Transfer of Training Between 3-D Computer-aided Design (CAD) Systems

Eric N. Wiebe
North Carolina State University
Raleigh, NC
eric _wiebe@ncsu.edu

During the late 1980's and early 1990's many mechanical design groups in manufacturing were faced with the movement from computer systems based around 2-D Computer-aided Design (CAD) systems to 3-D modeling systems.  A common issue at the time centered around 2-D to 3-D CAD system transfer of training issues (Gattiker, 1992).  More recently, many of these companies are faced with a new transfer of training issue, this time surrounding the move from one 3-D CAD system to another.  This issue has become particularly pronounced with the increased mobility of industry professionals and the emergence of more cost-effective (thus, more widespread) Windows-based modeling systems.
Shneiderman (1998) has proposed a model of human-computer interaction, the Object-Action Interface (OAI) model, which provides a structure for understanding how users transfer real-world tasks to the computer interface of a software package.  Part of this transfer process involves the translation of higher-level task strategies (the semantic level) into software specific commands (the syntactic level).  Previous research has indicated that many of the most popular 3-D modeling packages all contain commands/tools which support the same higher-level modeling strategies for modeling simple parts.  In contrast, all of the packages have different interface elements, which create different syntaxes for achieving these higher-level goals.
The purpose of this research was to test the hypothesis that, with minimal training, users can transfer their established high level modeling strategies between packages (thus preserving an important element of efficiency and effectiveness).  This transfer will take place once they have mastered the basic syntax necessary to execute their strategies.
Individuals holding at least an intermediate level of expertise on a specific modeling software package (Pro/ENGINEER) were recruited to model a specific part on this package.  In a second phase of the experiment, they were then asked to model the same part on a different modeler that they had never worked on before (SolidWorks).  These two modelers were very similar in their part modeling functionality, but had markedly different surface interface characteristics.  The second phase of the experiment was run as a 'coached walkthrough' (Mack & Robinson, 1992).  The investigator stayed with the user and acted as a resource when the user had questions about how to use the software.  The user's interaction with the software was recorded on videotape while all utterances were recorded simultaneously on the audio track of the video.
The results of the study clearly showed the initial conflict when there were syntactic differences between the software tools in executing a specific modeling strategy.  A majority of these conflicts, however, did not re-occur once they were resolved by the investigator/coach at the time of first encounter.  There was also clear evidence as to the ease in which original higher-level task strategies could be preserved on the new system.  The results of this study demonstrated clear differences between the transfer of syntactic and semantic level knowledge from one 3-D modeling system and another.  The results of this study also has implications for the design of training programs, pointing to the importance of delineating between software package-dependent and -independent skills and knowledge.
Acknowledgements
This research work was supported by funding from the Furniture Manufacturing and Management Center, NC State University, Raleigh, NC, USA.
References
Gattiker, U. E. (1992). Computer skills aquisition: A review and future directions for research. Journal of Management, 18, 547-574.
Mack, R., & Robinson, J. B. (1992). When novices elict knowledge: Question asking in designing, evaluating, and learning to use software. In R. R. Hoffman (Ed.), The psychology of expertise: Cognitive research and empirical AI, (pp. 245-268). New York, NY: Springer-Verlag.
Shneiderman, B. (1998). Designing the user interface: Strategies for effective human-computer interaction. (3rd ed.). Reading, MA: Addison-Wesley.

Presented at the IES 2000/HFES 2000 Congress. August, 2000, San Diego, CA, USA