In lieu of an abstract, here is a brief excerpt of the content:

1 Perceptions of Inevitability In August of 2005, a team of experienced technology developers at Autoworks—the pseudonym that will be used here for a major American automaker1 —filed into a dark conference room for yet another meeting with consultants from a young and eager software startup firm. In the past two years, Autoworks’ team of developers had met with representatives from the startup at least two dozen times to discuss the specifications for CrashLab , a computer-based simulation technology that was supposed to “revolutionize product design work at Autoworks.” Looking at the faces of the consultants from the startup, one could see they were excited to be there. Helping to build a mathematically sophisticated computer-based simulation tool for a major automobile company was not only good work; it also helped to raise the company’s profile in the industry. The developers from Autoworks looked less than overjoyed. This particular startup was the third vendor with which Autoworks had contracted in a ten-year period to bring CrashLab to the engineering-analysis community. Patience was running thin at various levels of the Autoworks hierarchy. One of the company’s senior directors asked “Where is the damn technology? Why is it taking so long? They’ve been working on this thing for almost ten years and we still haven’t been able to use it in our vehicle design work.” The developers from Autoworks were well aware of these concerns and also felt the increasing pressure of a diminishing budget in a company that tightened its purse strings with every new quarterly announcement of falling sales. Just after 7 a.m., a discussion of several problems that had surfaced in a recent round of CrashLab’s user testing began. The lead consultant from the startup took copious notes as two of Autoworks’ developers recounted the bugs they had found as they tried to replicate the actions an engineering analyst would take as she interacted with the technology. In response to one of these problems, a young consultant from the startup made what would seem, from the outside, to be a simple suggestion: 2 Chapter 1 We could probably fix this problem pretty easily if we didn’t constrain the location where engineers put their section cuts. I mean a section cut is just a cross-sectional view of the part where engineers measure the energy produced in a crash, right? So, if they want to make those section cuts in other places it’s not that much of a change. I think we should be able to accommodate it easily and get it taken care of. Would that work for you guys? The room fell silent. The developers from Autoworks looked at each other, surprised that the young consultant would propose something that, to them, seemed preposterous. Martin Yuen, one of the senior technology developers from Autoworks, leaned back in his chair and answered coolly: There’s no way we can change the location like you suggest. We just can’t do it. We have to define a specific process of making section cuts. CrashLab is about making sure everyone works the same way. That’s what it’s supposed to do; it’s not up to us; it’s just what this technology is supposed to do. That’s just the next logical step in the sequence of where we’ve been with technologies. We can’t argue with that. We’ll just have to find another way around this issue because that procedure will stay intact. That’s just the way it is. Six months later, Suresh, an engineering analyst at Autoworks who was responsible for conducting computer-based simulations of frontal-impact crash tests, sat down at his workstation and opened a software application that had been in development at the company for about ten years. Suresh hoped that CrashLab would help him prepare his simulation model for a series of tests. He had taken a three days of training on how to use the technology when it had first come out, and he recalled that the instructor demonstrated how, with no more than a few mouse clicks, one could automatically set up section cuts and then use them to determine the energies produced at the sections indicated in the model. The idea of using a technology that would perform this tiresome procedure automatically was very appealing to Suresh. He opened CrashLab, pulled out a reference guide that had been given to him during his training, and...

Share