<?xml version="1.0" standalone="yes"?>
<Paper uid="W97-1412">
  <Title>A Syndetic Approach to Referring Phenomena in Multimodal Interaction</Title>
  <Section position="6" start_page="7" end_page="7" type="concl">
    <SectionTitle>
8 Conclusions
</SectionTitle>
    <Paragraph position="0"> Traditional approaches to evaluating or comparing input devices have focussed either on the logical behaviour of the device, or ergonomic aspects of its use. This paper has presented a framework that allows analysis of the cognitive ergonomics of interaction, in terms of the mental resources needed to utilise a particular device for a specific task. We have used the model to present a systematic account of the diffences between mouse and touch screen. The example was chosen for familarity, rather than for novelty. Hoowever, the approach is one that can be extended to rather more sophisticated and problematic techniques.</Paragraph>
    <Paragraph position="1"> 92 G.P. Faconti and M. Massink One argument raised against the wider use of syndetic modelling for human factors evaluations is the level of formality involved. This is a reasonable concern, and there are two responses. The first is that the work on syndesis carried out so far has been primarily concerned with establishing its feasibility as a model for explaining interaction, rather than as a practical tool for industrial use. We are now beginning to explore means by which the level of formality can be tamed, both by supporting development of formal models with software tools, or by encapsulating the technique within a tool to support scenario-driven analysis of interaction.</Paragraph>
    <Paragraph position="2"> The second response to concern about formalism is that the complexity of modern interfaces, and the subtle demands that they place on users' cognitive abilities, calls for an expressive and analytically powerful method for modelling and evaluation. Such a method needs to be able to span both the user and the system, in order to capture the interplay between the information available from the system, the actions that can be taken, and the tasks and knowledge of the user. We are not advocating syndetic models as a replacement for other design representations. There is an inherent trade-off between the power and generality of a notation (Blanford and Duke, 1998), and there are important issues, for example based on social factors or domain requirements, for which syndetic models are either inappropriate or inadequate. Likewise however, syndesis brings considerable analytical power and authority (in the form of the underlying cognitive theory) to bear on problems whose complexity makes the use of less formal design techniques problematic.</Paragraph>
  </Section>
class="xml-element"></Paper>