[ODE] Quickstep and patents

[Klaus Backert]

Hello,

Am 07.06.2007 um 16:13 schrieb Patrick Enoch:

> Hello all,
>
> this is really funny!
>
>
> How exactly do these patents work? Is only the *full interaction of
> all parts* patented (e.g. SOR + Systemmatrix + Mass-matrix), or *all
> single parts seperately*?
>
> For each single part, there should be thousands of references BEFORE
> 2004. For example, Gauss-Seidel originates at least back to 1855
> (that is when Gauss died), SOR is a modification also known a long  
> time.
>
> Massmatrix should be covered in any "applied engineering" book
> covering physical properties.
>
> "Jacobian" of joint-freedoms, OMG how old is that exactly? I think
> the introduction of lambda is a smart move, but there should be a
> reference for that, too. BTW, Mr. Jacobi died 1851.
>
> Best regards,
> Patrick

that is so true, Patrick, and let me add this:

 From 1980 on I worked for a big company as a mathematician. We  
implemented a lot of "these" algorithms in a software package. At  
that time several important players in the software market did the  
same, including U.S. companies. Several of these software packages,  
or derivatives of them, are still in use worldwide.

Most, if not all, algorithms used by ODE are based on knowledge,  
which is told as standard at the applied mathematician departments of  
european universities. If I scan the websites of american and  
canadian universities - I did that some time ago because of other  
reasons -, then it looks just the same. The "prior knowledge" resides  
in the textbooks, scripts and brains of many applied mathematicians.

For the european market, by the way, Ageia's supposed patent is  
completely worthless, as far as I know.

Cheers
Klaus