Usability Basics for Software Developers Usability Engineering

focususability engineering Usability Basics for Software Developers

Xavier Ferré and Natalia Juristo, Universidad Politécnica de Madrid Helmut Windl, Siemens AG, Germany Larry Constantine, Constantine & Lockwood

n recent years, software system usability has made some interesting ad- vances, with more and more organizations starting to take usability seri- ously.1 Unfortunately, the average developer has not adopted these new I concepts, so the usability level of software products has not improved.

Contrary to what some might think, us- Usability engineering defines the target us- This tutorial ability is not just the appearance of the user ability level in advance and ensures that the examines the interface (UI). Usability relates to how the software developed reaches that level. The system interacts with the user, and it includes term was coined to reflect the engineering ap- relationship five basic attributes: learnability, efficiency, proach some usability specialists take.3 It is between usability user retention over time, error rate, and sat- “a process through which usability character- and the user isfaction. Here, we present the general us- istics are specified, quantitatively and early in interface and ability process for building a system with the the development process, and measured desired level of usability. This process, which throughout the process.”4 Usability is an is- discusses how most usability practitioners apply with slight sue we can approach from multiple view- the usability variation, is structured around a design- points, which is why many different disci- process follows a evaluate-redesign cycle. Practitioners initiate plines, such as psychology, computer science, design-evaluate- the process by analyzing the targeted users and sociology, are trying to tackle it. Unfor- and the tasks those users will perform. tunately, this results in a lack of standard ter- redesign minology. In fact, the term usability engineer- cycle. It also Clarifying usability concepts ing is not universally accepted—other terms discusses some According to ISO 9241, Part 11, usability used include usage-centered design, contex- management is “the extent to which a product can be used tual design, participatory design, and goal- by specified users to achieve specified goals directed design. All these philosophies adhere issues an with effectiveness, efficiency, and satisfaction to some extent to the core issue of usability organization in a specified context of use.”2 This definition engineering: evaluating usability with real must face when ties a system’s usability to specific conditions, users from the first stages of development. applying usability needs, and users—it requires establishing certain levels of usability based on the five basic Usability attributes techniques. attributes. We can’t define usability as a specific as-

22 IEEE SOFTWARE January/February 2001 0740-7459/00/$10.00 © 2001 IEEE pect of a system. It differs depending on the be carefully designed if it requires both high intended use of the system under develop- learnability and high efficiency—for exam- ment. For example, a museum kiosk must run ple, using accelerators (a combination of keys It’s important a software system that requires minimum to perform a frequent task) usually solves this training, as the majority of users will use it conflict. The point is that a system’s usability to carefully just once in their lifetime. Some aspects of us- is not merely the sum of these attributes’ val- consider the ability—such as efficiency (the number of ues; it is defined as reaching a certain level for interaction tasks per hour)—are irrelevant for this kind each attribute. of system, but ease of learning is critical. We can further divide these attributes to not just when However, a bank cashier’s system would re- precisely address the aspects of usability in designing quire training and would need to be highly ef- which we are most interested. For example, ficient to help reduce customer queuing time. performance in normal use and advanced the visible Because usability is too abstract a term to feature usage are both subattributes of effi- part of the study directly, it is usually divided into the ciency, and first impression is a subattribute user interface, attributes we mentioned at the beginning of of satisfaction. Therefore, when analyzing a the article:5 particular system’s usability, we decompose but also when the most important usability attributes designing the Learnability: How easy it is to learn the down to the right detail level. main system functionality and gain pro- Usability is not only concerned with soft- rest of the ficiency to complete the job. We usually ware interaction. It is also concerned with system. assess this by measuring the time a user help features, user documentation, and in- spends working with the system before stallation instructions. that user can complete certain tasks in the time it would take an expert to com- Usability and the user interface plete the same tasks. This attribute is We distinguish between the visible part of very important for novice users. the UI (buttons, pull-down menus, check- Efficiency: The number of tasks per unit boxes, background color, and so forth) and of time that the user can perform using the interaction part of the system to under- the system. We look for the maximum stand the depth and scope of a system’s us- speed of user task performance. The ability. (By interaction, we mean the coordi- higher system usability is, the faster the nation of the information exchange between user can perform the task and complete the user and the system.) It’s important to the job. carefully consider the interaction not just User retention over time: It is critical for when designing the visible part of the UI, but intermittent users to be able to use the also when designing the rest of the system. system without having to climb the learn- For example, if a system must provide ing curve again. This attribute reflects continuous feedback to the user, the devel- how well the user remembers how the opers need to consider this when designing system works after a period of nonusage. the time-consuming system operations. Error rate: This attribute contributes neg- They should design the system so it can fre- atively to usability. It does not refer to sys- quently send information to the UI to keep tem errors. On the contrary, it addresses the user informed about the operation’s cur- the number of errors the user makes while rent status. The system could display this in- performing a task. Good usability implies formation as a percentage-completed bar, as a low error rate. Errors reduce efficiency in some software installation programs. and user satisfaction, and they can be Unfortunately, it is not unusual to find de- seen as a failure to communicate to the velopment teams that think they can design user the right way of doing things. the system and then have the “usability team” Satisfaction: This shows a user’s subjec- make it usable by designing a nice set of con- tive impression of the system. trols, adding the right color combination, and using the right font. This approach is clearly One problem concerning usability is that wrong. Developers must consider user inter- these attributes sometimes conflict. For ex- action from the beginning of the development ample, learnability and efficiency usually in- process. Their understanding of the interac- fluence each other negatively. A system must tion will affect the final product’s usability.

