|
Engineering design is a complex technical pursuit mediated by social
processes such as communication, negotiation, and shared agreements.
In this study, we regard engineering design practice as a process of
"story telling". We define "story telling" in design as establishing
a coherent story about the design process and the designed artefact
by bringing coherence to the perspectives and interests of each design
team member. The "stories" generated explore various aspects of the
design process and the designed artefact. They capture the real-world
context in which the design concepts evolved in the product design
lifecycle. They reflect the conflicting interests and resulting
reconciliation and shared agreements of the design team members.
Studying engineering design from the viewpoint of "story telling"
frames design in an emergent perspective to understand how collective
properties of design - technical problem solving, social networks,
information processing - contribute to a narrative of the design
process and designed artefact.
The research focuses on the study of the design process by analyzing
time variant patterns of "story telling" in engineering design teams
using computational linguistics. Three research issues are discussed
in relation to examining the time variant patterns of "story telling".
First, based on our definition of design "story telling", how can we
identify, represent, and visualize "story telling" of engineering design
teams? Second, how does the "story telling" evolve over time? And finally,
are there any correlations between patterns in design "story telling" and
the outcome of the team? That is, what insights into the design process
will quantitatively depicting patterns in "story telling" provide?
A multidisciplinary graduate design course at the University of California,
Berkeley provided the test bed for this study. We examined the oral and
written histories left by the designers through their documentation,
presentation material and e-mail communication. We applied an established
computational linguistic technique, Latent Semantic Analysis (LSA), for
analyzing time variation of "story telling" in eight student design teams
based on the metrics of semantic coherence. The coherence metric applied in
this research adopted a "cosine" measurement, which is appropriate for
measuring the pair-wise coherence of documents consecutive in time to detect
the change in coherence over discrete periods of time. Intra-stage and
cross-stage semantic coherence were then defined to study patterns of the
semantic coherence within each design stage over time and the variation of
semantic coherence between design stages respectively. We also measured the
e-mail communication for each team, including the volume of messages, the
frequency of messages for each of the product design stages, and individual
team members to compare how the amount of communication might have affected
"story telling" and to compare the "story telling" capacity between the e-mail
and the documentation data sets.
This research provides empirical evidence of the phenomena of changing levels
of coherence in "story telling" in design and the scope of design concepts
explored by design teams. Results from the analysis suggest a positive correlation
between design outcomes and patterns of the average semantic coherence over
time as well as with variation in semantic coherence between design stages. We
also show that both the email data set and the documentation data sets contain
approximately the same amount of "story telling" capacity. This study establishes
a formal methodology for providing a real-time window into the design process
and coherence of design thinking of the design teams. The methodology is proposed
as means for contextualizing portraiture of design in terms of unique teams and
situated design contexts. Design teams could also apply the method to reflect on
their own performance with the ability to address important process questions.
Are we too focused when we should be expanding our scope? Have we been diverging
for too long? Our goal is to eventually sort the metrics of local performance of
design teams to global metrics of performance.
|
