News for hackers. Stuff that matters.
Re:Proving false in Coq using an implementation bug by agumonkey on Tuesday March 24, 2015 @06:59PM (#9260134)
How much does that invalidates Coq results ? like the 4 colors theorem.
Re:Proving false in Coq using an implementation bug by j2kun on Tuesday March 24, 2015 @07:47PM (#9260371)
AFAIK the four color theorem was not first proved with Coq, and people consider it correct regardless of whether Coq has bugs.
Re:Proving false in Coq using an implementation bug by cheatsheet on Tuesday March 24, 2015 @08:34PM (#9260568)
Yes, the four color theorem was proven by Appel and Haken on an IBM 370-168 at the University of Illinois in June 1976. Coq was released in May 1989.
Re:Proving false in Coq using an implementation bug by repsilat on Wednesday March 25, 2015 @09:02AM (#9262607)
I'd be surprised Appel and Haken did a formal proof on their machine -- I thought they made a conventional proof with a machine-assisted enumeration of special-cases.
Re:Proving false in Coq using an implementation bug by cheatsheet on Wednesday March 25, 2015 @10:41AM (#9263228)
My apologies for the lack of clarity: I was following the thought path that was concerned with the validity of the four color theorem proof because of the Coq bug.
Re:Proving false in Coq using an implementation bug by CHY872 on Tuesday March 24, 2015 @08:36PM (#9260571)
It shouldn't, really. It'll only invalidate results that

a) rely on this bug, and

b) haven't been formally proven any other way.

It is a valid concern when talking about theorem provers, though.