January/February 2001 IEEE SOFTWARE 23 Usability in software development ment cycle—when major usability problems The main reason for applying usability are very costly, if not impossible, to fix. Usability testing techniques when developing a software sys- Therefore, it is crucial to evaluate all results tem is to increase user efficiency and satis- during the product development process, alone is not faction and, consequently, productivity. Us- which ultimately leads to an iterative devel- enough to ability techniques, therefore, can help any opment process. A pure waterfall approach output a highly software system reach its goal by helping the to software development makes introducing users perform their tasks. Furthermore, good usability techniques fairly impossible. usable product, usability is gaining importance in a world in All software applications are tools that because it which users are less computer literate and help users accomplish certain tasks. However, can’t afford to spend a long time learning before we can build usable software tools— uncovers but how a system works. Usability is critical for or, rather, design a UI—we need information does not fix user system acceptance: If users don’t think about the people who will use the tool: design the system will help them perform their tasks, they are less likely to accept it. It’s possible Who are the system users? problems. they won’t use the system at all or will use it What will they need to accomplish? inefficiently after deployment. If we don’t What will they need from the system to properly support the user task, we are not accomplish this? meeting user needs and are missing the main How should the system supply what objective of building a software system. they need? For a software development organization operating in a competitive market, failure to The usability process helps user interac- address usability can lead to a loss of market tion designers answer these questions dur- share should a competitor release a product ing the analysis phase and supports the de- with higher usability. Also, a software prod- sign in the design phase (see Figure 1). uct with better usability will result in re- There are many usability methods—all duced support costs (in terms of hotlines, essentially based on the same usability customer support service, and so forth). process—so we have abstracted a generic us- Even if a system is being used, it does not ability process from the different approaches necessarily mean it has a high level of us- to usability mentioned earlier. We hope this ability. There are other aspects of a software makes it easier for the reader to understand product that condition its usage, such as the different usability techniques we will be price, possibility of choice, or previous describing. training. In addition, because users are still more intelligent than computers, it is usu- Usability analysis phase ally the human who adapts to the computer First, we have to get to know the users and in human–computer interaction. However, their needs, expectations, interests, behav- we shouldn’t force the user to adapt to soft- iors, and responsibilities, all of which charac- ware with poor usability, because this adap- terize their relationship with the system. tation can negatively influence efficiency, effectiveness, and satisfaction. Usability is a User analysis. There are numerous ap- key aspect of a software product’s success. proaches for gathering information about users, depending on each individual system The usability process under development and the effort or time As we mentioned, a system’s usability de- constraints for this phase. The main meth- pends on the interaction design. Therefore, ods are site visits, focus groups, surveys, we must deal with system usability through- and derived data. out the entire development process. Usabil- The primary source for user information ity testing alone is not enough to output a is site visits. Developers observe the users in highly usable product, because usability test- their working environment, using the sys- ing uncovers but does not fix design prob- tem to be replaced or performing their tasks lems. Furthermore, usability testing has been manually if there is no existing tool. In ad- viewed as similar to other types of software dition, developers interview the users to un- quality assurance testing, so developers of- derstand their motivation and the strategy ten apply the techniques late in the develop- behind their actions. A well-known method

