|time ordering for PID contributions [message #4599]
||Mon, 02 July 2007 11:09
Registered: March 2006
In my current macro-based simulations I do some simple analysis
of hit patterns on detector elements, but I do not care yet when
some external ingredients into the analysis become available.
A couple of months ago I heard about the requirement to provide
a fast kaon trigger from the Cherenkov detectors. Implementation
of the Endcap detector simulation has also started, and I feel
that the simulation analysis to be implemented in PandaRoot
should also reflect some of the online requirements.
For instance one could analyse the Endcap DIRC standalone and
then combine results including covariance matrix with similar
results from tracking. Standalone beta is poor, but the
off-diagonal elements linking to vertex direction are large.
So this analysis could already be done without tracklets
(I understand as being short straight stretches in a tracking
detector nearby - to be extrapolated into the DIRC)
being there yet.
Another question is how to seed the Cherenkov analysis.
Some roughly defined punch-through points externally provided
might already be enough. When would they be available?
The alternative would be to do track-finding with the
Cherenkov data, but beyond some ideas about the signature
to look for no work has been done yet, and I cannot say
how fast this will be compared to neighbouring devices.
So what I suggest is to think about the time-ordering and
logical dependence of data analysis chunks across detectors.