Simplifying data integration: the design of the Desert software development environment

Abstract

Abstract

An Exploratory Study of How Developers Seek, Relate, and Collect Relevant Information during Software Maintenance Tasks

“…For example, some tools have used a virtual file, which allows developers to combine a set of line ranges [41] or methods [9] and edit it as if it were a single file. Robillard proposed the concept of a concern graph [43], which represents information as a set of relationships between program elements (such as declares and subclasses relationships).…”

“…Tools proposed in the past have tended to be heavyweight. For example, Desert [41] requires users to know the line numbers beforehand and enter them manually. FEAT [42] requires users to navigate a program by its relationships and add "relations" to a concern graph through a contextual menu, which is a somewhat oblique way for a developer to say "this code is important to me."…”

“…Desert [41] presents relevant lines of code as a single integrated file; FEAT [42] represents concern graphs as a hierarchical tree and requires developers to select an element in this hierarchy, and use a pop-up menu to request the source file; Concern Highlight [36] highlights relevance code in a conventional editor. There are several problems with these approaches: 1) they treat all information as if it had the same role in the task, 2) they do not support side-by-side comparison of information, and 3) they incur much of the same interactive overhead we observed in relying on file tabs and scroll bars.…”

“…A number of important contributions have been built around this concept, including ways of representing task contexts [41], [43], tools that enable developers to manually build a task context by selecting program elements [42], and methods of automatically inferring the relevant task context based on a developer's investigations in a development environment [44], [46].…”

Leveraging visio for adoption-centric reverse engineering tools

“…(The graph in Figure 5 took about 10 minutes to lay out on a reasonably equipped machine, but we have not yet made detailed performance measurements or attempted to optimize the operations.) Figure 6 s h is to craft software engineering functionality on top of COTS products that offer end-user pro y coupled components [9]. One of these compo are artifacts [10 nt in implementing a software visualization an class, and its value may therefore exceed the CBO value [8].…”