24 IEEE SOFTWARE January/February 2001 Figure 1. The usability process.

User Task Evaluation Usability Conceptual EvaluationVisual Evaluation analysis analysis benchmarks design design

Analysis phase Design phase

for doing user analysis jointly with task importance and frequency to get a small analysis is contextual inquiry.6 This method task set. This approach guarantees that provides a structured way for gathering and you’ll build the most important functionali- organizing information. ties into the system and that the product A focus group is an organized discussion will not suffer from “featuritis.” These tasks with a selected group of users. The goal is to should be the starting point for developing gather information about their views and the system. One approach to analysis is to experiences concerning a topic. It is well build a task model within the Usage- suited for getting several viewpoints about Centered Design method, a model-driven the same topic—for example, if there is a approach for designing highly usable soft- particular software product to discuss—and ware applications, where tasks, described as gaining insight into people’s understanding essential use cases, are the basis for a well- of everyday system use. structured process and drive UI design.9 In a survey, the quality of the information Task analysis ends when we evaluate the depends on the quality of the questions. discovered task set, which is best done collab- Surveys are a one-way source, because it is oratively with users. When the user popula- often difficult or even impossible to check tion is already performing a set of tasks, we back with the user. Don A. Dillman’s book perform task analysis during user analysis to Mail and Internet Surveys provides a struc- apprehend the tasks the user performs rou- tured method for planning, designing, and tinely and how the user perceives these tasks. conducting surveys.7 After the optional first analysis, we identify Derived data includes hotline reports, the tasks our system will support, based on a customer complaint letters, and so forth. study of the goals the user wants to attain. It can be a good source of usability impli- Then, we break the tasks into subtasks and cations but is often difficult to interpret. into particular actions that the user will per- The most important limitation is that such form and take the identified tasks as the basis sources are one-sided. They report only for building the usability specifications. We problems and say nothing about the features then instantiate them to real-world examples that users liked or that enabled efficient use. and present them to test participants in a us- The most important thing about user ability test. analysis is to record, structure, and organize the findings. Usability benchmarks. We set usability benchmarks as quantitative usability goals, which Task analysis. Task analysis describes a set are defined before system design begins.10 of techniques people use to get things done.8 They are based on the five basic usability at- The concept of a task is analogous to the tributes or their subattributes. concept of a use case in object-oriented soft- We need these benchmarks because, if we ware development; a task is an activity want to assess the value of the usability at- meaningful to the user. User analysis is tributes for the system under development, taken as input for task analysis, and both we need to have a set of operationally de- are sometimes performed jointly. fined usability benchmarks. We analyze tasks because we can use the We establish usability benchmarks by located tasks to drive and test UI design defining a set of benchmarks for each usabil- throughout the product development cycle. ity attribute we want to evaluate—that is, for Focusing on a small set of tasks helps ra- each usability attribute we consider impor- tionalize the development effort. Therefore, tant for our system. We must define the we suggest prioritizing the set of tasks by benchmarks in a way that makes them calcu-

