[racket] Thanks for Math

Recently I used it for estimating orientation parameters of telescopes. Inputs are the locations of stars, outputs are telescope coordinates, parameters describe the relationship of axes to each other and to the earth.

In matrix-solve, MX=B are
M : J[transpose]J, where the J matrix is filled with numerically computed derivatives of each parameter vs outputs.

B : J[transpose]E, where E is filled with the errors in each axis as measured by observing stars

X : the changed in parameters that are to be solved on each iteration of the estimator

Due to numerous orthogonal components and parallel parameters (for certain pointing directions), the matrix library has to be rigorously written, else the iterations die of round-off errors etc. I had several choices, but given the high caliber people that write Racket code, I figured I could go with matrix-solve and expect it to work.

rac


On Apr 21, 2013, at 11:11 AM, Joe Gilray wrote:

> Yes, the math library is awesome.  Many thanks.
> 
> Richard, what are you using it for? (if you are liberty to say, of course).
> 
> -joe
> 
> 
> On Sat, Apr 20, 2013 at 8:28 PM, Richard Cleis <rcleis at me.com> wrote:
> Thanks for the math library. I started using some of the linear algebra functions for my paid programming, and have had zero trouble with any of it. Matrix-solve works like a dream, and the docs are on the mark.
> 
> I have to wonder what would happen if a program manager were looking over my shoulder on the rare occasion that I commit a Contract Violation for issues like non-invertible matrices. PMs are a humorless lot in regards to violation of contractual terms.
> 
> rac
> 
> 
> ____________________
>   Racket Users list:
>   http://lists.racket-lang.org/users
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.racket-lang.org/users/archive/attachments/20130422/0ee37a4c/attachment.html>