Skip to Content.
Sympa Menu

ssreflect - RE: defining an executable version of rank

Subject: Ssreflect Users Discussion List

List archive

RE: defining an executable version of rank


Chronological Thread 
  • From: Georges Gonthier <>
  • To: Vincent Siles <>, "" <>
  • Subject: RE: defining an executable version of rank
  • Date: Tue, 21 Jun 2011 13:09:13 +0000
  • Accept-language: en-GB, en-US

This is probably the simplest way, but by no means the only one. You could
also establish other properties that characterise the rank; the mxalgebra
library provides many of these for the ssr rank function, so by proving
my_rank = /rank you'll get them all. Conversely these properties show that
any "rank" function characterized by, say, a Smith normal form, must coincide
with the ssr rank function.
Note that as the ssr rank is computed by Gaussian elimination in a field,
you will have to start with an independent characterization if your
definition is more general (say, over a PID domain). Also note that in full
generality the several "rank" variations do not always coincide.

Georges

-----Original Message-----
From: Vincent Siles []
Sent: 21 June 2011 13:31
To:
Subject: defining an executable version of rank

I have defined a function to effectively compute the rank of matrix and want
to prove it correct in ssr. Is the only way to prove is to prove that is
something like

forall n m (A : 'M[R]_(n,m)), \rank A = my_rank n m A

?

Cheers,
Vincent




Archive powered by MHonArc 2.6.18.

Top of Page