On Wed, Feb 11, 2009 at 8:32 AM, Greg Kuperberg
<[email protected]> wrote:
>
> On Feb 11, 8:10 am, William Stein <[email protected]> wrote:
>> By echelon form do you mean Hermite Normal Form over that base ring,
>> or do you mean echelon form over the fraction field?  If you just want
>> the dimension, is their any chance you can specialize one of the variables 
>> and
>> compute the rank there -- if you do this for enough values it's likely
>> to give you the correct rank, and of course is very fast.
>
> Yes, I meant Hermite Normal Form.
>
> The code that I have now basically specializes the (lone) variable d.
> I make 14x14 matrices e[0] through e[6] and then the last two lines of
> the code are this:
>
> L = gap.LieAlgebra('Rationals',e)
> print 'Dimension:',gap.Dimension
> (L)
>
> What I really wanted to say in the paper is that L is the full 14x14
> matrix algebra for all real values of d with |d| >= 2.  I was already
> 90% sure that this is true.  I wanted to use SAGE to provide a
> computational proof, because if I worked out a human proof then I
> would have to digress into a topic for a later paper.  The problem, I
> realized, is that with this code SAGE cannot provide a proof for all
> relevant d, only for specific values of d.  (Moreover there exist
> algebraic values of d for which L is smaller.)
>
> In truth, SAGE is missing two things for what I wanted.  First, there
> is no equivalent of gap.LieAlgebra.  Second, the submodule method of
> the FreeModule class is not implemented for Euclidean domains, even
> though it is implemented for ZZ.  Of course, if gap.LieAlgebra worked
> over a Euclidean domain, that would also solve the problem, but I
> don't think that it does.
>

I hope somebody implements Hermite normal form over a general PID. It
would probably take one day for somebody who knows what they're doing.
 One student did do it once about 18 months ago, but lacked follow
through.

I wrote a lot of the code in FreeModule in such a way that as long as
there is a hermite_form command then the algorithms should work --
i.e., the one and only reason all the general free module machinery
doesn't work over anything but ZZ is the lack of a Hermite form
command.

 -- William

--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/sage-support
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---

Reply via email to