cover multiple scientific disciplines. The enhanced publications for the demonstrator are chosen to cover typical material from the three scientific disciplines: a hebrew database from the Alpha-studies, measurements from the Beta-studies and survey data from the Gamma-studies.
• Use OAI-ORE. OAI-ORE, as suggested in the report about Object Models, and RDF are used to aggregate and serialize enhanced publications.
• Display data. Data is too often just a compressed archive of files. This demonstrator should present the data behind a publication.
• Navigation. Users can browse through all objects of the enhanced publication.
• Identification. All elements of the enhanced publications are atomic and citable web resources that are identified by a persistent identifier.
• Show interrelations. Enhanced publications connect existing resources. Seen from the other side: existing resources will connect two enhanced publication. Enhanced publications can create flocks of related research activities.

Bad experience with demonstrations has lead to the decision to :

• Independence. Create a demonstrator that is independent from webservers and/or the internet (this explainsfor example why we did not use a real resolver for persistent identifiers).

Technology

The demonstrator aggregates scientific web resources via OAI-ORE v0.9 and RDF. XSLT is used to transform these into XHTML. CSS and Javascript do the rest of the presentation. A Java applet is used to dynamically display the relations between resources. Although these relations can be fed to the applet as parameters, they are not yet automatically interpreted from the RDF-serialisation.

The demonstrator is currently built to work optimally with Firefox 3.0 integrated with a Java VM. Unfortunatly there is a bug for people using OSX and Java: the applet displaying relations between objects doesn't work.

Overview of Enhanced Publications

The demonstrator consists of the following Enhanced Publications:

• Flame: "Multiple-impinging jets: flow and heat transfer" from 3TU Datacentre / Delft University of Technology
• Generic presentation: how? Does it limit a researcher in showing what he wants?
• Annotations and Feedback, are they part of EP? Sharing thoughts, discussions about the publications, should they be part of an enhanced publication?
• Hebrew: "I and your people: syntax and dialogue in exod 33" from DANS, Eep Talstra and WIVU
• Granularity of referencing. All queries of a DB, all variables of a survey? all questions too? How persistent are they if the amount of identifiers becomes unmanagable?
• Consistency of referred resources. Databases change continuously, should we cite them? Under what conditions?
• Journalism from DANS and the University of Amsterdam
• Order. There needs to be a way to define the order of e.g. chapters.
• Physics of Fluids from Twente University
• Availability/Accessibility of referred resources. Who is to guarantee? What is allowed in an enhanced publication?
• Youth EP1 and Youth EP2 from DANS. These show how reuse of existing resources improves discovery.
• Discovery. Referencing existing resources improves the discovery of related materials.
• Imiscoe from the University of Amsterdam
• Persistent identifiers. They are needed for identification. Can't they be used for referral? OAI-ORE doesn't allow non-protocol based URIs. How many should there be?

Note: The demonstrated enhanced publications try to reflect the original publication, data and metadata as much as possible. For this demonstrator some metadata had to be added by the developers.

Other Issues:

• Versioning. An increased number of items increases the chance for updated, additions, etc. Versioning might become very complex.
• Longterm preservation of enhanced publications. How to preserve all items, who is responsible? Joint effort?
• Creation of enhanced publications: the need for tools. OAI-ORE leaves many room for solutions on how to form enhanced publications. Tools and/or agreemenets are needed ensure the required structure.
• Abstraction. What level of abstraction do we need? We can create complex hyrarchical structures according to an ideal data model, but is this what researchers want? Is there a compromising solution?

Recommendations

• Integrate different levels of hierarchy into one page
• Integrate additional material with publication itself
• General agreement on usage of structure and metadata
• Use tools to maintain relations. Don't maintain XML files