Re: [Fwd: "VIS Optimization Question"]

From: Fabio Somenzi (Fabio@colorado.edu)
Date: Tue May 28 2002 - 10:23:40 MDT

  • Next message: Fabio Somenzi: "Re: Needing for more information about counter examples in approximate model checking"

    >>>>> "HW" == Hartmut Wittke <Hartmut.Wittke@Informatik.Uni-Oldenburg.DE> writes:

     HW> Hi Fabio,

     HW> In your answer to the mail of Hadi Afshar you wrote the next
     HW> version of VIS will also support LTL model checking. This sounds
     HW> very interesting to us. How will this be realized? Are you
     HW> planning to re-use the language emptyness check for this feature
     HW> with an externally generated buechi-automaton or will VIS
     HW> directly accept LTL formulas as input?

    VIS 2.0 reads LTL formulae directly, translates them internally into
    Buechi automata, and then uses a slightly modified version of the
    language emptiness check already in VIS 1.4 to verify the property.
    VIS 2.0 will also have bounded model checking for a subset of LTL
    (safety properties). For that, VIS interfaces to zchaff.

     HW> Have you thought about mixing LTL for assumptions with CTL for
     HW> commitments?

    The new LTL parser is in effect a CTL* parser, but CTL* model checking
    is not part of VIS 2.0. With CTL* you'll be able to do what you
    describe, if my understanding is correct.

     HW> If there are concrete plannings, when will this version of VIS be
     HW> available?

    We are in the testing phase. Release is expected in a few weeks.

     HW> If you use Buechi automata in either way, what will be the basis
     HW> for the tableau construction? The ordinary way or/and the
     HW> generation you suggested in the CAV 2000 paper?

    Right now (i.e., for 2.0), the algorithm is a simplified version of
    the one in the CAV 2000 paper.

    Fabio



    This archive was generated by hypermail 2b30 : Tue May 28 2002 - 10:42:01 MDT