Users of software make many mistakes, and correcting these mistakes may be time-consuming. Research shows that users may spend between 25% and 50% of their time correcting errors. Helping users deal with mistakes is therefore important. Reducing mistakes and streamlining error detection, diagnosis and recovery means more than saving time. It can also substantially reduce the frustration and anxiety of learning about and using computers. In addition, it can make such learning and use more productive.
The design strategies that support error recognition and recovery are:
- Heuristic 3.1: Prevent mistakes whenever possible
- Heuristic 3.2: Provide error information when actions are error-prone or when correction is difficult
- Heuristic 3.3: Provide error information that supports detection, diagnosis and recovery
- Heuristic 3.4: Provide on-the-spot error information
The best place to prevent errors is in the application itself. There is no substitute for an application that has been tested thoroughly and its inadequacies treated as a measure to prevent users from making mistakes. But no matter how well-designed, errors undoubtedly arise and designers should give these due attention.
An effective way to prevent users from committing mistakes is to block certain options in the initial stages when users are getting acquainted with an application. In such a Training Wheels approach the user cannot access all options on the interface but he or she can still see these. Blocked options can be grayed to signal that they cannot be activated. Typically one would block options that: (a) lead to mistakes that are difficult or time-consuming to correct and (b) options that can seriously hamper the proper use of the program or destroy (parts of) a database. It is important to let users practice with the regular interface including the blocked options. That way they learn to see which menu option belongs where.
There are different ways to prevent users from making mistakes by creating good documentation. Following general and preferably empirically validated guidelines is one of these ways. For example, writers are advised to provide a safe trajectory of activities, use short and simple sentences, signal action-information, minimize jargon, and so on. Mistakes can also sometimes be prevented by including hints in the manual. As long as the manual is not crowded with hints, they are likely to be attended to and serve both as a means to prevent error and as information that users may turn to later, after having made a mistake.
Heuristic 3.2: Provide error information when actions are error-prone or when correction is difficult
Error information is always needed for actions that users are likely to misunderstand. The difficult question here is how one can discover these actions. The best approach is usability testing. Nothing beats testing to discover the hot spots in a manual; there is no real substitute for a well-designed usability test. Research shows that even experienced designers generally do not do very well in predicting the problems in documentation. Designers should therefore always take the opportunity to test their documentation. Only a few users from the main audience can suffice to detect the major problems. Preferably such testing takes place in the early phases of document development.
Usability testing is an empirical approach. For some situations one can be almost sure without testing that they lead to mistakes. One such moment is when the user’s standard routine for handling an application does not work. Another error-prone moment is when the users’ real-world experience conflicts with the way in which the application should be operated. An example is the user who forgets to select a text block before underlining it. Users forget to select information in advance to acting on it because the procedure does not match with how they work on paper. Selecting a bit of information before manipulating it is not part of the standard way of handling this matter on paper.
A stagewise model for problem-solving suggests the following stages: seeing the problem, expressing the problem and processing the problem-solving information. As error handling is a form of problem solving, these stages also apply and giving error support can be modeled accordingly. Error-information should thus help the user in detecting, diagnosing and correcting mistakes.
Information about detection and correction are always necessary. Diagnostic information is optional in the sense that users may not always need it to solve their problem. Designers do well to include this information too, however, because this is the moment in which users tend to be highly motivated to learn more about an application. Errors provide an opportunity to offer valuable conceptual information that users are likely to read when they need it.
Figure 1 presents the kind of error information presented in some of the earlier studies (see also Guided Exploration cards). The information has its own presentation format to distinguishes it from other information types. In addition, there is information to help the user in detecting, diagnosing, and correcting a mistake.
Figure 1. Error information as presented in a minimal manual about WordPerfect.
In recent work, shown in Figure 2, three aspects of the original design have been changed. One. a first-aid icon is added to enhance the distinctive nature of the help. Two, the constituent elements of the problem-solving information have been visibly separated to better facilitate scanning. Users can now skip more easily what they do not want to read. Three, the three moments in problem-solving are mentioned explicitly. This too facilitates scanning, but more importantly, it also models problem decomposition. By showing users which steps to take in handling mistakes they may be inclined to handle their own problem solving in the same fashion.
