more on readme

Author: Andrei Popescu

README.md

@@ -128,7 +128,6 @@ Most of our examples and case studies consist of three distinct types of theorie
    An exception to the rule of using `binding_datatype` is the (non-recursive) datatype of commitments for the pi-calculus (described in Sect. 7.1), for which we use some Isabelle/ML tactics to the same effect in thys/Pi_Calculus/Commitments.thy (the reason being that our parser does not yet cover the degenerate case of non-recursive binders). 
 
 (2) Those introducing the relevant binding-aware inductive predicates, usually via our `binder_inductive` command described in Sect. 9 and App. G.2) -- the exceptions being the instances of the binder-explicit Thm. 22, where we instantiate the locale manually. In particular, we have:
-
     * In thys/Untyped_Lambda_Calculus, the theories LC_Beta.thy and LC_Parallel_Beta.thy, containing the inductive definitions of lambda-calculus beta-reduction and parallel beta-reduction respectively, referred to in Sects. 2 and 5. In particular, Prop. 2 from the paper (in the enhanced version described in Remark 8) is generated and proved via the `binder_inductive` command from LC_Beta.thy; it is called `step.strong_induct`. The corresponding theorem for parallel-beta is called `pstep.strong_induct`, which is generated and proved from the `binder-inductive` command from LC_Parallel_Beta.thy. A variant of parallel-beta decorated with the counting of the number applicative redexes (which is needed in the Mazza case study) is also defined in LG_Beta-depth.thy (and its strong rule induction follows the same course).
     * In thys/Pi_Calculus, the theories Pi_Transition_Early.thy and Pi_Transition_Late.thy use the `binder-inductive` command to define and endow with strong rule induction the late and early transition relations discussed in Sect. 7.1; and the theory Pi_cong.thy does the same for both the structural-congruence and the transition relations for the variant of pi-calculus discussed in App. B. 
     * In thys/POPLmark, the theory SystemFSub.thy is dedicated to defining (in addition to some auxiliary concepts such as well-formedness of contexts) the typing relation for System-F-with-subtyping discussed in Sect. 7.2. Here, because (as discussed in Sects. 7.2 and 7.3) we want to make use of an inductively proved lemma before we prove Refreshability (a prerequisite for enabling strong rule induction), we make use of a more flexible version of `binding_inductive`: namely we introduce the typing relation as a standard inductive definition (using Isabelle's `inductive` command), then prove the lemma that we need, and at the end we "make" this predicate into a binder-aware inductive predicate (via our command `make_binder_inductive`), generating the strong induction theorem, here named `ty.strong_induct` (since the typing predicate is called `ty`). Note that, in general, a `binder_inductive` command is equivalent to an `inductive` command followed immediately by a `make_binder_inductive` command. We have implemented this finer-granularity `make_binder_inductive` command after the submission, so it is not yet documented in the paper. (In the previous version of the supplementary material we had a different (less convenient) solution, which inlined everything that needed to be proved as goals produced by `binder_inductive`.)
@@ -150,7 +149,7 @@ As discussed in Sect. 9 and App. G, we have automated the production of binding-
    * todo: the tactic for heuristic too. 
 
 
-### Mapping of the results from the the main paper to Isabelle theorem names
+### Mapping of the results from the main paper to Isabelle theorem names
 
 Prop 1 --> subsumed by Prop. 2 (also generated and proved automatically by the standard inductive definition)