No abstract

For decades, automatic test case generators driven by context-free grammars usually have adopted a derivation method designed to obtain the minimum set of shortest sentences.<div>Within this paper, a <i>pure weight-based</i> approach is described: such approach uniquely uses <i>weights </i>to select a rule while expanding a nonterminal symbol; we will assess how it is always possible to identify <i>balanced </i>weights which ensure the convergence of the process and we will see how the generated sentences are usually more complex than the shortest ones, and having additional chances to exercise the underlying system more extensively as confirmed by experimental results.</div>

For decades, automatic test case generators for context-free grammars usually have adopted a derivation method designed to obtain the minimum set of shortest sentences.<div>Within this paper, a <i>pure weight-based</i> approach is described: such approach uniquely uses <i>weights </i>to select a rule while expanding a nonterminal symbol; we will assess how it is always possible to identify <i>balanced </i>weights which ensure the convergence of the process and we will see how the generated sentences are usually more complex than the shortest ones, so having more chances to exercise the underlying system more extensively.</div>

For decades, automatic test case generators for context-free grammars usually have adopted a derivation method designed to obtain the minimum set of shortest sentences.<div>Within this paper, a <i>pure weight-based</i> approach is described: such approach uniquely uses <i>weights </i>to select a rule while expanding a nonterminal symbol; we will assess how it is always possible to identify <i>balanced </i>weights which ensure the convergence of the process and we will see how the generated sentences are usually more complex than the shortest ones, so having more chances to exercise the underlying system more extensively.</div>