Full title

Dependability as ordinary action

Keywords

production management, ethnography,

Summary

An ethnographic study of work in an industrial production environment has highlighted the importance situated action in achieving dependability of the production process. The same setting has afforded an investigation of the practices of IT professionals concerned with managing the various systems supporting production work. This latter part has highlighted the role that a biographical familiarity with the setting can play in systems development and ongoing management.

DEPENDABILITY AS ORDINARY ACTION

We draw on an ethnographic study of work in a production plant that illustrates how dependability is realised in and as part of peoples' everyday ordinary activities. The case study organisation, ENGINECO, produces mass-customised diesel engines from 11 to 190kW. Production in its plant is designed to work along a strict production orthodoxy and large parts are automated. An ethnographic study of working practices in the control room has been conducted, followed by an investigation of the practices of IT professionals who are charged with procuring, installing, operating and maintaining the various IT systems that control production processes.

While production work is highly regulated and carried out in an environment that reflects the 'production orthodoxy' in its physical and social arrangements, we found a range of situated practices that constituted a core part of peoples' work in the control room. An example is provided by this extract from the shift book that is used to record important information about the production process and to communicate this information to people on other shifts:

As soon as crankcases for 4-cylinders are available, schedule ~ order number 87965576 (very urgent for Company X)

In this example, workers exhibit their knowledge of the state of production and the priorities that apply in specific situations as well as the general rules for production management. The case described by this shift book entry constitutes a local modification of a pre-computed production schedule that takes into account not only the ordering of delivery dates and bills of materials (i.e. certain engines requiring a particular crank case) but also the relative importance of particular customers to the company and the relations between these companies. In principle, it would have been sufficient for the purposes of ensuring that an order be fulfilled as soon as possible to state that it is urgent (i.e. close to the delivery date) but the worker chose to include information about who the customer is and thereby making available to his colleagues information that enables to draw on a repertoire of resources to handle the situation.

Given that the order is for Company X, they can make a judgement about, for example, what kinds of measures are appropriate to get the job done and speed up the production process, e.g. by assigning higher priorities relative to other orders, by having someone work extra hours or simply by closely monitoring this order's progress in order to spot potential problems and delays as early as possible.

The practices described above are important in that they are crucial in establishing an orderly production process. In contrast to a mechanistic view that would emphasise the importance of prescriptions for action and or planning, this example shows how the prescriptions and plans are instantiated through peoples' resourceful action. Dependability of the production process is something that is only partially achieved by the design of the plant and its formal arrangements but is rather crucially dependent on ordinary, everyday, seen-but-unnoticed activities of production workers. A last important point is that these activities are inherently social in nature and thus need to be described in terms of the working division of labour in the workplace and in terms of what people know and use as they go about their business.

DESIGN as a team MEMBER

In addition to the ethnographic study of production work, we have had the opportunity to get involved in an study the work of IT professionals in the same setting. The local IT department is co-located with production in the same building and its staff are intimately familiar with the details of the production process through their day-to-day involvement in the management of production and of the IT systems that support it. The importance of this 'biographical familiarity' with the setting becomes clear in the following example, where a member of IT staff and a production worker discuss a new system designed to record production statistics as well as information about breakdowns. Its development has been contracted out to an external company and while this supplier has staff on-site, they are much less familiar with the biography of the setting.

Michael: Normally, the stations should be ordered as they are on the assembly line... ~ Barbara: Yes. That's exactly how I've entered them but someone has changed the sorting order.

Barbara and Michael were both involved in drawing up the requirements document for the system which was then handed over to the supplier. The order of station identifiers in a drop-down list was implicitly specified by Barbara giving the list in this particular order which corresponds to the physical arrangements in the plant. Not realising the significance of this ordering, the developer involved changed the order to an alphabetical one. As production workers may well use different names to identify a station, this ordering produces a problem of potential ambiguity and makes searching for a station harder.

Biographical familiarity, i.e. knowing how things get done in a place, is an important resource for designers. Requirements specifications can only capture some details of how a system will fit in with working practices. Various methods have been devised to improve the process of 'informing design' but as long as the separation between designers and users is maintained, they will only be patches for the problem. Our investigation of corealisation as an alternative orientation to design aims to tackle the problem through a fundamental respecification of the practice of design as a collaborative activity that both designers and users are routinely involved in (as opposed to exceptionally as in many participatory design methods).

links

Co-realisation

PUBLICATIONS

Mark Hartswood, Rob Procter, Roger Slack, Alex Voß, Monika Buscher, Mark Rouncefield, Philippe Rouchy. Co-realisation: Towards a Principled Synthesis of Ethnomethodology and Participatory Design. Scandinavian Journal of Information Systems, 14(2), 2002.

Mark Hartswood, Rob Procter, Roger Slack, James Soutter, Alex Voß, Mark Rouncefield. The Benefits of a Long Engagement: From Contextual Design to The Co-realisation of Work Affording Artefacts. Proceedings of NordiCHI. 2002.

Monika Büscher, Dan Shapiro, Mark Hartswood, Rob Procter, Roger Slack, Alex Voß, Preben Mogensen. Promises, Premises and Risks: Sharing Responsibilities, Working Up Trust and Sustaining Commitment in Participatory Design Projects. T. Binder, J. Gregory, I. Wagner (eds.) PDC'2002 Proceedings of the Participatory Design Conference. June, 2002. pages 183--192.

Alexander Voß, Roger Slack, Rob Procter, Robin Williams, Mark Hartswood, Mark Rouncefield. Dependability as Ordinary Action. Stuart Anderson, Sandro Bologna, Massimo Felici (eds.) Computer Safety, Reliability and Security: Proceedings of the 21st International Conference, SAFECOMP 2002. September, 2002. pages 32--43.

Alexander Voß, Rob Procter, Robin Williams. Innovation in Use: Interleaving day-to-day operation and systems development. T. Cherkasky, J. Greenbaum, P. Mambery (eds.) Proceedings of the Participatory Design Conference. 28 November to 1 December, 2000.