January/February 2001 IEEE SOFTWARE 25 Table 1 A Sample Usability Specification Table4 Usability Measuring Worst acceptable Best possible Observed attribute instrument Value to be measured Current level level Planned target level level results Performance “Answer Length of time taken 2 min, 53 sec 2 min, 53 sec 1 min, 30 sec 50 sec in normal use request” task to successfully perform the task (minutes and seconds) First Questionnaire Average score— 0 1 2 impression (range –2 to 2)

lable in a usability test or through a user sat- able to paint reasonable pictures if they use isfaction questionnaire. Table 1 shows the the principles they learned. Another way to format of a usability specification table. (The improve design ability is to examine UIs. “Observed results” column is filled with the Analyzing the UIs of every software appli- data gathered during the usability tests.) cation you can access is very helpful and can We take task analysis as an input for this sometimes be a source of inspiration for activity, because most usability benchmarks finding innovative, alternative solutions. are linked to a task specified in task analysis. The conceptual design phase also ends with evaluating the results. It is a good idea to Usability design test the paper prototypes against the defined Once we have analyzed the tasks our sys- task set to check that all the prioritized tasks tem will support, we can make a first at- can be enacted. The last test in this phase is tempt at the UI’s conceptual design, which run together with users as a usability test or we will evaluate and possibly improve in the usability inspection of the paper prototype. next iteration. Visual design. Having completed the concep- Conceptual design. During the conceptual tual design, the final step in our process is vi- design phase, we define the basic user–sys- sual design, where we define the UI’s appear- tem interaction and the objects in the UI ance. This covers all details, including the and the contexts in which interaction takes layout of screens and dialog boxes, use of place. The findings of the user and task colors and widgets, and design of graphics analysis are the basis for the conceptual de- and icons. There are also rules and principles sign. The deliverables from this phase are for visual design, addressing use of color, typically paper prototypes, such as pencil text, screen layout, widget use, icon design, drawings or screen mockups, and a specifi- and so forth. It pays to have a professional cation, which describes the UI’s behavior. screen designer, especially in this phase. Rec- Conceptual design is the most crucial ommended readings about visual and con- phase in the process, because it defines the ceptual design are About Face12 and Soft- foundation for the entire system. Unfortu- ware for Use,9 which both include numerous nately, design is a very creative process, and design tips. Designing Visual Interfaces fo- it can’t be automated with a method. There is cuses on screen design and graphics design in a set of design principles and rules that we the context of UIs, as well as the underlying must creatively adapt for a certain design principles of visual design.13 problem. (A good reading for any designer— The deliverables of this phase are proto- not just software designers—is The Design of types that must be tested, an exact specifi- Everyday Things,11 which presents general cation of the UI appearance, and behavior design principles by evaluating the design of plus the specification for new widgets that everyday objects.) must be developed. The main principles of UI design cover feedback, reuse, simplicity, structure, toler- Prototyping ance, and visibility in UIs. Knowing usabil- Prototypes are not exclusive to UI design, ity design principles is the basis for good de- but they are valuable for performing usabil- sign. Compare this to an adult drawing ity testing in early development phases. We class. Not everyone will be Picasso by the need to build prototypes because abstract end of the course, but the students will be technical specifications are not a good way of

