Scilab code differentiation toolbox (Aug 2002). 					

This toolbox enables Scilab code differentiation using operators and
primitive functions overloading.

Given a Scilab code computing a variable y depending on a variable x
and a direction dx it allow evaluation of y together with the
directional derivative Grad(y)*dx.

It is far from complete, but supports   basic computations including
matrix inversion.  It is quite easy to complete adding new overloading
functions in the macro directory. See the functions already defined
and the help on overloading. 

See the help and demos files for more details

Authors:
------- 
Xavier Jonsson INRIA
Serge Steer INRIA Serge.Steer@inria.fr

Installation
============
To Install this toolbox: 

We Suppose here that  stands for  the path of the directory
containing this README file.

- On Unix/Linux systems
     * Administrator
        Has to execute, once and for all, the following instruction
        within Scilab:
        exec /builder.sce
     * User
        Should execute the following instruction within Scilab:
        exec /loader.sce
        before using the toolbox, he  can also put it  in his
        .scilab startup file for automatic loading.

- On Windows systems
     * Administrator
        Has to execute, once and for all, the following instruction
        within Scilab:
        exec  \builder.sce
        This operation requires permission to write in \macros
        to generate *.bin, names and lib files in \macros
        directory.
     * User
        Should execute the following instruction within Scilab:
        exec \loader.sce
        before using the toolbox, he  can also put it  in his
        .scilab startup file for automatic loading. 

					

Ho hum. The code may be ok, but the documentation and the examples are terrible.
I am a Scilab novice, but well accustomed to Matlab and Maple.

The documentation of this package is poor (for example it doesn't tell you what
is going on: does 'automatic calculation of derivatives' mean that derivatives
are calculated analytically, giving a full Scilab function or just through
repeated application of thechain rule, resulting only in *numerical* answers),
hard to read, too much tied-up with Scilab mechanics, and too short. 

The examples shed some light on the matter, but not much. In the end I had to
look at the code to figure out what was going on. After 10 minutes or so I gave
up.

The high-level interface 'der' is not bad when all you want to see are
numerical derivatives, but what I want to see from an automatic derivation
package is symbolic output so that I can check it, and compile it. As it is I
didnt't take the time to study either the output or the code.

Nice idea, but with an hour or two in improving the documentation the whole
package would improve immeasurably. As it is, you would have to have a *really*
pressing for this type of stuff to take the time to dive into the code of this
package to figure out what it does exactly. 

I appreciate the effort though.		

This package is good but not up to date.  People should email Mr. Steer to get
latest release.

I also suggest visiting Mr. Benoit Hamelin site for the SciAD library :
http://www.dmi.usherb.ca/~hamelin/autodiff/html/autodiff_en.html

Both packages are great stuff.		

Best from the waest movies on our site are like yours. Put some more info.		

Links to your site now published on the net. Wellcome aboard, and put some more
info.		

Many peoples use this way of culture. Nice article but need more comment for
discuss.		

Your site is da best, give me more info about the theme.		

Wellcome more info...		

i need some examples for my report because my instructor did'nt even bother to
teach us with numerical integration		

Your email address :
Your rate :	Love it  Useful Just OK none
Comments :
Confirmation code : Enter the code exactly as you see it. The code isn't case sensitive.