These documents are work in progress and are therefore not available directly. Much of the work has been adapted for inclusion in the deliverables. Requests for copies of any of these papers should be directed to Jacqui Forsyth in the first instance.
C. Simone, B. Hewitt, K. Schmidt and A. Pozzoli
Risø-3-10
In order to handle the high degree complexity that characterizes cooperative work in modern work settings, and handle it efficiently, the articulation of the distributed activities requires a category of symbolic artifacts that stipulate and mediate articulation work and thereby reduce the complexity of articulation work. We call these artifacts `mechanisms of interaction'. From the evidence of sociological field studies, computational mechanisms of interaction must be malleable and linkable. The paper describes a three level architecture for a notation that enables cooperating actors to manipulate computational mechanisms of interaction.
L. Navarro
UPC-3-5
The Shared Object Services and the Shared Interface Services (SOS/SIS: S*S) is an architecture to computer support the complex interactions of multiple actors working in a complex work arrangement. The S*S has grown as a result of the work in the COMIC Strand4 Sharing Objects group (C4SO) from a set of prototype systems to be an architectural model of the components, primitives, dependencies, flows of information among the components of such large scale computer system. In other words, it provides a notation(s) at the semantic level of cooperative work to design and incorporate different mechanisms of interaction built on a common set of primitives, a common notation in order to avoid artificial boundaries and support articulation work [Milano-3-3].
An overview of the architecture with a focus on the Resource Manager, a component that provides primitives to manipulate the objects of articulation work and reduce the complexity of articulation work by providing orderly access to resources in their organisational context.
The Aleph-Tcl is a prototype notation with the required malleability and linkability to specify mechanisms of interaction based on the C4SO architecture.
P. H. Carstensen, T. Tuikka and C. Sørensen
RISØ/OULU-3-11
In large software development projects, software testing is a complex process involving a multitude of cooperating actors with different perspectives and competences. Mutual interdependencies between participants in the process requires articulation of the work, e.g., meshing, allocating, coordinating, negotiating, etc. of activities. The nature of the organizational aspects of software testing, hence, lead to a number of requirements for computer-based mechanisms supporting cooperative aspects of software testing. Based on action research of software testing in a large international software company and a field study of the software development process in a large instrument development project at a manufacturing company, a characterization of the organizational complexity and the cooperative nature of software testing is presented. These characteristics provides the basis for an initial set of requirements for computer-support of the cooperative aspects of software testing.
C. Sørensen
RISØ-3-17
Traditionally, the Bill of Material is used to manage the production of components and units. The Bill of Material specifies the number and types components which goes into making sub-assemblies, and the type and number of sub-assemblies that goes into making units or finished products. Hence, the Bill of Material is basically a tree-structure making it easier to cope with the complexity of a multitude of interconnected parts. Material Resource Planning (MRP) production scheduling systems are based on Bill of Materials. At Foss Electric they had over the years used an ordinary Bill of Materials in order to manage the production of prototypes and test series of instruments. Because Foss Electric has implemented a concurrent engineering mode of operation, developent projects test the fundamental ideas, the detailed designs, and the production and assembly system in a series of prototypes. There is, hence, a substantial element of production planning and control even in the development phase. As the instruments got more and more complex, this work became more and more complex. In order to cope with the complexity, the participants engaged in weekly meetings. In these meetings the participants reviewed Bills of Materials and blue-prints in order to obtain a common understanding of the state-of-affairs and in order to make decisions regarding changes to the various specifications. In order to reduce the need for this weekly meeting, the Augmented Bill of Materials was designed. Apart from containing all the information on the regular Bill of Materials, it also includes a number of symbolic references to objects of articulation work and to external systems of reference. A written organizational procedure stipulates the process of reviewing, annotating, updating, and sending the Augmented Bill of Materials and blue-prints.
P. Carstensen
RISØ-3-15
In software development the testing of the software is often a demanding and difficult task involving many actors that need to coordinate their activities, distribute and be aware of information on identified errors, negotiate on the classification of an error, etc. To support the distributed nature of the testing activities different forms, schemes, etc. combined with explicit procedures for how to handle the errors are often used.
This paper report in a very descriptive manner from a field study conducted at Foss Electric. The field study especially addressed how the software designers in a large project articulated their activities. An interesting finding was that the software designers involved in the project realized problems in monitoring, coordinating, and handling the testing activities. To cope with the increasing complexity in monitoring, coordinating, and handling the testing activities they invented and used a standardized form (the Bug Form) that all testers had to fill in whenever they identified an error (bug) plus a set of procedures and conventions for how to handle the process of registering, diagnose, and correcting the bugs.
In this paper we consider, analyze, and describe the Bug Form, and the procedures and conventions as a social mechanism of interaction ( Schmidt, 1993 ). The purpose of this is first of all to get a good understanding for how the articulation activities, required in the software testing work, are handled and how they are supported and stipulated by the mechanism. A second purpose is to establish an empirically based foundation for evaluating and improving the concept of Mechanisms of Interactions ( Schmidt et al., 1993 ).