26 IEEE SOFTWARE January/February 2001 communicating when we want to involve ability without testing it with real users. users in the design process—users understand First, we must decide which groups of tangible system prototypes much better.5 users we want to use to test the system and We can’t predict Some prototyping techniques help perform how many from each group we will try to usability testing and require little implemen- recruit as test participants. Then, we must a software tation effort. We create prototypes to test design the test tasks we’ll ask the partici- system’s them on the user through usability evaluation pants to perform. We usually take them usability techniques. The prototyping techniques with from the results of the task analysis activity which software developers usually are not fa- and apply them to hypothetical real-life sit- without testing miliar include uations. Some characteristics of the test re- it with real quire consideration, such as Paper mock-ups: At the beginning of the users. design process, the designer creates pa- whether the participant can ask the per prototypes—usually pencil drawings evaluator for help; or printouts of screen designs—for the should two participants jointly perform user. The designer will act as the com- each test task to observe the remarks puter, showing the user the next element they exchange in the process; when a transition between graphical el- what information participants will receive ements occurs.8 about the system prior to the test; and “Wizard of Oz” technique:8 A human whether to include a period of free sys- expert acts as the system and answers tem access after completing the prede- the user’s requests, without the user’s fined tasks to get the user’s overall im- knowledge. The user interacts normally pression of the system. with the screen, but instead of using software, a developer sits at another com- After we prepare the test and recruit test puter (network-connected to the user’s participants, we run the tests, optionally computer) answering the queries. The user recording them with video cameras or audio gets the impression of working with a recorders, and log the users’ actions in the real software system, and this method is system for further analysis (also optional). cheaper than implementing a real soft- Once we have performed all the tests, we ware prototype. analyze the data and gather results to apply Scenarios, storyboards, and snapshots: them in the next iterative cycle. A scenario describes a fictional story of a user interacting with the system in a par- Thinking aloud. Formative evaluation seeks ticular situation; snapshots are visual to learn which detailed aspects of the inter- images that capture the interaction oc- action are good and how to improve the in- curring in a scenario; and storyboards8 teraction design.5 This opposes summative are sequences of snapshots that focus on evaluation, which is performed at the end of the main actions in a possible situation. the development process, after the system They make the design team think about has been built. The results of summative the appropriateness of the design for a evaluation do not help shape the product. real context of use, and they help make Thinking aloud helps perform formative the process user-centric. evaluation in usability tests. We ask the test participant to think aloud while using the Usability evaluation system in a usability test,8 to verbalize his or Usability evaluation is a central activity her actions so we can collect the remarks. in the usability process. It can determine the For example, a participant might say, “First, current version’s usability level and whether I open the file, and I click once on the file the design works. icon. Nothing happens. I don’t know why this is not working like the Web. I press the Usability testing. The term usability testing Enter key, and it opens. Now I want to describes the activity of performing usability change the color of the label, so I search in tests in a laboratory with a group of users the Tools menu, but I can’t find any option and recording the results for further analy- for what I want to do.” User remarks ob- sis. We can’t predict a software system’s us- tained in usability tests can provide signifi-

