Collaborative research poses important challenges and opportunities to scientific communities, such as the realization of effective interactions between different groups and organizations. Data sharing is growing increasingly important, as repositories are growing in size, number, and variety in the different domains. The underlying challenge is interoperability: between data, services, applications, models, and ultimately people. Abstraction and standardization (e.g. of data and services) have historically played an essential role in enabling interoperability. However, they proved not to be sufficient alone to let large-scale collaborative research thrive. For different spheres (e.g. social/economic realms), intermediation approaches were successfully applied to pursue similar goals in the so-called brokering approach. This chapter argues that brokering is useful to enable collaborative research as well, both by addressing technological interoperability and supporting socio-organizational challenges. A technological brokering framework, implemented to help the earth system science collaborative research, is finally presented, along with success stories.
TopCollaborative Research Challenges
Collaborative studies can be characterized and categorized according to the organizations involved and their goals. Organizations span from academic to private entities and collaborations can be either tight or loose. In keeping with that, taxonomies distinguish among different types of research collaborations (e.g. disciplinary, multidisciplinary, participatory, interdisciplinary) (Balsiger, 2004) (Tress, Tress, & Fry, 2006). Research groups can belong to one or more organizations and Communities of practice (CoP). They may setup one or more of these collaborative research types, depending on the scale and scope of their research.
Relations among different organizations (or CoPs) have been often viewed as a high burden with unclear benefits. The prominent motivations for general inter-organization relations to emerge are an internal need for resources or the achievement of external benefits (Van de Ven, 1976). Bos et al. (Bos et al., 2007) identify three types of barriers specific to large scientific collaborations (that are to be added to the normal challenges of working across distance):
- •
Scientific knowledge is difficult to aggregate (e.g. information semantics and data structures heterogeneity);
- •
Scientists usually work independently (e.g. large meetings, hierarchical structures, resource transfers, written contracts are perceived as barriers);
- •
Institutional related barriers (e.g. legal, economic, cultural issues).
In addition to these organizational and human/cultural interoperability challenges, in the present Web era, issues typical of complex distributed systems are to be taken into account as well; especially, when the research gets larger such as, building global and multidisciplinary e-infrastructures. Technological and semantics interoperability challenges are mainly related to the e-infrastructure:
- •
Scale: It is increasingly difficult to make a large number of systems and organizations work together. Organization patterns need to be applied in order to guide the activities of the entities involved.
- •
Heterogeneity: It can be difficult to make systems and organizations that are different work together. Mediation between parties is often a fundamental requirement. For instance, the real challenge in complex systems such as the Internet and The World Wide Web is not the number of entities involved, but their heterogeneity (Gordon & Grace, 2012). Standards and policies issued by dedicated international organizations (such as the World Wide Web Consortium) ease such interoperability processes to take place.
Due to the growing importance of (big) data sharing, great attention and effort have been dedicated to address technological and semantic interoperability, overlooking important organizational and human/cultural barriers. However, data and interoperability scientists are not expert in data content. Therefore for example, to develop cross-domain data models or semantics, organizational and cultural barriers must be addressed in addition to data structure mapping. Data sharing is another noticeable example, where the technological challenges are not as significant as policy ones. In turn, data sharing represents a fundamental building block for subsequent achievements in collaborative research requiring both technological and cultural interoperabilities, such as:
- •
Creation of inter-domain models and applications making use of the datasets;
- •
Creation and composition of workflows orchestrating heterogeneous models and datasets;
- •
Creation of web-based platforms for collaborative research.
The rest of the chapter argues about the role that some technological interoperability solutions can play to facilitate domain experts lowering human and cultural interoperability barriers.