C. Sørensen
RISØ-3-18
Developing an instrument at Foss Electric is a very complex process. In most instruments mechanics, electronics, software, and chemics interacts. Before an instrument is ready to be
produced and sold, an abundance of detailed decisions has to be made. Each decision made by a participant can potentially affect decisions to be made by others. It is therefore important to ensure both participants awareness of state-of-affairs as well as a fully distributed process of finding errors and promoting changes. In order to support this process, Foss Electric has adopted a Japanese system for reducing the number of defects in manufacturing by enabling distributed accumulation of knowledge and experience. The Cause and Effect Diagram with the Addition of Cards (CEDAC) system basically consists of a large board with categories for classifying observation cards. This CEDAC board is updated by participants and regular meetings are held where the observations posted are discussed. An observation might result in a corresponding suggestion. This suggestion for change is classified according to it's current status, e.g. whether the suggestion is being tested, or whether the test performed has reached a successful result. Because the CEDAC system contains two classification structures, it reduces the complexity of distributed classification. The regular meetings, therefore, can be conducted in a more effective manner.
K. Schmidt
RISØ-3-22
This paper discusses the requirements for a general notation for computational mechanisms of interaction.
The Construction Note
H. Andersen
RISØ-3-16
This paper reports in a descriptive manner on how a form is used in a large scale manufacturing company for the articulation and propagation of changes regarding products, parts, processes and informational objects The Construction Note is a mechanism used within the company to handle and distribute semi-structured messages and notes regarding product changes and to handle and distribute proposals for product changes. Also it is used as a basis for delegation of responsibilities and tasks, and to a certain degree control of stock, processes, machining and measuring tools.
K. Schmidt and C. Simone
RISØ-3-20
This short paper presents and discusses a new definition of `mechanisms of interaction' that solves the problems encountered in applying the initial definition in field studies of the use of symbolic artifacts for coordination purposes.
S. Herskind and H. Nielsen
RISØ-3-13
The present report documents an experiment conducted with the purpose of designing a mechanism of interaction.
The report presents a short review of the theory outlined in the COMIC Deliverable 3.1, with a focus in the notation concept. The notation concept is used as the design tool for a mechanism of interaction for emergency management.
The actual content of four objects of articulation work are designed at the notation levels, and a specification of what happens the levels is given. Further the notation design is transformed to a semistructured message system.
In the conclusion the experiences by performing the design process are given, and the design is evaluated in a further CSCW perspective.
P. Carstensen and C. Sørensen
RISØ-3-21
When organizations embark on manufacturing complex products, a multitude of actors representing different areas of competence cooperate. Because of the complexity of the work due to, for example, the nature of the product and the large number of interdependent participants, part of their work concerns articulation, e.g., coordination, management, allocation, negotiation. This chapter provides a general introduction to the field study at Foss Electric, a Danish manufacturing company. The field study surveyed a large-scale project involving mechanical, electronic, software, and chemical design of a complex instrument for testing the quality of raw milk. We argue that in this particular project a matrix organization, scheduled project meetings, informal meetings, and paper-based artifacts are the primary means of managing the complexity of articulation work. This chapter serves as a general introduction to the following four chapters analyzing the social mechanisms of interaction used at Foss Electric in the S4000 project: 1) The Bug Report; 2) the Augmented Bill of Materials; 3) the CEDAC Board; and 4) the Product Classification Scheme.
J. Pycock
MAN-3-3
This paper re-examines the case study described in MAN-3-2 but in terms of concepts drawn from a literature on Technologies of Representation. In exploring the features, functions and interelation of representations in this case and using concepts from this literature, the paper provides a number of inputs to the developing notion of Mechanism of Interaction.
W. Sharrock and J. Pycock
MAN-3-2
This report consists of three main sections. The first discusses the ethnographic approach we have taken to analysing and describing work, courses of action, their organisational context, their relationship to plans and projections and, finally, to those mechanisms of interaction evident in work situations.
The second section details those representations and mechanisms of interaction (eg. 'Fault Report Forms') found to be at play in the organisation of a design and development project we studied. Our report also describes some of the working practices associated with these mechanisms of interaction.
The third section presents a discussion of aspects of the fieldstudy using notions of the 'social distribution of knowledge' taken from the work of Schutz. In this section we address, among other topics, how different groups within the project come together to interpret or 'read' the Fault Report Forms and the information about the field of work that they represent. We also discuss some how some of the practices of the 'writers' and the 'readers' of these forms are related.
B. Krogh
RISØ-3-14
Is the object oriented paradigm suitable for eliciting and expressing requirements for computer systems to support cooperative aspects of a work domain? In this article, a project is outlined in which object oriented analysis and design was used to develop and evaluate a computer based group calendar. A perspective on CSCW is then presented, based on the concepts of articulation work and Mechanism of Interaction (MoI). The group calendar case is reviewed according to a set of requirements for MoIs, introduced by this perspective. The relation between the support of articulation work that this type of computer application must provide and the object orientation of the development approach is discussed briefly. It is concluded that important aspects of MoIs are not handled sufficiently well by the object oriented approach. As an outline of further research activities, it is suggested that a combination of the object oriented paradigm and the MoI perspective is one effective way of further developing powerful approaches to the design of computer support for cooperative work.
