Internal Strategies in a Reeective Logic Strategies and Rewriting in Elan Tacticals in Ilf -and Why They Are Superruous Bernd Ingo Dahn 1.40pm -2.10pm Break Session 2: Strategy Analysis 2.10pm -2.40pm Strategies for Logic Programming Languages

i Preface The concept of strategy allows describing and guiding computations and deductions in automated theorem provers, proof checkers, logical frameworks. Strategies are used for various purposes, e.g. for-proof search in theorem proving,-combining diierent proof techniques and computation paradigms,-program transformation,-developing heuristics for playing games and nding proofs. Rules are not suucient to capture every proof development. A mechanism is needed for instance to formalise the search for diierent solutions, the check of context conditions, the request for user input to instantiate variables, the processing of subgoals in a particular order. Strategies are used to guide the application of rules, but may also involve iteration, case analysis, deterministic and non-deterministic choices. One may want to program strategies, to transform them, to prove certain properties on the computations or the proofs that they describe. The workshop aims at gathering diierent experiences on the use of strategies, under various terminology (tactic, tactical, method, heuristic, proof planning...), and in various application domains ((rst-order / higher-order / inductive theorem proving, program transformation, operational semantics...). Based on these experiments, several points are addressed: { Strategy languages: Which basic constructs are needed? Which are relevant choices: expressive-ness versus eeciency, metalanguage versus reeexivity, higher-order versus rst-order syntax? { Strategy analysis: Which properties are required for strategies? How to automate as much as possible the design of strategies in some logical fragments? How to evaluate applicability and performance? { Strategies in theorem proving: Which kind of architecture, modular design is suitable for such systems? How to deal with user-interaction, input-output? How to reuse past proof experience for search-guiding heuristics? The contributions included in these proceedings deal with several aspects of these topics. Papers have been available before the workshop on the web, to encourage reactions and discussions at the workshop. We thank the CADE-14 Program Committee and the Local Arrangements Committee for giving us the opportunity to hold this workshop. We also thank the participants for presenting and discussing their experiences and points of view on strategies in automated deduction.