Page 78 -

IDENTIFYING THE DRIVERS OF ISD METHOD EMERGENCE 63
while a local project manager was responsible for the Danish team members. All project team
members had long formal educations in computer science, but varying degrees of practical experi-
ence in the relevant areas of SPI, ISD, and MMIS.
From the beginning, the application was envisioned as an MMIS consisting of components
representing relevant themes of SPI in the form of hypermedia-linked textual and animated slide
shows, a video film presenting a case company, a videotaped expert panel, and a large annotated
bibliography. The planned development process covered a time period of eighteen months. It was
outlined as an in-house developed method and formally included in the EU contract. The method
specified (1) a number of sequential steps, (2) milestones and deliverables, (3) development activi-
ties, (4) the application of MMIS analysis and design techniques for information content model-
ing as well as functional, interface, and technical design (such as treatment writing, storyboards,
flowcharts, and scripting languages), and (5) the use of supporting software tools for multimedia
development. The in-house developed method was inspired by a waterfall model approach to
MMIS development and multimedia development techniques recommended in available books
and papers (e.g., Bergman and Moore, 1990; Sørensen, 1997).
The actual ISD process lasted twenty-two months and only partially followed the path outlined
in the method description. Two months into the project, communication problems between the IT
consultancy and the academic team led to adaptation and standardization of information-content
templates and the accompanying storyboards. Moreover, after four months, it became apparent
that the planned sequential process could not appropriately deal with a number of complex tasks. It
proved difficult to develop storyboards for the individual information elements without specifying
the overall information architecture. In addition, attempts to define the relationships between the
different interlinked hypermedia elements were difficult. This made the related task of determining
the user’s access to the features describing the navigational possibilities equally problematic. Thus,
from month four, the overall information architecture, the hyperlink navigation structure, and user
interface modeling activities were performed in parallel. Furthermore, in month six, these activi-
ties were officially assigned to and performed by one subteam (i.e., the Danish academic team),
which was geographically located in the same place. From that period on, the other subteam (i.e.,
the IT consultancy team) largely provided input for the modeling and design tasks in the form of
literature reviews. This specialization of tasks and the resulting reallocation of resources in favor
of the academic team had not been anticipated in the original project plan. Thus, even though an
elaborate in-house–developed method that incorporated the unique requirements of the particular
situation had been outlined, it was necessary to devote much time during the project to negotiating,
revising, and revisiting the way of (co)working.

The Web Case

The Web project was performed in a UK-based market research company. It concerned the de-
velopment of a technically complex Web-based research data repository for collection, storage,
processing, and formatting of a large volume of market data about companies in the drinks industry,
such as manufacturers, packagers, and distributors. The RDR project was undertaken as in-house
development, but was performed collaboratively by the market research company and academic
staff within a government-funded program, the Teaching Company Scheme (TCS), which promotes
collaboration between industry and university. TCS specified the project duration (two years)
and the formal project organization, which comprised a steering committee (quarterly meetings),
a project team (monthly meetings), and the involvement of support from academic researchers
(weekly). The steering committee consisted of six to eight people, including the project team,