Towards query formulation and visualization of structural search results
Oleksandr Panchenko, Arian Treffer, Alexander Zeier · 2010
Source code search goes far beyond simple textual search. One possibility of improving code search is the utilization of structural information in form of abstract syntax trees (ASTs). However, developers usually work with the textual representation of source code and, thus, have difficulties in expressing their queries as fragments of abstract syntax trees and in interpreting the results. This paper addresses assistance of query composition and search result visualization. Query formulation is considered to be an iterative process. After one query is run, the AST vertices neighbored to the result vertices are analyzed to propose refinement options for the next query. Search results are visualized in a tree view which aggregates all matches in a compact way instead of showing a small number of ranked matches.