Faulty Human Communication: Its Impact on Dependability and Ways to Control It
Speaker: John C. Knight , Dept. of Computer Science, University of Virginia
Date: May 8 2006
Time: 4:00PM to 5:00PM
Location: 32-D463 (Stata Center-Star Conference Room)
Host: Brian C. Williams, MIT CSAIL
Contact: Marcia Davidson, 617-253-2448, marcia@csail.mit.edu
Relevant URL: http://dependability.cs.virginia.edu/info/Welcome
For computing systems requiring ultra-high dependability, such as those in
the aerospace domain, a significant fraction of defects that are found during
testing or after deployment are the result of defective communication between
engineers. Many defects arise because of misunderstandings between engineers
during requirements analysis, and many others arise because of
misunderstandings about the impact of engineering decisions on dependability
goals.
Natural language is an essential medium for engineering communication, including
during requirements analysis, yet it is understood to be problematic for
high-precision communication. However, natural language possesses its own body
of research results that can be brought to bear on this issue. We have analyzed
the communication problem in software engineering from the perspective of
current cognitive linguistic theory, and this analysis has yielded a model that
helps to explain sources of ambiguity and other problems with the use of natural
language in software development.
A second source of defects is misunderstandings about the effect of engineering
decisions on system dependability. Arguments are made about the effects of
specific engineering choices, and flawed arguments are often accepted because
the flaws are not recognized. As with natural language, there is a body of
theory about rigorous argument that can attack this problem. We have analyzed
this problem from the perspective of modern philosophical argument, and
developed a taxonomy of fallacies that arise in assurance arguments.
In this presentation, I will summarize the linguistic model and the fallacy
taxonomy, and I will present insights derived from them. I will explain how
these insights are exploited and some details of our applications of the
concepts discussed in safety-critical systems.
For additional information, see:
http://dependability.cs.virginia.edu/info/Welcome and
http://www.cs.virginia.edu/~jck/recentpapers.htm
See other events that are part of
See other events happening in May 2006
