- To: mathgroup at smc.vnet.net
- Subject: [mg61473] GradPack v1.1
- From: "Ofek Shilon" <ofek at simbionix.com>
- Date: Wed, 19 Oct 2005 23:07:21 -0400 (EDT)
- Sender: owner-wri-mathgroup at wolfram.com
Hi dear Mathgroup.
I just submitted to the MathSource a notebook called GradPack v1.1,
that personally made my life a lot easier -
The original task it addressed was formulating gradients and hessians
of complex potentials in R^n (i.e., real smooth functions on R^n). It
evolved into a symbolic abstract-differentiation system, which enables
computation of gradients of scalar fields and jacobiabs of vector
Both input and output are given at a coordinate-free abstract level -
e.g., once the symbol v is declared as a vector, (using the appropriate
syntax), you can immediately obtain results such as
Grad[ ||v|| ^ 14 ] = 14 (||v||^12) Transpose[Jacobian[v]].v ,
and of course - variants up to arbitrary complexity. Elaborate examples
are contained in the notebook.
(this isn't the exact notation the notebook uses - it can't pass as
The main design emphasis was on ease of use and intuitive notations.
The notebook is available at -
I'd much appreciate any feedback & comments.
Prev by Date:
Re: Display sound but suppress graphics
Next by Date:
Re: Re: Stylesheets vs. DTDs or XML Schemas
Previous by thread:
Next by thread:
JLink share kernel from multiple Java Applications