Skip to Content.
Sympa Menu

coq-club - RE: [Coq-Club] "Nested proofs are deprecated and will stop working in the next Coq version" - Can we keep this?

coq-club AT inria.fr

Subject: The Coq mailing list

List archive

RE: [Coq-Club] "Nested proofs are deprecated and will stop working in the next Coq version" - Can we keep this?


Chronological Thread 
  • From: "Soegtrop, Michael" <michael.soegtrop AT intel.com>
  • To: "coq-club AT inria.fr" <coq-club AT inria.fr>
  • Subject: RE: [Coq-Club] "Nested proofs are deprecated and will stop working in the next Coq version" - Can we keep this?
  • Date: Tue, 4 Aug 2015 16:49:08 +0000
  • Accept-language: de-DE, en-US

Dear Coq Team and Users,

the reason why I sometimes keep nested lemmas where they are is that with
proof automation it is not possible to get to the state where the lemma is
used otherwise. When the lemma is there, the automation will use it and
finish the goal. The only way to get to the point where one can see why the
lemma is required is to remove it. Then the automation will leave the
interesting cases for inspection. It is sometimes really difficult to explain
why a lemma is required without seeing the proof context. Of cause I could
copy the context to a comment to motivate the intermediate lemma. Or I could
write a comment "please delete this lemma, run the proof below and you will
see why you want it back". I think in such cases one may argue that it is
not bad style, but a reasonable help for others, to keep the lemma where it
is, so that proof automation will stop there, one can look at the case(s) and
then proof automation continues happily with the new lemma. There are cases
where the lemma is useful in similar cases outside of the proof, so using
local hypothesis in hint databases is not an option.

When I look at Coq developments, I frequently find fairly unmotivated rather
obscure lemmas without a single line of comment on why they are there. Maybe
mathematicians (I am physicist) are better trained to see through such
things, but I would think if we want to make Coq accessible to a broader
audience, we need to accept that some things are a bit harder to understand
for others, and that ways which make it easier to see through what is going
on are important. I use nested lemmas rarely for this purpose, but sometimes
I do.

Best regards,

Michael

Intel Deutschland GmbH
Registered Address: Am Campeon 10-12, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Christin Eisenschmid, Prof. Dr. Hermann Eul
Chairperson of the Supervisory Board: Tiffany Doon Silva
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928



Archive powered by MHonArc 2.6.18.

Top of Page