Involving users in the design cycle for parallel tools second period September 1, 1994--January 31, 1995
- Oregon State Univ., Corvallis, OR (United States). Dept. of Computer Science
Parallel programmers do not use software tools, in spite fact that parallel development is a difficult and time-consuming task that could benefit from tool support. It has become increasingly clear that the simple availability of elegant, powerful software tools employing the latest technology is not enough. Usability is the real key to success; users simply do not adopt tools that fail to respond to their needs. Research in the area of usability engineering indicates that five design principles can have significant impact on parallel tool usability: tools must be based on demonstrable user requirements; actively involve users throughout tool design; minimize tool complexity to reduce the learning curve support the tool across multiple machine platforms to amortize the user`s investment employ iterative refinement techniques to improve tool usability. Those principles served as the starting point for a Parallel Tools Consortium project to develop a tool that will help users determine the final state of a program that crashes or is terminated forcibly. Carried out over a period of ten months, the project involved the collaboration of tool researchers, and implementors, and users. This report describes how user-centered design techniques were applied to ensure that the tool would provide simple, intuitive support for the programmer`s task. Users were recruited for the project working group so that they could have direct input to design decisions, even in the earliest sets of user trials. Additional feedback was acquired from a broader user base, in three distinct phases. These were spaced out over a period of six months so that feedback could be analyzed, then applied to refine the tool before the next trial. In some cases, user input kept us from investing substantial effort in features that would not have been used or appreciated. In others, feedback showed us where our conceptions of usefulness did not quite align with those of the user community.
- Research Organization:
- Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States); Oregon State Univ., Corvallis, OR (United States). Dept. of Computer Science
- Sponsoring Organization:
- USDOE, Washington, DC (United States)
- DOE Contract Number:
- W-7405-ENG-48
- OSTI ID:
- 82529
- Report Number(s):
- UCRL-CR-120462; ON: DE95013275
- Resource Relation:
- Other Information: PBD: 31 Jan 1995
- Country of Publication:
- United States
- Language:
- English
Similar Records
Automated Cache Performance Analysis And Optimization
DIVA V2.0