The Transportable Applications Environment Plus (TAE Plus) is a NASA-developed user interface development environment (UIDE) for the rapid prototyping, evaluation, implementation, and management of user interfaces. TAE Plus provides an intuitive What You See Is What You Get (WYSIWYG) WorkBench for designing an application's user interface. The WorkBench supports the creation and sequencing of displays, including real-time, data-driven display objects. Users can define context-sensitive help for a target application. They can rehearse the user interface and also generate code automatically. In addition TAE Plus contains application services for the runtime manipulation and management of the user interface. Based on Motif and the MIT X Window System, TAE Plus runs on a variety of Unix- or VMS-based workstations. TAE Plus is an evolving system. User-defined requirements and new technology guide the development of each new version. Advances in virtual operating systems, human factors, computer graphics, command language design, standardization, and software portability are monitored and incorporated as they become available.

The success of any software development project depends in part on the quality of the communication among the individuals involved: users, designers, coders and managers. On large systems, a variety of individuals perform various tasks at different points in time. The efficiency and correctness with which later tasks are performed depends critically on the documentation supplied during previous phases of the development cycle. This paper describes a series of three experiments designed to examine the effects of documentation format on the performance of programmers on different software-related tasks.

Three software complexity measures (Halstead's E, McCabe's u(G), and the length as measured by number of statements) were compared to programmer performance on two software maintenance tasks. In an experiment on understanding, length and u(G) correlated with the percent of statements correctly recalled. In an experiment on modification, most significant correlations were obtained with metrics computed on modified rather than unmodified code. All three metrics correlated with both the accuracy of the modification and the time to completion. Relationships in both experiments occurred primarily in unstructured rather than structured code, and in code with no comments. The metrics were also most predictive of performance for less experienced programmers. Thus, these metrics appear to assess psychological complexity primarily where programming practices do not provide assistance in understanding the code.