[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: [mizar] A puzzle



On Thu, Jan 03, 2002 at 02:03:20PM +0100, Andrzej Trybulec wrote:
> In a new submitted article - AFINSQ_1 (The Authors:Tsunetou, Bancerek,
> Nakamura) the theorem Th2 says
> 
> theorem Th2:
>    k <= n implies k = k /\ n
> 
> the next theorem has a straightforward justification:
> 
> theorem
>   k = k /\ n implies k <= n by Th2;
> 
> ("/\" is the new notation for the intersection of sets, and the von
> Neuman construction is used in MML, so both are true)
> 
> Mizar does not allow, as a rule, to justify "alpha implies beta" by
> "beta implies alpha" :-)
> 
> Why it works in this case?

One explanation is that the following facts are obvious to the checker:

k > n implies k <> n;

k > n implies k >= n;

and intersection is commutative.

-- 
Piotr Rudnicki               CompSci, Univerity of Alberta, Edmonton, Canada
email: piotr@cs.ualberta.ca                 http://web.cs.ualberta.ca/~piotr