January/February 2001 IEEE SOFTWARE 27 Further Reading

The following are books about human–computer interaction (HCI) and usability that are more likely to interest software practitioners with little or no knowledge about the field (the information for each book is in the Refer- usually more applicable, and they are more ences section of this article). precise about the underlying usability problems, such as a lack of consistency or poor B. Shneiderman, Designing the User Interface: Strategies for Effective Hu- navigation. On the other hand, usability man–Computer Interaction—This book summarizes all aspects related to testing must be performed with real users to interactive systems from a serious scientific viewpoint, although some identify specific usability problems. Heuris- readers might prefer a more engineering-focused approach. It includes a tic evaluation can complement but not re- valuable set of guidelines for designing the user interface. There have place usability testing. been some interesting additions in the third edition about issues such as hypermedia and the Web and Computer Supported Cooperative Work. Collaborative usability inspection. A collaborative usability inspection is a systematic ex- D. Hix and H.R. Haertsen, Developing User Interfaces: Ensuring Usability amination of a finished system, design or 9 Through Product and Process—Despite its title, this book is not just about prototype from the end user’s viewpoint. A the user interface; it focuses on the process of user interaction design. Writ- team of developers, end users, application ten in a very practical style, it provides a hands-on approach to designing or domain experts, and usability specialists the interactive part of a software system. Software practitioners might want collaboratively perform the review. Collab- to skip the chapters devoted to the User Action Notation—a technique for orative usability inspections (CUIs) use fea- representing interaction designs—which is too formal for non-HCI experts. tures and techniques from heuristic evaluation, pluralistic usability walkthroughs, and L.L. Constantine and L.A.D. Lockwood, Software for Use: A Practical Guide expert evaluations and are less expensive to the Models and Methods of Usage-Centered Design—The most recent of and faster than usability testing. Behind this the books reviewed here, it presents a process for designing usable soft- technique is a set of strict rules to avoid the ware systems based on one of the current trends in software engineering: problems that typically arise if end users dis- use cases. The book is written in a practical style that is likely to appeal to cuss their work together with designers or software practitioners. developers. CUIs uncover more—albeit different—usability defects (up to 100 defects J. Nielsen, Usability Engineering—This book provides a good introduc- per hour) than usability testing. tion to the issue of usability engineering. It is easy to read and includes Apart from efficiency, one advantage is stories of real situations. It deals with a wide variety of issues related to that people with multiple perspectives and usability engineering—but none are addressed in depth. expertise examine the test object. Another advantage is that the participating developers build skills and know-how about how to make software more usable. cant insight into the best way of designing the system interaction. By detailing their Management and organizational mental process, test participants can un- issues cover hidden usability problems. When introducing usability, an organiza- Formative evaluation is the usual form of tion must first commit management to the evaluation in a usability process, combining ideas behind the usability process and con- qualitative data gathered from user com- vince them of its benefits.4,5 The newest ments with quantitative data to check against concepts they need to accept include creat- previously defined usability benchmarks. ing conceptual design in the first stages of development and evaluating usability Heuristic evaluation. A usability expert can throughout the development process. Cost- perform a heuristic evaluation of the system Justifying Usability presents cost-benefit ar- to make some development iterations guments in favor of performing usability shorter and to perform more iterations in the practices, which can be used when trying to development process. The expert will make get management commitment.15 Another a critique founded on both his or her inter- option to convince management is to take a action design experience and on generally recently developed system or one that is cur- accepted usability guidelines, like the ones rently being developed and to perform by Ben Shneiderman14 and Jakob Nielsen.5 videotaped usability tests with a few users Experts provide a different kind of feed- who are novel to the system. Showing the back than final users through usability test- results to management and the development ing. Expert suggestions for modification are team can produce a change of attitude to-