Figure 2. Error information as presented in a minimal manual about WWW-searching.
Probably one of the most underrated issues in user documentation is the provision of timely error information. In nearly all applications, errors can be a nightmare to restore when the user is not attended to their presence quickly enough. Error information should therefore be presented as near as possible to the source of an error-prone action or method.
Giving error information on the right spot is important to help the user catch mistakes before they can lead to yet other mistakes and before the user executes other actions which make it difficult to track down what went wrong. A timely provision of error information also facilitates the task of the designer who does not have to worry about the usual problem of presenting meaningful contextual information. When, as is the more common approach, all error information is placed in a separate section, it is very difficult to describe the problem state in such a way that the information is both easily accessible and readily understood by the user.
When error information is positioned close to where things may go wrong, users also tend to explore error correction out of curiosity. Even when they have correctly executed a particular task, they sometimes redo the task and make a mistake on purpose. Research indicates that this exploratory usage is neither incidental nor unusual. In minimal manuals about 15% of all sections with error information are used in this fashion. Because it supports the development of a deeper understanding of the program, this can be an important side-effect of a timely presentation of error information.
The minimalist approach to supporting error recognition and recovery is a unique point of view that distinguishes it from other instructional design theories. It is still quite typical for these theories to ignore errors as a central phenomenon of using information, to implicitly assume that errors do not occur. As indicated earlier, the facts go another way: error detection, diagnosis, and recovery can consume somewhere between a quarter and half of the user’s time.
Commentators have suggested that minimalism over-emphasizes error and error support, making a fuss about something that “most of us would endorse as good, old-fashioned, common sense”. Indeed some even state that “it is fairly common for manuals to anticipate user’s errors and to provide correctives for errors”. This is wishful thinking. Most conventional manuals do not provide adequate support for handling mistakes. A survey on the presence and design of error information in conventional manuals concludes that most manuals offer way too little error-information. In addition, users had to surmount formidable obstacles to accessing that information.
Minimalism provides lots of help for dealing with mistakes. In one document error information appeared with a frequency of about once for every three user actions. Too much of a good thing, you may ask? Perhaps. Thanks to its presentation format the users could easily ignore the error-information. In addition, tests conducted after training showed that the help boosted the user’s performance in a statistically significant way.
It would be wrong to equate minimalism with an abundance of error-information, however. It promotes various ways of fencing off users from making mistakes that are difficult, distracting, and not meaningful to recover from. Minimalism also does not advocate an approach that concentrates on helping users deal with error as that does not fit the primary goal for which the user turns to the documentation. In short, minimalism is a complete view of how to deal with potential user mistakes in an application and in the documentation that accompanies it.
- Carroll, J.M., & Van der Meij, H. (1998). Ten misconceptions about minimalism. In J.M. Carroll (Ed.), Minimalism beyond the Nurnberg funnel (pp. 55 - 90). Cambridge, Mass: MIT Press.
- Lazonder, A. W. (1994). Minimalist computer documentation. A study on constructive and corrective skills development. Doctoral Dissertation. Twente University, Enschede.
- Lazonder, A.W., & Van der Meij, H. (1994). The effect of error-information in tutorial documentation. Interacting with Computers, 6(1), 23-40.
- Lazonder, A.W., & Van der Meij, H. (1995). Error-information in tutorial documentation: supporting users' errors to facilitate initial skill learning. International Journal of Human Computer Studies, 42, 185-206.
- Van der Meij, H. (1996). Does the manual help? An examination of the problem solving support offered by manuals. IEEE Transactions on Professional Communication, 39(3), 146-156
- Van der Meij, H. & Carroll, J.M. (1998). Principles and heuristics for designing minimalist instruction. In J.M. Carroll (Ed.), Minimalism beyond the Nurnberg funnel (pp. 19 -53). Cambridge, Mass: MIT Press.