M. Dam
SICS-3-7
A critical dimension in the handling of change in computer-based systems for cooperative work is whether mechanisms for change should be explicitly embedded into systems, or whether change should be handled in a global and uniform manner, for instance by a process of editing and recompiling programs or scripts on the fly. We argue that to reflect the structure of organisations, powers of change must be local, structured, and dynamic. Thus a global and uniform handling of change is in general insufficient. We propose a formal basis for the description of dynamically modifiable objects, and explore its applicability in the field of CSCW by exposing it to three examples of increasing complexity: A system for dynamic communication channel creation; an adaptable conversation manager; and a rudimentary, yet quite general, awareness model.
C. Sørensen
RISØ-3-19
The product classification scheme developed and used at Foss Electric supports the articulation of distributed storage and retrieval of CAD models in and across development projects. The classification scheme has 16 classes and a total of approximately 340 categories. The classification structure is build mainly based on the functional properties of the components and units in the field of work, e.g., springs/dampers, optics, and transmission components. There are, however, also entries in the structure which are defined based on the structural properties of objects in the field of work, e.g., console, plate, and rod. It is used as a mixture of a paper and computer-based social mechanism of interaction. Use of the classification scheme is not stipulated in any written procedure, the distributed activities are made according to conventions. It is, however, argued that the classification scheme itself in the form of a tree structure stipulates the articulation of distributed storage and retrieval of CAD models. The mere existence of a category stipulates that any component or unit with this functionality must be placed in the category.
T. Tuikka
OULU-3-2
There are many papers pointing out ways to reach reliability in the product. Those papers are mostly dealing with the issue in quantitative manner. However, the reliability we in our everyday software production try to reach depends more or less on an agreement between people producing a product. This agreement is reached through cooperation during the process of software production. This is a paper to take a look at software testing and reliability with a view of cooperative work. Surely, all work is essentially cooperative work in that it depends upon others for its successful performance and software testing is no exception of this. Open question is what is the role of cooperative work in software testing and reliability of the product? The ideas presented here are examples based on the experience of the writer in the field of software testing. As a result cooperative work is seen to play essential role for reliability. Furthermore, a view of articulation work is suggested to be useful for further studies of cooperative work in software testing.
T. Tuikka, J. Kokkoniemi, T. Lalli and K. Kuutti
OULU-3-1
This report describes a tool which can be used to demonstrate cooperative systems. We have named it as MOI-tool. The tool has been built for Apple Macintosh computers in network using HyperCard. The tool has two purposes: research and practical purpose. The research purpose is connected with the primitive operations and their relationship with the notations, describing mechanisms of interaction used in cooperative work. The practical purpose is to support quick and easy creation of demonstrations of cooperative interfaces in multiple workstation. Thus, it should be easy to create demonstrations for example of calendar, flow control or state control applications between many workstations. The demonstrations should work in a way sufficient enough to support discussions with possible users of similar systems, furthermore users should better understand how proper application could help them in their work.
The report is divided into two main sections. First section explains the idea of how tool has been designed, what kind of ideas we have had and what are the conclusions where we are at the moment. Second section (third chapter) is a manual and an example of how a user can build a demonstration.
J. Pycock and W. Sharrock
MAN-3-5
We here make two proposals about the way in which the notion of 'mechanism' of interaction might be elaborated, ones which will make more apparent its consistency with themes and concepts being developed in other Strands of the COMIC project. We first argue that it is necessary to maintain the 'anti-essentialist' frame of mind which informs much of COMIC work, and thus avoid fruitless debates about whether some particular artifact is, or is not, intrinsically a Mechanism of Interaction. Secondly, we suggest some ways in which the study of 'Mechanisms of Interaction' may be informed by the employment of the 'performative' approach to organisation as developed by Bowers in Strand 1.
H. Borstrøm, P. H. Carstensen and C. Sørensen
RISØ-3-9
When organizations embark on manufacturing complex products, a multitude of actors representing different areas of competence cooperate. Because of the complexity of the work due to, for example, the nature of the product and the large number of interdependent participants, part of their work concerns articulation, e.g., coordination, management, allocation, negotiation. This paper is based on a field study at Foss Electric, a Danish manufacturing company. The field study surveyed a large-scale project involving mechanical, electronic, software, and chemical design of a complex instrument for testing the quality of raw milk. We argue that in this particular project a matrix organization, scheduled project meetings, informal meetings, and paper-based artifacts are the primary means of managing the complexity of articulation work. The aim of this paper is to investigate the origination and use of artifacts in articulation work in the project. Artifacts stipulating and mediating articulation work in the project are presented and discussed from the perspective of Mechanisms of Interaction (Simone and Schmidt, 1993) . The analysis supports the hypothesis that when confronted with an abundance of detailed decisions that need to be articulated, organizations invent and adopt artifacts which can be interpreted as stipulating and mediating articulation work in order to reduce the complexity of this work.