LinkZoo: A Collaborative Resource Management Tool Based on Linked Data

1. Introduction

Collaborative management of communal resources is required in many contexts and scenarios, such as collaboration between and within professional organizations, research teams, ad-hoc social groupings that fulfil various purposes and so on. Providing a common representation model for different types of artefacts, such as files, documents, locations, websites, people and events, enables them to be managed and organized under shared principles that ultimately allow for the creation and association of intricate aggregations of different kinds of items with dynamic processes, flows and contexts.

Currently, the linked data (LD) initiative focuses on either providing representations of singular entities, or publishing large datasets in flux, created through loosely defined, context-dependent workflows. In this sense, publication of linked data becomes a complicated, non-universal process, where often custom publishing tools and systems need to be defined in order to deal with domain-dependent complexities. W3C’s general guidelines for publishing linked data, as reviewed in the Best Practices for Publishing Linked Data Note¹, are mostly concerned with good data modelling (i.e. using existing ontologies and vocabularies where possible), following good Universal Resource Identifier (URI) design practices, and providing machine access; in real world applications, most of these issues are primarily solved with ad-hoc design decisions. Except for the publishing needs, most popular semantic tools and systems focus on the storage, interlinking and presentation (e.g., browsing, visualization) aspects of whole linked datasets rather than on the user-oriented manipulation of single information resources. This creates a lack of technical support and functionality when it comes to applying linked data techniques on social and collaborative environments, where data-driven user interaction is intuitive and dynamic. Hence, existing semantic techniques and systems must be complemented by tools that target the operational layer, promoting user-oriented, dynamic interactions between resources at most granular levels.

With the re-use of ontologies, codelists and commonly used linked data URIs to annotate artefacts we can provide a layer of semantics that is inherently shared between stakeholders, and helps manage, connect, organize and explore these resources in various meaningful ways. Publication of these resources as linked data enables further exposal for external reference and processing.

For example, a scenario becoming very popular in collaborative reproducible open science activities concerns a team of researchers which collaborates on a new hypothesis. They start by collecting the literature for reviewing, which can be stored locally in the form of PDF or in web repositories. They annotate and categorize papers based on several properties such as their authors, subjects, venues and journals they were published in and enrich these properties by providing references to external web and linked data resources. Furthermore, they gather relevant dataset sources, experimental data, and information concerning the conference / journal they intend to submit their work. They share these resources with each other, show part of them to a proof-reader, and collaboratively work and manage their on-going research along all needed resources. In their individual accounts, each team member organizes the same resources differently, by creating distinct folder structures and workspaces that they find more intuitive. Finally, they make available the process, or the results of their work as linked data and they enable citation and dereferencing for all resources they managed. In this scenario, the main research and technical challenges that arise include a) how do we offer end users a uniform and extensible way of modelling and annotating heterogeneous scientific resources which can be shared between members of a research community, b) how do we offer a keyword search interface such that users can easily and intuitively search over publicly available information spaces using multiple criteria, c) how do we enable “non-tech savvy” users to collaboratively work on resources, to form ad-hoc collaborative information spaces and make them available for reuse using persistent identifiers such that they promote reproducible scientific results.