previous_group previous up next next_group
Previous: 7. A Sixth Issue: Up: 1. The Need for Next: 2. HQ Interface

8. Conclusion

We have presented our motivations for a common generic interface, which could be used in static analyzers. Then, we have briefly compared six interfaces. We did not consider the interface of the interval and BDDs domains here, since those two domains are used in a different way.

As we have seen, there exist many problems towards a common interface, even if we study only the polyhedral domain with its existing polyhedra implementations (4_subsec:polyhedra_implementation), or with its closest domains, the octagons (4_subsec:octagons_polyhedra) and the Presburger formulae (4_subsec:presburger_polyhedra). These problems are:

In the next section, we present our approach to analyze in greater details and then to deal with these problems. The Parma library PPL also proposed a general-purpose interface and implemented it. However, in our point of view, it is not satisfying (see our related work discussion in 4_subsec:related_work:ppl). Therefore, the construction of a prototype of a new interface seems to be required if we want to move forward toward our goal. APRON, another approach with a prototype developed by Bertrand JEANNET, which is also presented in our related work, 4_sec:related_work, is based on our approach.


previous_group previous up next next_group
Previous: 7. A Sixth Issue: Up: 1. The Need for Next: 2. HQ Interface
Nguyen Que Duong
2006-09-16