Extracting Attributes

Our working definition of compatibility is thus based on mutually agreed objectives, procedures and tools. Mutually agreed is not necessarily the same as "shared" — in some cases, a partnership will be formed to achieve complementarity, not consolidation. In order for users of the proposed online partnering tool to assess each other on relevant dimensions, the project team designed a series of structured interactions that relate to the initial planning tasks of a given partnership. The dialogue that ensues may be supported by a video-conference, by audio input, or by online text. In such a structured interaction, potential partners are invited to discuss topics that articulate focal points of trust — such as shared vision, or mutual understanding of roles. We suggest that a number of "assessment objects" can be linked to a number of "assessment activities." For example, an initial dialogue between two potential partners may discuss their "vision" of the project. In this case, the "assessment activity" is a "vision discussion" and the "objects"' that are rated might be "motivation," "investment," "realism" and so on. A further "assessment activity" might be a discussion of "technology alignment"; this would also be associated with a number of attributes for rating mutual compatibility, such as "comparable bandwidth," or "common application toolkit." A further dialogue might explore alignment of work practices by asking questions such as the following: "What systems design approach do you normally use?," "Do you use project management tools?", "Which ones?", "How do you think this will work in the current project?", "What role are you comfortable with in this project?", "Should we work in parallel or serially on this project?", "Are there areas that you think might need special care in the project?", "How do you think we might apply current standards in the design approach?"

A similar process can identify "activities" and "attributes" that are the basis of judgements about "confidence." Sets of activities, attributes and assessments can be tagged and stored in an XML database. As we indicate above, the attributes have been selected partly on the basis of advice from the key informant or practitioner partners in the project and reflect their work experience. The resulting database will support flexible searching and profiling. A project or team leader, for example, may reflect on the focal points of trust for a given project, and construct a benchmark profile for a potential partner. After interaction with, and assessment of, a number of candidates, a visualisation of their trust profiles may be compared with the ideal pattern, tradeoffs made and the most promising candidate selected. The practitioner partners in the project expect that an archive of such templates, held for an appropriate period (long or short) can support a more reflective approach to judgements about future collaborators. The coding and externalisation of what were previously fleeting and intuitive judgements should broaden the field of view, as remote exploration saves time and allows a larger number of candidates to be examined. If the templates can be viewed in a number of ways (to check a single attribute or to check a composite profile that fits the task in hand), partners may be able to consider a range of possibilities and make tradeoffs that reflect particular sets of circumstances. They can thus identify areas where a partnership may be vulnerable and assess if they have the resources to take due care. The efficacy of judgements about trust early on in the partnership process cannot be assessed until after the completion of a project. Over time, an archive of assessments will accumulate that functions as a "trust pattern library" that allows managers to consider what type of profile in what type of situation leads to what type of outcome. Such an archive, however, raises privacy issues that are discussed in the next section. Judgements about interpersonal trust made in the context of one specific project must not be held as absolute assessments.