Re: sparse objects

On Thu, 11 Oct 2001, Leonardo B. Lopes wrote:

> Dear Friends,
> 
> 	About a year ago I asked around in this list and the OM lists 
> about sparse matrices and vectors, which are central to my field of
> research. I also (thanks to generous support from WRI) attended the
> MathML conference in Urbana. The answer I got was basically that MathML
> was very strongly focused on K12 and that I shouldn't hold my breath. 
> 
> 	Well, it turns out that our community is now sufficiently
> organized that it looks like our standard communication format is actually
> going to become reality. Of course we would love to be able to use a
> MathML structure for representing our sparse objects. We would at least
> like to convey our structures in a way that MathML-aware applications
> would understand them.
> 
> 	I searched the MathML 2.0 recommendation and found no reference to
> sparse objects. But apparently there are new features in Mathematica 4.0
> to deal with sparse objects. So that gives me some more hope. 
> 

Hi,

There is an OpenMath CD, which deals with structured matrices, which to a
large extent are pretty sparse
(http://www.openmath.org/cdfiles/html/extra/cd/linalg5.html). It doesn't
deal with general sparse matrices, though it would not be a big deal to
create a private one that did, (I would suggest representing them as a
list of tuples, the coordinates and the values of the non zero entries +
some dimension and rank (though this is redundant) information for the
whole matrix). Of course if you want these to be MathML compatible, then
you use a semantics element or a csymbol element.

hope this has been of help,

Bill.

--

Received on Friday, 12 October 2001 10:01:49 UTC