No matter how much code-generation automation goes into CASE tools, sometime during software development, a real, live, "flesh and blood" engineer will need to edit code. Furthermore, how much useable information can be learned from an automatically generated reverse-engineering design graph? When presented with such a cluttered diagram, an engineer will pick out one or two of the modules of interest and revert to command line tools and editors to read the code directly, trying to note important design features or "what function calls what" on a neighboring sheet of paper...

collaGra's ModuleGrapher tool enables Semi-Automatic Graph Generation. This approach requires the involvement of the engineer in the generation of the graph. This insures that only code elements relevant to a particular use case or bug fix design are included in a particular design graph. This involvement also insures that the engineer builds his or her own knowledge of the system; after all, the engineer will be changing the code.