Re: [Brahms-dev-l] T2-T3 matching

Hi
Just a short comment.
I am clueless why T2-T3 matching depends on H1, H2 calibrations. This should
only be a function of
tracking detectors - so I will go along with a solution for matching that
looks like BrModuleMatchTrack, but will
strongly disagree to let this module dela with T2,T3 matching (the is no
filed involved). I have sai this before
and will thus repeat. One coul make a matching approximately like this.
We should also recall that due to the independence of FFS an BFS there may
be small systematic geometry
difference, most likely in the x-offsets an not in angles that can effect
how we do t2-t3 matching and t2-t4 matching.

Note this is not a complete note, but summarizes do something like solution
3 but with a module parralle to BrModuleMatchTrack.. I do not like the root
file solution which would add a 3. method for storing calibration/db
information.

flemming
----------------------------------------------------------------
Flemming Videbaek
Physics Department
Brookhaven National Laboratory

e-mail: videbaek@bnl.gov
phone: 631-344-4106
----- Original Message ----- 
From: "Pawel Staszel" <ufstasze@if.uj.edu.pl>
To: "Djamel Ouerdane" <ouerdane@nbi.dk>
Cc: "Brahms Devel List" <brahms-dev-l@lists.bnl.gov>
Sent: Monday, June 14, 2004 3:07 PM
Subject: Re: [Brahms-dev-l] T2-T3 matching


> Hi Djamel,
> You are right, it is very old file, and we certainly have to update both
> the parameter file and the part of the software
> which deals with it.
> Let me explain you how was it done for now and then we can decide about
> the updates.
>
> offsets5361_5972.FS contains matching parameters in ascii format for all
possible detector combinations in FS.
> BrFsTrackingModule utilizes only parameters for T2-T3 combination. Files
like offsets5361_5972.FS
> are generated using the stuff from ps_app/fs/receff is a first part of the
detector efficiency calculation.
> This calculation uses local tracks as an input and depends only on H1 and
H2 calibration.
> (the bad thing is that H1 and H2 calibrations depends on global tracking
but for T2-T3 we don't need it)
> In the mean time I moved from ascii to root format (see
~ufstasze/effic/EfficParamFile.root).
>
> As for now I see the following 3 solutions:
> 1: update BrFsTrackingModule, so it can read root format files: the
advantage of this solution is that we don't
> have to update a lot. The disadvantage is however, that you generate
offsetes for all detector combinations
> but still these offsets have to be redone because of H1 and H2.
> 2. Extract from ps_app/fs/receff what is needed and create a dedicated
module for generating T2-T3 offsets only.
> Then it will depend only on the localtracks.
> We can decide about the format, or maybe, for this 8 float numbers we can
create a db entry.
> 3. Generate T2-T3 offsets at the same time when the other parameters
utilized in BrModuleMatchTrack are generated.
> The last solution would required some updates in BrModuleMatchTrack but I
thing that this solution would
> significantly simplified a book keeping and would made the whole analysis
process more solid
> (less possibility for a mistake). It would automatically took the
advantage of your latest updates (brat version 2.17.8).
>
> So, I vote for the 3rd solution.
> I can provide the updates but let us first agree on of how to proceed.
>
> Cheers
> Pawel.
>
>
>
>
>
> Djamel Ouerdane wrote:
>
> >Hello,
> >
> >I was checking the BrFsTrackingModule and I saw the following: FFS and
BFS
> >tracks are combined by checking some matching between T2 and T3 tracks.
> >The module still uses an offset file called
> >
> >trackmatch/offsets5361_5972.FS
> >
> >How good is that ?? runs 5xxx are almost prehistory compared to 11xxx.
> >I guess this question is for Pawel. How did you generate this file ??
> >
> >Djam
> >
> >
> >
> >
>
>
>
> _______________________________________________
> Brahms-dev-l mailing list
> Brahms-dev-l@lists.bnl.gov
> http://lists.bnl.gov/mailman/listinfo/brahms-dev-l
>


_______________________________________________
Brahms-dev-l mailing list
Brahms-dev-l@lists.bnl.gov
http://lists.bnl.gov/mailman/listinfo/brahms-dev-l