Skip to Content.
Sympa Menu

coq-club - Re: [Coq-Club] [Oktoberfest 2017] Sad news: Kosta Dosen

coq-club AT inria.fr

Subject: The Coq mailing list

List archive

Re: [Coq-Club] [Oktoberfest 2017] Sad news: Kosta Dosen


Chronological Thread 
  • From: "1337777.OOO" <1337777.ooo AT gmail.com>
  • To: psh AT uni-tuebingen.de, categories AT mta.ca, coq-club AT inria.fr, types-list AT lists.seas.upenn.edu
  • Cc: Steve Awodey <awodey AT cmu.edu>
  • Subject: Re: [Coq-Club] [Oktoberfest 2017] Sad news: Kosta Dosen
  • Date: Sun, 29 Oct 2017 10:37:12 -0400
  • Authentication-results: mail3-smtp-sop.national.inria.fr; spf=None smtp.pra=1337777.ooo AT gmail.com; spf=Pass smtp.mailfrom=1337777.ooo AT gmail.com; spf=None smtp.helo=postmaster AT mail-io0-f193.google.com
  • Ironport-phdr: 9a23:i6ZuuBamYJ6GeLycFdSYPRv/LSx+4OfEezUN459isYplN5qZpcm+bnLW6fgltlLVR4KTs6sC0LWG9f24EUU7or+/81k6OKRWUBEEjchE1ycBO+WiTXPBEfjxciYhF95DXlI2t1uyMExSBdqsLwaK+i76xXcoFx7+LQt4IPjuUs6X1pzvlrP6x5qGex5IrCe7YPZuMRi8rAPeu89QjYYmYq041hbGpWFJdv9+wH9hY1ma21734d7195p++QxRvegg/ohOS+GyY7UgQKcdCDU9L0gx5db3rl/YQAKUoGYEX2MQ1BdEHlvr9hb/C92o6XCl6bUing69Goe+GelvAmv9sPgtTxjvoCgCPj89tmrQj5oj3+pgvBu9qkknkMbva4aPOa87J/uFcA==

Proph

https://gitlab.com/1337777/dosen/blob/master/dosenSolution101.v

solves some question of Dosen which is how to program the
logical/particular/rewriting/(one-step) polymorphism/cut-elimination
and the confluence/commutation of this rewriting for the grammatical
polymorph pairing-projections ( "product" ) ...

The computational/total/asymptotic/reduction/(multi-step)
polymorphism/cut-elimination of many polymorph mathematics were
already programmed in the earlier COQ texts : polymorph
reflections/adjunctions , polymorph comonad , polymorph
grammatical-metafunctors , polymorph localizations ... The form of
these programs was the same : primo to define grammatical (multi-step)
reduction relation , then secondo to define some non-structural
recursive grade function from the morphisms to the integers , and
finally tertio to define some non-structural recursive cut-elimination
function from any two-morphisms input ( "topmost cut" ) by choosing
some syntactic-match cases precedence when there are ambiguities ref
which reduction shall apply . Memo that the grade function was linear
arithmetic ( addition , maximum , less than ) over the integers .

Now to proceed to the logical consequences of such things shall
require some alternative logical description : some
logical/particular/rewriting/(one-step) polymorphism/cut-elimination
resolution and the confluence/commutation lemma of this rewriting
which shows that this logical resolution is indeed some alternative
description of the computational function . Memo that the grade
function here will ( necessarily ? ) be non-linear arithmetic (
addition , multiplication , distributivity , less than ) over the
integers which mimic/simulate the constructors of the morphisms , in
particular : the polymorphism ( "distributivity" ) of the
composition/cut constructor over the pairing constructor . More
precisely : grade ( f o> g ) = 3 * grade f * grade g ; grade ( 1 ) =
1 ; grade ( ~_1 o> f ) = 1 + grade f ; grade ( ~_2 o> f ) = 1 +
grade f ; grade ( << f , g >> ) = 1 + grade f + grade g .

This COQ program and deduction is mostly-automated , and the COQ [nia]
automation-tactic is sufficient for this non-linear arithmetic . The
confluence lemma produce some total of ( ( 7 * 8 ) * ( 7 * 8 ) = 3136
goals ) , but only ( 37 classes of goals) are critical via the
automation-tactics ; moreover the COQ computer ( parallel dual core ,
4 gigabytes memory ) motion-time for ( 93 + 27 = 120 minutes ) .

For instant first impression , the lemma that the « resolved-cut » is
congruent modulo solution-conversion is written as :

#+BEGIN_EXAMPLE
Fixpoint solveCoMod_PolyCoMod_cong_Post_ len {struct len} :
forall (B1 : obCoMod) (B1' : obCoMod) (g'Sol g'Sol0 : 'CoMod(0 B1 ~>
B1' )0 %sol)
( Hconv : ( g'Sol0 ~~~ g'Sol )%sol_once ),
forall B2 (g_Sol : 'CoMod(0 B2 ~> B1 )0 %sol),
forall (H_gradeParticular : ( (Sol.gradeConv Hconv +
Rewrite.gradeParticular ((Sol.toCoMod g_Sol) o>CoMod (Sol.toCoMod
g'Sol))%poly)%Z <= Z.of_nat len)%Z ),
( ( ( g_Sol o>CoMod g'Sol0 ) ~~~ ( g_Sol o>CoMod g'Sol ) ))
%sol %sol_conv .
#+END_EXAMPLE

This file [[1337777.OOO//dosenSolution101.v]] is the epilogue of Dosen
book « Cut-Elimination in Categories » .

Keywords : 1337777.OOO//dosenSolution101.v ; pairing-projections
grammar ; logical cut-elimination ; confluence ; non-linear arithmetic

Outline :

* Grammatical description of polymorph pairing-projections
** Importing the non-linear arithmetic COQ automation-tactics
** Common-inductive polymorph morphisms, with overloaded notations
for the composition/cut constructor

* Solution
** Grade of solution-conversion for combined non-structural induction
** Destruction of solution-morphisms with inner-instantiation of
object-indexes

* Congruent-Rewriting relation
** Non-linear mimicking grade function
** Degradation lemma and the non-linear arithmetic COQ automation-tactic
[nia]

* Polymorphism/cut-elimination by congruent-rewriting

* Confluence
** Goals accounting
** Unique solution for the logical-rewriting relation

* Polymorphism conversion
** Cut-elimination resolution is congruent from the
polymorphism-terminology to the solution-terminology

-----

PS : Oktoberfest 2017 continues as NovemberSchätz 2017 at the CMU
Schätz weekends 11:37-14:00 until november ...

-----

memory amazon.com/hz/wishlist/ls/28SN02HR4EB8W , eth
0x54810dcb93b37DBE874694407f78959Fa222D920 , paypal
1337777.OOO AT gmail.com
, wechatpay
2796386464 AT qq.com
, irc #OOO1337777


  • Re: [Coq-Club] [Oktoberfest 2017] Sad news: Kosta Dosen, 1337777.OOO, 10/29/2017

Archive powered by MHonArc 2.6.18.

Top of Page