[ODE] Iterative solution

[Gary R. Van Sickle]

> Gary R. Van Sickle wrote:
> >>Gary R. Van Sickle wrote:
> >> > BUT: instead of relying on islands, could each constraint be
> >>
> >>>provisionally removed from the loop when it meets the error criteria?
> >>
> >>I'm a bit dubious -- the nature of iterative constraint solving
> >>means that any particular constraint could have a metric truckload
> >>of propogated error dumped on its head from 'upstream' at any point.
> >
> > True.  What about the sorting idea I floated a while back?  Sort from worst
> > violator to least violator, solve them in order, and do this "provisional
> > culling"?
>
> I don't remember reading that.  It sounds reasonable, or at least of
> interest for experimentation.  Given that I can't find the mail in
> which you floated this, how did you propose that 'violation' be
> evaluated?  Rate of convergence?
>

Uh... well... I left that exercise to the reader ;-).  I was thinking more along
the lines of sorting by how much force+torque was going to be applied if this
was the first constraint to be solved in isolation.  This would appear to
require one essentially wasted run through the constraints to determine this,
but maybe it'd pay for itself by reducing the number of iterations required.

>  > Hehehehehe!  I ain't 'fraida no ghosts!
>
> I don't remember (I know there seems to be a trend in my lack of
> memory) exactly what prompted me to add that to my signature, but
> it has served me well. :)

Just don't cross the streams.  That would be bad. ;-).

--
Gary R. Van Sickle
Brewer.  Patriot.