[Halld-offline] recent change to tracking code

[Simon Taylor]

Hi, Richard.

For now I will implement a method in DKinFit that converts from the 5x5 
matrix to the 7x7 matrix.  In the future Paul and I are considering just
using the 5x5 representation in DKinFit directly, but this is not ready 
to go yet.

Simon

On 07/09/2012 10:04 AM, Richard Jones wrote:
> Simon,
>
> Igor showed me that the DKinFit class is still using the 7x7 error 
> matrix as input.  Because of this, and because the 5x5 was heretofore 
> in a non-standard representation, I implemented the 7x7 matrix into 
> the REST model.  I save 30% in data volume if I can use your new 5x5 
> representation instead.  Do you have a standard method for massaging 
> the new standard 5x5 matrix for presentation to DKinFit?
>
> -Richard J.
>
>
> On 7/9/2012 8:01 AM, Simon Taylor wrote:
>> Hello, everyone.
>>
>> I have made the following change to the Kalman Filter code:  instead of
>> providing 5x5 covariance matrices and 5x1 state vectors whose meaning
>> depended on which parametrization was used in the fit, if the forward
>> parametrization was used I have now rotate the 5x5 matrix and 5x1 state
>> vector from this space into the central parametrization; in other words
>> all the tracking results are now stored in DKinematicData using the
>> state vector {q/pT,phi,tanl(lambda),D,z}.
>>
>> Simon
>> _______________________________________________
>> Halld-offline mailing list
>> Halld-offline at jlab.org
>> https://mailman.jlab.org/mailman/listinfo/halld-offline
>
>
>