Authors:
At its heart, our philosophy is simple: All software systems are tools, so, to empower users, give them better tools. Since good tools fit the work, we employ task modeling within a disciplined engineering process. That planes are designed to fly and bridges to stand does not mean that they cannot also be innovative, elegant, or even beautiful. Breakthrough designs arise from the tensions inherent in addressing competing needs, objectives, and constraints and from the conviction that a creative synthesis resolving such apparent conflicts is possible.
We also believe in the power of models and modeling. Abstract models help us design smaller, simpler systems that are both inventive and effective, that provide users with what they truly need rather than what they merely want.
We distinguish our usage-centered approach from mainstream practices of user-centered design. Our focus is usage more than users, models more than user input and feedback. Instead of design-by-discovery through multiple rounds of rapid prototyping and user feedback, our objective is to get it essentially right the first time out of the gate.
We find that world-class solutions require up-front design by a dedicated team, preferably an interdisciplinary one involving visual and interaction designers along with technical specialists and problem-area experts. Believing that usability is everyone's job, we also teach developers usability principles and techniques and collaborate with them as partners in the design process.
We practice usage-centered design, a systematic, outside-in design process driven by three simple models: a model of roles users can play in relationship with the system, a model of tasks users need supported, and a model of the content and organization of the user interface. We use task cases rather than scenarios because greater abstraction and finer resolution promote deeper understanding of the essence of user tasks. (See "Essential modeling: use cases for user interfaces," interactions 2, 2, pp. 3446.) Abstract rather than representational prototypes help our designers, clients, and users to focus on what needs to be present where on the user interface without preconceptions or preoccupations with what it looks like or how it will work. We consistently find that abstract modeling promotes innovative solutions that reflect the real nature of a problem.
Our process has evolved over nearly a decade of application and continues to be refined. The demands of "Web-time" development and compressed release cycles, for example, have led to simplified "agile" modeling techniques based on index cards and sticky notes, and large-scale industrial applications have led to software tools for managing complex models.
Siemens' STEP 7 Lite is a new integrated development environment (IDE) for applications on specialized computers called programmable logic controllers (PLCs) that operate modern automation equipment. End-to-end, usage-centered design enabled us to devise a breakthrough in PLC programming that is immediately usable by first-timers yet highly efficient and flexible in the hands of expert programmers. For this work, Helmut Windl, who headed the Siemens team, and I were honored with first place in the 2001 Performance-Centered Design Competition.
Although large, this 50-person-year project is representative of our basic design process. The design team started by identifying, describing, and prioritizing 19 different user roles. From these we constructed a task model consisting of 342 distinct task cases. Questions raised in the course of modeling were answered with the aid of user interviews and observations. From the task model, we were able to organize the user interface to collect related tools and information according to how they were most likely to be used. User efficiency was improved not only through a close fit to the supported tasks but also through reducing the overhead associated with window management. Overall, the number of user steps for a representative mix of real-world tasks was cut in half compared with earlier systems.
We used abstract prototypes to model the contents of 25 interaction contexts, then developed annotated paper prototypes to detail the visual and interaction design. Assisted by our three consultants, the five-person Siemens modeling and design team completed the user interface design, ready for inspection, just 14 weeks after the first team meeting!
Inspection of the paper prototype uncovered 117 usability defects, mostly minor and easily fixed; a handful of more significant defects also proved to be easily corrected. User participants in the inspections were not only able to complete all tasks in representative scenarios but found the novel design to be immediately understandable and easy to use. Later, extensive usability tests of the beta release in both the United States and Germany further proved the robustness of the design. Overall, the system achieved the highest level of measured usability of any product ever tested by Siemens's U.S. lab.
It is impossible to do full justice to this sophisticated design in the space available here (more details are available at http://foruse.com/pcd/), but a few aspects bear highlighting. Although the user interface resembles a conventional Windows application, it relies on novel organization and incorporates numerous innovations to better support user tasks. An Explorer-like tree view (Figure 1) is actually a custom control tailored to simplify project navigation and provide quick access to sets of closely related views. Rollout panels automatically supply tools and materials as needed, anticipating Microsoft's task panes. Cascading tool tips, another invention of ours, present progressive guidance tied directly into the integrally designed help system.
The solution makes extensive use of instructive interaction, a body of techniques for making innovative interfaces intrinsically self-teaching (Figure 2). For example, the complex task of configuring a project to conform to the target hardware is simplified by visual presentations corresponding to actual equipment and by dynamic visual cues that link synchronized views and convey affordances based on complex rules embodied in the system.
Director of Research and Development
Constantine & Lockwood, Ltd.
Adjunct Professor
University of Technology
Sydney, Australia
larry@forUse.com
Phone (United States): +1 (978) 948-5012
A pioneer of modern software engineering theory and practice, Constantine is a leading exponent of task modeling in user interface design. A consultant, educator, and prize-winning author and designer, his publications span more than 150 papers and 16 books, including the 1999 Jolt Award winner, Software for Use: A Practical Guide to the Models and Methods of Usage-Centered Design, written with Lucy Lockwood.
Figure 1. Hardware configuration in the STEP 7 Lite
IDE
Figure 2. Examples of instructive interaction in STEP 7
Lite.
Constantine & Lockwood, Ltd., is a small-town team playing in the big leagues. We often go head-to-head with prominent consultants and design houses, and our clients are mostly large multinational corporations and leaders in their fields. We are a rarity in that we both teach what we do and do what we teach. For example, after a leading analyst firm discovered that its new portal from one of the biggest Web design shops was monumentally unworkable, the firm asked us to solve the usability problems while preserving the unique aesthetic of the original design. Later, an in-house team trained by us in usage-centered design produced the much-improved next generation of the site. By drawing on a network of highly talented consulting associates, we find we can field a team and devise a strategy tailored to meet client needs.
Sidebar: Practitioner's Workbench
Resources
Many, but the handbook of usage-centered design is (of course) Constantine and Lockwood, Software for Use (Addison-Wesley, 1999). We mostly rely on numerous Web resources to keep current, including weekly updates from EPSSCentral.com.
Tools
Low-tech rulesindex cards for inventorying and prioritizing user roles and task cases, sticky notes for modeling interface contents (but we prefer digital whiteboards with computer connections (e.g., mimio, eBeam) for problem-solving sessions and sketching prototypes. Microsoft PowerPoint® gives us nice renderings with simulated behavior when needed. Abstract models that leverage human thought are our most ubiquitous and valued tools.
Sources of inspiration
Bill Buxton. We aspire to the unique blend of invention and pragmatism so evident in his work.
Work. We love the work we do and strive through dogged determination to define visual and interaction designs that truly fit the work our users are doing, making it easier and more fun. That absolute commitment to supporting work and to workability is what has consistently led us to design innovations that push the envelopecascading tool tips, drop-enabled "hover" tabs, instructive interactionand to a string of patents.
Process. The process works and keeps proving itself again and again. In the hands of talented people who "get it," the envelope just keeps expanding.
Teaching. Teaching both tests and inspires us. We thrive on empowering professionals and students with new ideas and techniques and watching their energy and excitement grow as they "get it."
©2002 ACM 1072-5220/02/0300 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2002 ACM, Inc.
Post Comment
No Comments Found