- advances towards strong normalization

[helm.git] / matita / matita / contribs / lambdadelta / basic_2 / rt_computation / csx_lfpx.ma

diff --git a/matita/matita/contribs/lambdadelta/basic_2/rt_computation/csx_lfpx.ma b/matita/matita/contribs/lambdadelta/basic_2/rt_computation/csx_lfpx.ma
new file mode 100644 (file)
index 0000000..47d066a
--- /dev/null
+++ b/matita/matita/contribs/lambdadelta/basic_2/rt_computation/csx_lfpx.ma
@@ -0,0 +1,102 @@
+(**************************************************************************)
+(*       ___                                                              *)
+(*      ||M||                                                             *)
+(*      ||A||       A project by Andrea Asperti                           *)
+(*      ||T||                                                             *)
+(*      ||I||       Developers:                                           *)
+(*      ||T||         The HELM team.                                      *)
+(*      ||A||         http://helm.cs.unibo.it                             *)
+(*      \   /                                                             *)
+(*       \ /        This file is distributed under the terms of the       *)
+(*        v         GNU General Public License Version 2                  *)
+(*                                                                        *)
+(**************************************************************************)
+
+include "basic_2/rt_computation/cpxs_lfpx.ma".
+include "basic_2/rt_computation/csx_drops.ma".
+include "basic_2/rt_computation/csx_cpxs.ma".
+
+(* STRONGLY NORMALIZING TERMS FOR UNCOUNTED PARALLEL RT-TRANSITION **********)
+
+(* Advanced properties ******************************************************)
+
+(* Basic_2A1: was just: csx_lpx_conf *)
+lemma csx_lfpx_conf: ∀h,o,G,L1,T. ⦃G, L1⦄ ⊢ ⬈*[h, o] 𝐒⦃T⦄ →
+                     ∀L2. ⦃G, L1⦄ ⊢ ⬈[h, T] L2 → ⦃G, L2⦄ ⊢ ⬈*[h, o] 𝐒⦃T⦄.
+#h #o #G #L1 #T #H @(csx_ind_cpxs … H) -T
+/5 width=3 by csx_intro, lfpx_cpx_trans, lfpx_cpxs_conf/
+qed-.
+
+lemma csx_abst: ∀h,o,p,G,L,W. ⦃G, L⦄ ⊢ ⬈*[h, o] 𝐒⦃W⦄ →
+                ∀T. ⦃G, L.ⓛW⦄ ⊢ ⬈*[h, o] 𝐒⦃T⦄ → ⦃G, L⦄ ⊢ ⬈*[h, o] 𝐒⦃ⓛ{p}W.T⦄.
+#h #o #p #G #L #W #HW @(csx_ind … HW) -W
+#W #_ #IHW #T #HT @(csx_ind … HT) -T #T #HT #IHT
+@csx_intro #X #H1 #H2
+elim (cpx_inv_abst1 … H1) -H1
+#W0 #T0 #HLW0 #HLT0 #H destruct
+elim (tdneq_inv_pair … H2) -H2
+[ #H elim H -H //
+| -IHT #H lapply (csx_cpx_trans … o … HLT0) // -HT
+  #HT0 lapply (csx_lfpx_conf … HT0 … (L.ⓛW0)) -HT0 /4 width=1 by lfpx_pair/
+| -IHW -HT /4 width=3 by csx_cpx_trans, cpx_pair_sn/
+]
+qed.
+
+lemma csx_abbr: ∀h,o,p,G,L,V. ⦃G, L⦄ ⊢ ⬈*[h, o] 𝐒⦃V⦄ →
+                ∀T. ⦃G, L.ⓓV⦄ ⊢ ⬈*[h, o] 𝐒⦃T⦄ → ⦃G, L⦄ ⊢ ⬈*[h, o] 𝐒⦃ⓓ{p}V.T⦄.
