The increasing development of information and communication technology is causing fundamental changes to today’s society, changing the communication and interaction of people in their daily lives. For example, geographical distributed co-workers experience collaboration using a broad spectrum of shared workspace systems for communication and interaction despite their physical separation to share data, documents and contextual information.The objective of this thesis is to design Elgar, a shared workspace system, that flexibly coordinates teamwork for such geographically distributed teams. In particular this thesis focuses on the design and evaluation of the system for remote industrial machine diagnosis by a team of distributed engineers.This thesis identifies and describes a spectrum of coordination mechanisms, to structure and provide flexible coordination support. The two extremes of this spectrum are explored and implemented in Elgar. In the context of diagnosis, this thesis proposes Rectio, based on existing models of diagnosis to analyse observed anomalies, explore potential causes, and propose a diagnosis. Elgar supports distributed teams in diagnosis tasks providing different coordination mechanisms and various functionalities that allow engineers to analyse, discuss and document diagnosis for machines. Elgar is evaluated in two different experiments one of Hägglunds Drives and another of Volvo Construction Equipment. Both experiments evaluate the usefulness of implemented diagnosis functionalities of the system and the prescribed and ad-hoc coordination mechanisms in a collaborative diagnosis task. The first experiment evaluates the use of Elgar with engineering students, whereas the second experiment evaluates Elgar with experienced diagnosis engineers.In conclusion, this thesis shows that it is possible and necessary to design different coordination mechanisms and integrate them in a shared workspaces system, supporting flexible coordination for collaborative diagnosis tasks. In addition, the evaluation of the two experiments indicates that even the use of dissimilar coordination mechanisms provides similar teamwork outcomes, making the selection of a specific coordination mechanism a shared decision of team preferences.