The Origins of Errors in Rule-Based Programming Tasks

如果说“人非圣贤，孰能无过”，那么理解错误的起源、本质、检测和纠正应该是从事人机交互研究的人的一个重要目标。 我们可能犯错误的方式有很多;目前的建议解决了最温和的，研究最少，但最普遍的人为错误类型之一：在执行简单的，基于规则的任务时的错误。基于规则的任务包括诸如从ATM取钱、对VCR编程以录制节目或使用浏览器上网等任务。这样的任务涉及遵循有限的一组规则（即， 良好描述的程序），以实现有限的，但明确定义的目标。它们不涉及解决问题或创造力。 虽然它们对人类认知的要求似乎很低，但即使是有经验的用户，这些任务中的错误发生率似乎也很高。越来越多的此类任务在家里（例如，可编程语音邮件、恒温器和自行车里程表）和工作（例如，先进商用飞机上的飞行管理计算机，电子表格或数据库的日常使用， 和 电子邮件系统）表明， 一个 深入了解这些任务中的错误可以显著改进许多简单但重要的交互式工具的设计。

If "to err is human," then understanding the origin, nature, detection, and correction of errors should be an important goal for those engaged in the study of human-computer interaction. There are many ways in which we may err; the current proposal tackles one of the most modest, least studied, but most pervasive types of human errors: errors in performing simple, rule-based tasks. Rule-based tasks include tasks such as getting money from an ATM, programming a VCR to record a show, or using a browser to surf the web. Such tasks involve following a limited set of rules (i.e., well described procedures) to achieve limited, but well defined goals. They do not involve problem solving or creativity. Though seemingly modest in terms of their demands upon human cognition, the prevalence of errors in such tasks seems constant and high even with experienced users. The increasing number of such tasks at home (e.g., programmable voice-mail, thermostats, and bike odometers) and work (e.g., flight management computer on advanced commercial aircraft, routine use of spreadsheets or databases, and email systems) suggests that an indepth understanding of errors in these tasks could significantly improve the design of many simple but important interactive tools.