+#h #o #p #G #L #V #HV @(csx_ind … HV) -V
+#V #_ #IHV #T #HT @(csx_ind_cpxs … HT) -T #T #HT #IHT
+@csx_intro #X #H1 #H2
+elim (cpx_inv_abbr1 … H1) -H1 *
+[ #V1 #T1 #HLV1 #HLT1 #H destruct
+  elim (tdneq_inv_pair … H2) -H2
+  [ #H elim H -H //
+  | /4 width=3 by csx_cpx_trans, csx_lfpx_conf, lfpx_pair/
+  | -IHV /4 width=3 by csx_cpx_trans, cpx_cpxs, cpx_pair_sn/
+  ]
+| -IHV -IHT -H2
+  /4 width=7 by csx_cpx_trans, csx_inv_lifts, drops_drop, drops_refl/
+]
+qed.
+
+fact csx_appl_theta_aux: ∀h,o,p,G,L,U. ⦃G, L⦄ ⊢ ⬈*[h, o] 𝐒⦃U⦄ → ∀V1,V2. ⬆*[1] V1 ≡ V2 →
+                         ∀V,T. U = ⓓ{p}V.ⓐV2.T → ⦃G, L⦄ ⊢ ⬈*[h, o] 𝐒⦃ⓐV1.ⓓ{p}V.T⦄.
+#h #o #p #G #L #X #H @(csx_ind_cpxs … H) -X
+#X #HVT #IHVT #V1 #V2 #HV12 #V #T #H destruct
+lapply (csx_fwd_pair_sn … HVT) #HV
+lapply (csx_fwd_bind_dx … HVT) -HVT #HVT
+@csx_intro #X #HL #H
+elim (cpx_inv_appl1 … HL) -HL *
+[ -HV #V0 #Y #HLV10 #HL #H0 destruct
+  elim (cpx_inv_abbr1 … HL) -HL *
+  [ #V3 #T3 #HV3 #HLT3 #H0 destruct
+    elim (lift_total V0 0 1) #V4 #HV04
+    elim (eq_term_dec (ⓓ{a}V.ⓐV2.T) (ⓓ{a}V3.ⓐV4.T3))
+    [ -IHVT #H0 destruct
+      elim (eq_false_inv_tpair_sn … H) -H
+      [ -HLV10 -HV3 -HLT3 -HVT
+        >(lift_inj … HV12 … HV04) -V4
+        #H elim H //
+      | * #_ #H elim H //
+      ]
+    | -H -HVT #H
+      lapply (cpx_lift … HLV10 (L.ⓓV) (Ⓕ) … HV12 … HV04) -HLV10 -HV12 /2 width=1 by drop_drop/ #HV24
+      @(IHVT … H … HV04) -IHVT /4 width=1 by cpx_cpxs, cpx_bind, cpx_flat/
+    ]
+  | -H -IHVT #T0 #HLT0 #HT0 #H0 destruct
+    lapply (csx_cpx_trans … HVT (ⓐV2.T0) ?) /2 width=1 by cpx_flat/ -T #HVT0
+    lapply (csx_inv_lift … L … (Ⓕ) … 1 HVT0 ? ? ?) -HVT0
+    /3 width=5 by csx_cpx_trans, cpx_pair_sn, drop_drop, lift_flat/
+  ]
+| -HV -HV12 -HVT -IHVT -H #b #V0 #W0 #W1 #T0 #T1 #_ #_ #_ #H destruct
+| -IHVT -H #b #V0 #V3 #W0 #W1 #T0 #T1 #HLV10 #HV03 #HLW01 #HLT01 #H1 #H2 destruct
+  lapply (cpx_lift … HLV10 (L. ⓓW0) … HV12 … HV03) -HLV10 -HV12 -HV03 /2 width=2 by drop_drop/ #HLV23
+  @csx_abbr /2 width=3 by csx_cpx_trans/ -HV
+  @(csx_lpx_conf … (L.ⓓW0)) /2 width=1 by lpx_pair/ -W1
+  /4 width=5 by csx_cpxs_trans, cpx_cpxs, cpx_flat/
+]
+qed-.
+
+lemma csx_appl_theta: ∀h,o,a,V1,V2. ⬆[0, 1] V1 ≡ V2 →
+                      ∀G,L,V,T. ⦃G, L⦄ ⊢ ⬈*[h, o] ⓓ{a}V.ⓐV2.T → ⦃G, L⦄ ⊢ ⬈*[h, o] ⓐV1.ⓓ{a}V.T.
+/2 width=5 by csx_appl_theta_aux/ qed.