With all but one of the tutorials mostly finished, I should go ahead and announce the existence of my automata library, a dynamic state machine framework which you can either use to construct DFAs and DPDAs or extend to build your own types of automata. If you haven't done so already, you can download it from http://boost-consulting.com/vault/ (automata.zip is in the Data Structures directory). If you have an older version, here are some changes: * The automaton<> front-end class template used to require only one template parameter: a Transition Function. Now, it also needs a Unary Metafunction Class that positionally comes first and whose return type is an Automaton Base. The decoupling was done to allow for cases like a user-defined type instead of dfa_generator working together with a matrix_transition_function<>, etc. * Visitors/Traits are now Observers. * Input-Check Visitors are now Input Validators. * Renamed *_ostream_policy<> class templates to *_insertion_policy<> class templates, to allow for the use of a Boost.Serialization archive as an Insertion Target instead of an output stream. * Renamed *_check_policy<> class templates to *_validation_policy<> class templates. * builder_from_graph<> now lets you attach Observer Builders to it. * Added an option to enable undo functionality. * Removed unnecessary type definitions. Concept and reference documentation still need tons of work, but I think the code structure is stable enough for preliminary testing. Feedback is much appreciated, especially in the following areas: * Tutorials. Are they comprehensive enough? What other potential use cases deserve attention? * Interface. Is it too restrictive? Is it too big? To all of you who've been testing this library at its various stages, thank you for your continued interest! Cromwell D. Enage __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com