On Aug 27, 2008, at 1:19 PM, Robert Bradshaw wrote:
On Wed, 27 Aug 2008, Craig Citro wrote:
Thoughts?
I vote for fast.
I also vote for fast -- but couldn't there be a flag for the slow
option? Maybe consistent=True or something, in case someone really
wants it? I could at least
On Wed, Aug 27, 2008 at 10:04 AM, Martin Albrecht
[EMAIL PROTECTED] wrote:
Hi there,
Robert wrote at
http://trac.sagemath.org/sage_trac/ticket/3956#comment:7
Matrix hashes are specifically designed to be compatible with each other:
{{{
sage: M = random_matrix(GF(2), 10, 10)
On Aug 27, 9:05 pm, Craig Citro [EMAIL PROTECTED] wrote:
I also vote for fast -- but couldn't there be a flag for the slow
option? Maybe consistent=True or something, in case someone really
wants it?
AFAIK, the key requirement for a hash function is that equivalent (in
the sense of ==)
On Wed, Aug 27, 2008 at 12:21 PM, Simon King
[EMAIL PROTECTED] wrote:
On Aug 27, 9:05 pm, Craig Citro [EMAIL PROTECTED] wrote:
I also vote for fast -- but couldn't there be a flag for the slow
option? Maybe consistent=True or something, in case someone really
wants it?
AFAIK, the key
And how should one pass the argument consistent=True in that
situation?
What was malb's original definition of inconsistent? Did he *really*
have pairs of GF(2) matrices that are equal, but have different hashes?
Dense matrices over GF(2) have the same hash if they are equal. However, a
On Aug 27, 12:21 pm, Simon King [EMAIL PROTECTED] wrote:
On Aug 27, 9:05 pm, Craig Citro [EMAIL PROTECTED] wrote:
I also vote for fast -- but couldn't there be a flag for the slow
option? Maybe consistent=True or something, in case someone really
wants it?
AFAIK, the key requirement for
On Aug 27, 9:41 pm, William Stein [EMAIL PROTECTED] wrote:
AFAIK, the key requirement for a hash function is that equivalent (in
the sense of ==) objects must have the same hash.
Just for the record, we break that all over the place. However, we do
try to keep it if possible over a given
His hash function depends on the characteristic of the base ring,
the matrix entries, the matrix dimension,
I'll revise it so that it is more similar to the current model (while
preserving speed) and so that sparse=True and sparse=False still hash to the
same value.
and also on the size
On Aug 27, 9:52 pm, Simon King [EMAIL PROTECTED] wrote:
I understood he only has pairs of matrices over *different* rings that
are equivalent but have different hashes.
... i forgot the dense vs. sparse issue, sorry.
Do we agree that if a dense and a sparse matrix defined over the same
ring
On Wed, 27 Aug 2008, Craig Citro wrote:
Thoughts?
I vote for fast.
I also vote for fast -- but couldn't there be a flag for the slow
option? Maybe consistent=True or something, in case someone really
wants it? I could at least imagine a case where someone might care, in
which case we
10 matches
Mail list logo
