Stuttgart, Univ., Fakultät Informatik, Studienarb., 2000.

Stuttgart, Univ., Fakultät Informatik, Diplomarb., 1999.

Zugleich: Diss. no. 12075 techn. sc. SFIT Zürich. / Im Buchh.: Aachen : Shaker. Literaturverz.

Th. sc. écon. et sociales Neuchâtel, 2004. / Existe aussi en version papier et en version électronique sur CD-ROM.

Zugl.: Hagen, FernUniversiẗat, Diss., 2005.

Universiẗat, Diss., 1999--Saarbrücken.

This thesis investigates the design objectives for a collaborative workspace based on Sun Microsystems' Java programming language and JavaBeans component architecture. The feasibility of a collaborative component workspace based on these objectives is demonstrated by Sieve, a Javabased framework for collaborative applications. Sieve allows multiple users to collaboratively add, edit, and connect components on a shared two-dimensional workspace. Sieve introduces a technique that leverages standard JavaBeans mechanisms to support use of \collaboration-unaware" software components. With this technique, components need only conform to basic JavaBeans conventions in order to be shared across collaborating sessions { they need not be programmed specifically for collaboration. Sieve also allows component developers to provide custom mechanisms for sharing components. Sieve is extensible in other ways, allowing developers to introduce new mechanisms for creating, displaying, editing, and connecting components. Three collaborative applications built on this framework are presented: a visualization environment, a circuit simulation, and a set of tools for composing arbitrary software components. The visualization environment allows construction of dataflow networks from an extensible set of modules. Modules may read data from a variety of sources, filter and transform the data in various ways, and generate visualizations. The circuit simulation allows users to collaboratively construct and analyze simple direct-current circuits. Finally, the \BeanBox Emulation" application reproduces the basic component-linking functionality of Sun's BeanBox builder tool. With this application, users may collaboratively edit and link objects that conform to standard JavaBeans conventions. / Master of Science

The aims of the research are to study the effects of computer-mediated communication (CMC) on individual communication processes and group interaction under realistic conditions. This contrasts with previous research which has been conducted using inexperienced users in artificial situations. A study investigated communication issues in an organisation where a new electronic mail (e-mail) system had been implemented. Data regarding usage patterns and subjective evaluations of e-mail showed that usability of the system was not critical, but communication and social interaction were important issues not considered during implementation. In particular, the linking of groups within the organisation had been ignored. The second and third studies investigated the way that e-mail impacts on group interaction. Research on the effects of CMC on group processes has produced a number of contradictory findings and it has been proposed that differences in the e-mail context may be responsible for these findings. Based on social identity theory and the concept of de-individuation, it was hypothesised that the identifiability of users and the strength of group identity would be important factors. It was predicted that there would be less adherence to group norms in individuated groups, in terms of: more uninhibited communication (flaming), less group cohesion and less group polarisation. Study 2 compared subjects before and after discussion, whereas study 3 focused on the dynamic nature of communication and experience, using repeated assessment. The provision of extra identifying information was associated with increases in communication activity, self-disclosure and flaming, while limiting the amount of identifying information resulted in more balanced participation. Personal identifiability did not significantly affect the way users perceived themselves, but did affect the way they perceived other group members: there was more perceived group cohesion in groups which received extra identifying information. There was no significant support for the group polarisation phenomenon. There were very few significant effects. of group identity. The research findings are discussed in relation to social psychological theory, previous CMC research and theories of group development. Methodological issues and the practical implications of varying levels of identifiability are also considered. Recommendations are made for future research. One particular issue that needs addressing concerns whether 'flaming' is properly conceptualised as normative or antinormative behaviour.

No description available.

No description available.