Towards Specifying Reengineering Services Using Graph-based Schemas

GUPRO (Generic Understanding of Programs) is a generic tool for supporting querying and browsing of software source code. An overview of GUPRO, as well as hints for further reading, can be found in [EKRW02]. GUPRO is generic in the sense, that heterogenous software written in different languages can be treated on arbitrary levels of abstraction. The key concept to support this genericity is the use of conceptual schemas to define the appropriate conceptual view on sofware. To use GUPRO, appropriate extraction tools (e.g. light-weight parsers) must be supplied, which extract the facts from a given software into the graph-based GUPRO repository according to the respective schema. These facts are analyzed by graph algorithms and GReQL (Graph Repository Query Language) graph queries. Using appropriate extractors, the repository content can also be viewed by the generic GUPRO browser, which allows multidimensional browsing through the sources. Browsing is also guided by the given schema. Furthermore, to handle C or C++ source code, preprocessor facilities are included into the browser. This facility allows browsing on any level of detail arbitrarily chosen between perprocessor input and preprocessor output. The generic browser and the generic query component are fully integrated with each other. Thus, a selection of code in the browser can be used as a starting point for query. The output of a query can be visualized as a table containing hyperlinks or as a selection in the browser window. GUPRO heavily relies on querying the graph-based repository as its enabling technology. Also, a script language clg (Command Line GReQL) is supplied which allows to execute stored query suits for standard applications. 1