28 IEEE SOFTWARE January/February 2001 ward usability testing, as the results will man Factors in Computing Systems, ACM Press, New probably show that the system is not as York, 1986, pp. 241–246. 4. D. Hix and H.R. Hartson, Developing User Interfaces: good in usability terms as expected. Ensuring Usability Through Product and Process, John Integrating UI designers into the develop- Wiley & Sons, New York, 1993. ment team isn’t always easy, especially if 5. J. Nielsen, Usability Engineering, AP Professional, Boston, Mass., 1993. they are assigned to several projects at 6. H. Beyer and K. Holtzblatt, Contextual Design: A Cus- the same time. One approach to applying tomer-Centered Approach to Systems Design, Morgan usability techniques in some projects is Kaufmann, San Francisco, 1997. 7. D.A. Dillman, Mail and Internet Surveys: The Tailored to promote one member of each develop- Design Method, John Wiley & Sons, New York, 1999. ment team to usability champion,5 similar to 8. J. Preece et al., Human-Computer Interaction, Addison- process improvement champions. Usability Wesley Longman, Reading, Mass., 1994. 9. L.L. Constantine and L.A.D. Lockwood, Software for champions learn the basic usability skills Use: A Practical Guide to the Models and Methods of and are coordinated by a user interaction Usage-Centered Design, Addison-Wesley Longman, Reading, Mass., 1999. designer. The user interaction designer then 10. J. Whiteside, J. Bennett, and K. Holtzblatt, “Usability acts as a consultant in several projects but Engineering: Our Experience and Evolution,” Hand- can interact with the usability champion in book of Human-Computer Interaction, Elsevier North- Holland, Amsterdam, 1988. 4 each group. 11. D.A. Norman, The Design of Everyday Things, Dou- Don’t try to do a full-scale usability process bleday, New York, 1990. from the beginning. You can start by setting 12. A. Cooper, About Face: The Essentials of User Interface Design, IDG Books Worldwide, Foster City, Calif., 1995. a small set of usability specifications with a 13. K. Mullet and D. Sano, Designing Visual Interfaces: simple task analysis of the most prominent Communication Oriented Techniques, Prentice Hall, tasks, some conceptual design with paper Upper Saddle River, N.J., 1994. 14. B. Shneiderman, Designing the User Interface: Strate- prototypes and simple usability tests to be gies for Effective Human-Computer Interaction, Addi- carried out with a small set of users. You son-Wesley Longman, Reading, Mass., 1998. 15. R.G. Bias and D.J. Mayhew, Cost-Justifying Usability, can also act as usability expert performing Academic Press, Boston, Mass., 1994. heuristic evaluation on the system using the guidelines we mentioned (by Shneiderman14 and Nielsen5). Starting with modest objec- About the Authors tives will contribute more firmly to the final success of your endeavor. Xavier Ferré is an assistant professor of software engineering at the Universidad Politéc- nica de Madrid, Spain. His primary research interest is the integration of usability techniques into software engineering development practices. He has been a visiting PhD student at CERN (European Laboratory for Particle Physics) and at the HCIL (Human–Computer Interaction Lab- espite increasing usability awareness oratory) at the University of Maryland. He received an MS in computer science from the Uni- in software development organiza- versidad Politécnica de Madrid. He is a member of the ACM and its SIGCHI group. Contact him at [email protected]. D tions, applying usability techniques in software development is not easy. Soft- Natalia Juristo is a full professor in ware engineers and usability engineers have the Computer Science Department at the Universidad Politécnica de Madrid, where she directs master’s-level courses in knowledge engineering and software engineering. She is also an edi- a different conception of software develop- torial board member of IEEE Software and the International Journal on Software Engineering ment, and conflicts can arise between them and Knowledge Engineering. She has a BS and PhD in computer science from the Technical due to differences in terminology and pro- University of Madrid. She is a senior member of the IEEE Computer Society and a member of the ACM, the American Association for the Advancement of Science, and the New York Acad- cedures. To create acceptable usability con- emy of Sciences. Contact her at the Facultad de Informática UPM, Campus de Montegancedo, cepts, the software engineering community s/n, Boadilla del Monte, 28660 Madrid, Spain; [email protected]. must integrate usability techniques into a Helmut Windl leads the User Interface Design Group for Simatic Automation Software at software engineering process that is recog- Siemens’ Automation & Drives Division, where he has helped define and implement a struc- nizable from both fields. Use cases offer a tured usability process within the software development process. He is an experienced user- good starting point, as they are the software interface and visual designer for large-scale software applications and a project leader for usability-focused products. He is also a trainer and presenter in Siemens AG and with engineering construct closer to a usable soft- Constantine & Lockwood. He received a diploma in electrical engineering from the University ware development approach. of Applied Sciences Regensburg. Contact him at Siemens AG, A&D AS S8, PO Box 4848, D-90327 Nuremberg, Germany; [email protected].

Larry Constantine is director of References research and development at Constantine & Lockwood, a training and consulting firm. He is 1. L. Trenner and J. Bawa, The Politics of Usability, also an adjunct professor in the School of Computing Sciences at the University of Technology, Springer-Verlag, London, 1998. Sydney, where he teaches software engineering and managing organizational change, and he 2. Ergonomic Requirements for Office Work with Visual is on the faculty of the Cutter Consortium. He has authored or coauthored 10 books, including Display Terminals, ISO 9241-11, ISO, Geneva, 1998. Software for Use: A Practical Guide to the Methods and Models of Usage-Centered Design (Ad- 3. M. Good et al., “User-Derived Impact Analysis as a dison Wesley Longman, 1999). Contact him at Constantine & Lockwood, 58 Kathleen Circle, Tool for Usability Engineering,” Proc. CHI Conf. Hu- Rowley, MA 01969; [email protected]; www.foruse.com.

January/February 2001 IEEE SOFTWARE 29