[plt-scheme] DrScheme as alternative to Matlab

The most important thing is to get the underlying multi-dimensional array
structure and associated syntax for slicing, etc. These should allow me to
do things like adding the elements of the third dimension of A to the vector
B and storing the results in the third row of X. This should be done by
efficiently striding through the referenced array structures without needing
any temporary vectors, etc. Matlab and Numpy (the Python numeric package)
are good places to look.

Originally, my thought was to also allow the specification of an underlying
representation (I was using u8, u16, u32, u64, s8, s16, s32, s64, f32, f64,
c64, and c128 for various unsigned and signed integers, and single and
double precision floats and complexes). I have this code running and will
put in up on the Schematics site and post a reference later. It works fine,
but I was disappointed in it's performance relative to just plain Scheme
vectors. [It uses the SRFI 4 vector representations underneath, which I
think map to definitions in the foreign function interface (FFI). We could
work with Matthew and see if there are something we can do to improve the
efficiency.] I still think that giving the user the option to specify the
representation is the right thing to do from a numerical analysis
perspective.

I'll put some more time into writing up my own thoughts on the
multi-dimensional array structure and the syntax for referencing slices,
etc.

Doug

On Tue, Aug 11, 2009 at 8:15 PM, Matthias Felleisen <matthias at ccs.neu.edu>wrote:

>
> Doug, can you describe to someone like Jason how he can assist you?
>
>
> On Aug 11, 2009, at 8:57 PM, Doug Williams wrote:
>
> We actually do a lot of analysis work in Scheme - migrating from Python and
> Matlab. The science collection (
> http://planet.plt-scheme.org/display.ss?package=science.plt&owner=williams)
> and PLot (that comes with PLT Scheme) are the basis for this. We also have
> some additional plotting and animation capabilities that aren't (yet) on
> PLaneT. I had hoped to mature the science collection into something I was
> calling Schemelab, which would be more MatLab-like, but work has gotten me
> back over on the agent-based simulation side again.
>
> Anyway, depending on what you're looking to do, much of the framework is
> there. I do wish it was more integrated, but it takes time.
>
> Doug
>
> On Tue, Aug 11, 2009 at 3:26 PM, Jason <jason.lillywhite at gmail.com> wrote:
>
>> Anyone out there heard of someone using drscheme as an alternative to
>> matlab? With some libraries and plotting ability, it seems it could
>> work. I'll be starting an engineering programming class with a lab
>> that uses matlab. I might try doing some of the assignments in
>> DrScheme...
>> _________________________________________________
>>  For list-related administrative tasks:
>>  http://list.cs.brown.edu/mailman/listinfo/plt-scheme
>>
>
> _________________________________________________
>   For list-related administrative tasks:
>   http://list.cs.brown.edu/mailman/listinfo/plt-scheme
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.racket-lang.org/users/archive/attachments/20090812/6e015049/attachment.html>