I will add comments later to the specifics of the de/dx (as well as other issues). Ones people modify datastaructures that may be written onto root files, you MUST add a streamer method if not already there and update the version-number in the ClassDef macro. In general pointers should NOT be persistent i.e. use the / / ! in the comment for the def since otherwise the automatic root streamer may or may not output the object. Ownership considerations are quit important here. (This is in part why one uses IDs rather than pointer but at expense of time..) Flemming > Hi > > In BrTPCHitCluster I have added the pointer to the BrTPCCluster, so there > one does not have to match. > It would of course be nice if the web page was updated more often so that > this could be seen on the web. > > I like the idea with a pointer in BrDetectorHit and using the file://! option > so that it is not streamed out. > > Cheers, > Peter > > :-) --------------------------------- )-: > |Peter H L Christiansen aka PAN @ NBI | > |EMAIL : pchristi@nbi.dk | > |OFFICE : Tb1 @ NBI | > |PHONE : 353 25269 | > |SNAIL : Hans Tavsensgade 35, 4th | > |PHONE : 35 349336 | > :-D --------------------------------- \-: > > On Mon, 13 Nov 2000, Jens Ivar Jordre wrote: > > > On Mon, 13 Nov 2000, Claus O. E. Jorgensen wrote: > > > > > > > > Hello everybody > > > > > > I would like to start a discussion on how to get > > > dEdx information for the local tracks. Right now > > > the only way to do it is to get the BrDetectorHits > > > and look through all the BrTPCHitClusters to see > > > which hit matches which cluster and then get the > > > energy from the cluster. As I see it we have 3 > > > options: > > > > > > 1) Keep it the way it is. > > > > > > 2) Let the BrDetectorHit have an Float_t dEdx > > > member and corresponding get and set methods. > > > > > > 3) Let the BrDetectorHit have a void pointer to > > > the BrTPCHitCluster (or the BrDCCluster) and > > > corresponding get and set methods > > > (void* GetMotherCluster() and SetMotherCluster(*void). > > > > > > As I see it the latter is the best solution. It > > > will make it easy to get the energy, and we'll > > > avoid having a not-defined dEdx for detector hits in > > > the DCs (these detectors don't measure the energy, > > > as far as I'm informed). > > > > > > Any comments? > > > > Here comes the first one. > > > > I agree with you. The 3rd approach resembles what is implemented for > > tracks, where BrDetectorTrack has a pointer to its mother BrLocalTrack. > > > > Also present in the current version of BRAT is ID inheritage, i.e. new > > instances of BrDetectorHit inherit the ID from their mother > > BrTPCHitCluster, which again inherits ID from its mother BrTPCCluster. > > This may be what you refer to as matching hits and clusters. But it is > > arather time consuming process to match ID compared to getting the mother > > data structure from a pointer. I therefore support your view. > > > > JI > > > > --- > > Jens Ivar Jřrdre, Dep. of Phys., Allégt. 55, N-5007 BERGEN, NORWAY > > room 521, e-mail: JensIvar.Jordre@fi.uib.no, phone: (+47) 55 58 27 92 > > > > > > >
This archive was generated by hypermail 2b29 : Mon Nov 13 2000 - 10:57:42 EST