BRAT-2-0-28 (Was: "Re: BRAT-2-0-27" )

From: Jens Ivar Jordre (jens@fi.uib.no)
Date: Fri Aug 17 2001 - 15:15:29 EDT

  • Next message: Eun-Joo Kim: "RICH for supermon"

    Howdy brats.
    
    I updated BRAT after my not so successfull [:)] additions yesterday. New
    version bumbed 2.0.28 and tagged BRAT-2-0-28.
    
    On Fri, 17 Aug 2001, Christian Holm Christensen wrote:
    
    > Well, I don't mind the class BrSpectrometerTrackingPackage - tough
    > it's a bloody long name - so if you want to leave I have no real
    > problem with that.  Hmm, maybe I made it a bit to easy to make new
    > modules/packages by introducing the Elisp function 'brat-class' :-)
    
    I renamed BrMrsTrackingPackage to BrMrsPackage and BrFfsTrackingPackage to
    BrFfsPackage.
    
    For now I also removed BrSpectrometerTrackingPackage. Anyone adding more
    functionality to the above mentioned packages should however have in mind
    that interface similar to both spectrometer arms could go into a base
    class like BrSpecPackage.
    
    > Uh! g2root does a full conversion of the GEANT3.21 [1]
    > volume/material/position/rotation common blocks (sigh) which is far
    > far too much for an event display.
    
    It all depends what you want to display. If you want to have the option of
    displaying different materials differently, e.g. with color coding, then
    this is not farfetched.
    
    > You should rather define the volumes in the detector
    > specific classes, using the service of BrGeometryDbManager.  Also
    > note, that ROOT has already implemented the behaviour you're
    > addressing via globals gGeomtry [2].
    
    My initial idea was actually to generate geometries/nodes using
    BrGeometryDbManager. But why do things twice as the ROOT geometry is
    easily with g2root. That was my reason for using g2root generated macros,
    as well as my point above regarding materials.
    
    > Hmm.  All in all, I think all this belongs in a seperate package from
    > BRAT, since it's not likely that the code will be used outside of the
    > event display.
    
    The intention has never been to put TED in BRAT. Rather it should go into
    BROP as it will typically be used by shift operators. Some of the classes
    I made, however, I thought would be usefull in other contexts as well.
    Therefore they were put into BRAT.
    
    Anyhow, I've removed BrGeometryManager from BRAT. I'll see if I'll modify
    the implementation of detector geometries in TED or if I'll just leave it
    as it is, i.e. in TED with name BrVizGeomManager.
    
    > > > > BrTriggerManager		Easy managing of trigger mask
    >
    > I guess you pass a given BrEvent through the regular BRAT analysis
    > modules to do tracking, and ya di da.  Then, you might as well pass it
    > trough a BrTriggerFilter.
    >
    > Also, I see absolutly no reason why any analysis class needs access to
    > the trigger words, assuming you have filtered out the proper events
    > first.  We (Flemming, Kris, and I at least) had a long discussion on
    > that some time ago, and the consequence of that was that
    > BrEventManager was depriciated for exactly these reasons.
    >
    > > > > BrEventInputManager		Easy input of new events
    >
    > Well, it's really a cache isn't it?  Also, you could easily combine
    > the services of BrEventList and BrEventInputManager into a module or
    > something, say BrCacheEventModule and use it with bratmain as:
    >
    >   BrMainModule*       mainModule    = new ...
    >
    >   BrIOModule*         inputModule   = new ...
    >   mainModule->AddModule(inputModule);
    >
    >   BrTriggerFilter*    triggerFilter = new ...
    >   mainModule->AddModule(triggerFilter);
    >
    >   BrCacheEventModule* cacheModule   = new ...
    >   cacheModule->SetMaxEvents(10);
    >   mainModule->AddModule(cacheModule);
    >
    
    I removed BrTriggerManager and BrEventInputManger from BRAT, at least
    while I try your suggestions. I also removed everything related to
    BrTriggerManager from BrEventList so that it can be used by others. The
    internal storage of events in BrEventList I'll test with different
    solutions as I move on.
    
    I guess that's all for now.
    
    Best wishes from
    Jens Ivar
    
    -- 
    Jens Ivar Jřrdre
    e-mail:    JensIvar.Jordre@fi.uib.no
    usually:   Dep. of Phys., Allégt. 55, N-5007 BERGEN, NORWAY
    currently: Bldg 510D, P.O.Box 5000, Upton, NY 11973-5000, USA
    phone:     +1-631-344-4223
    



    This archive was generated by hypermail 2b30 : Fri Aug 17 2001 - 13:14:47 EDT