Adapting the ADS for High Volume Manufacturing Connor Upton, Gavin Doherty Distributed Systems Group, Dept of Computer Science, Trinity College Dublin Abstract. Cognitive Work Analysis (CWA) is a methodology for analyzing complex socio-technical systems. It aims to structure system information in a manner that is meaningful for human control and interaction. The Abstraction Decomposition Space (ADS) in an important tool used during the first phase of CWA to describe the work domain. In this paper we create an ADS for a Semi- conductor Fabrication Plant (Fab). This is a High Volume Manufacturing envi- ronment and its complexity necessitates a number of adjustments to the original ADS technique. The physical decomposition of the system is de-emphasized and a number of alternative decomposition hierarchies are used instead. The analysis aims to produce artifacts that can aid in the design of decision support systems. These artifacts not only help to assess the information needs of work- ers, but also structure the work domain in a manner that will inform display de- sign. 1. Introduction The correct visual representation of data has been shown to improve user performance and reduce human error in a range of domains [1] and many guidelines exist for the correct visual encoding of quantitative data [2,3]. Advances in sensor and communica- tions technology means that more data is now being generated than ever before. Automated control systems are frequently used to process this data but human opera- tors are often relied on to step in and assume system control if required. In these cases operators must examine data to evaluate the system state and make decisions. The complexity of these domains means that the challenge is not only how to encode the data visually, but also how to decide what data is required for the tasks at hand and how to navigate through the information space. These systems, described by Vicente [4] as Complex Socio-technical Systems, generally involve: large problem spaces, multiple users, conflicting constraints, dynamic data, coupled components and unan- ticipated events. These attributes make it difficult to apply a purely task-oriented analysis approach when designing user interfaces. Cognitive Work Analysis (CWA) [4] is an alternative approach that attempts to structure system information in a man- ner that is meaningful for human control and interaction. It produces a number of design artifacts that can inform a UI designer about both the system and the user’s information requirements. 2. The Abstraction-Decomposition Space CWA structures system information using multiple levels of abstraction. The aim is to support reasoning about a system rather than providing a set path of interaction to- wards a predefined goal. The approach is closely tied to Rasmussen’s SRK Taxonomy