
Spinal Tap Computing: An Interview with David M. Weiss

WOW caught up with David Weiss after his keynote address that capped the David L. Parnas Symposium, and asked him to expand on a few of his points. 
You mentioned that this was the David Parnas era in software engineering. What other eras are on the horizon? 
No single figure has, or indeed will have, as much influence in the field of software engineering as David Parnas. The field is expanding so explosively that it is simply too hard for one person to keep on top of it. As far as new themes go, in the applications area the next era will be defined by programs that interoperate on their own, without human involvement. Think of interconnected business networks populated by processes that communicate in an autonomous fashion.  Problems of security and trust will have to be overcome before this can happen, of course. In the area of software development, we need tools that reflect the shift from computing to information processing, that are better integrated with the way that development actually happens. Essentially, we need tools that create tools, like what has happened in the XML community. Software engineers are always the last people to develop tools for themselves.
You mentioned the recurring nature of many software problems, that many issues today are simply old problems in new clothing.  Are there any brand new problems out there?
Yes there are. Your readers may think Im crazy, but Im convinced that a crucial problem for the future is how to connect users directly to computers, such as providing a direct connection to the optic nerve.  We may think that current GUI technology is easy to use, but watch how a non-technical person uses a computerthey have an incredibly difficult time, at least initially!  We need to find a way to merge computing with the natural human sensessight, hearing, speakingsuch that there is a seamless path between the two. If we can decode the human genome, then we can decode human sensory signals such that direct input and output is feasible.

John M. Linebarger