Increasingly focus in the software comprehension community is shifting from representing the results of analysis in the graph and database domain to reflecting insights directly into source. The obvious modern representation for this reflection is XML markup. In the simplest case, XML markup of ract syntax tree itself can be represented in source, although the result is wordy, overly detailed and cumbersome to deal with. A more realistic solution is to use island or multiweight parsing to mark up the AST in only those sections of source of interest to the current task. In this paper we outline a method for extending and generalizing the partial markup idea to minimize source markup not only by marking only sections, but by selectively marking up the source with only a subset of the AST nodes relevant to each particular task as well. By exploiting agile parsing, this idea is further extended to allow for task-directed selective markup as a natural extension of selective syntactic marku...
James R. Cordy