Re: BRAT 1.13.2

From: Christian Holm Christensen (cholm@hehi03.nbi.dk)
Date: Tue Jan 30 2001 - 11:03:17 EST

  • Next message: Bjorn H Samset: "Error in BrTPMClusterVertexModule"

    Hi Kris et al, 
    
    On Tue, 30 Jan 2001 08:35:43 -0600
    hagel@comp.tamu.edu wrote
    concerning ": Re: BRAT 1.13.2":
    > In regards to Steve's remark at the end.  Not only is it ill
    > advised.  It was also STRONGLY advised against.
    
    As I've already explained to Steve, the reason for this quick commit,
    was motivated by the need for a uniform way of determing centrality. 
    
    The current situation is, that the constant cuts defined by J. H. are
    only valid in |Vz - <Vz>| < 3-7cm, once you go further way from the
    nominal center, then the 5% cut for example drops with around 10%. 
    
    A vertex cut of |Vz - <Vz>| < 3-7cm is just not good enough for doing
    analysis in FFS. I believe that Ian has a cut of approximate 
    |Vz - <Vz>| < 15cm or maybe even higher. 
    
    Now, I agree major changes - like the one in question - should not be
    done too quickly. However, given the current deadline for p/pbar
    paper, my tight schedule this week, and the fact that the old code was
    far too messy to be understood, I opted for a quick commit. In
    retrospect, I agree I could have been more forthcoming in this
    matter, and I apologize.
    
    However, you must appricate the frustation to see that the old code
    hadn't been revisted since August last year, and the it didn't
    deliever the functionality needed urgently. If one has new code, one
    should commit it, even though one may not be 100% sure of it. 
    
    Again, as I've previously pointed out, code doens't need to be 100%
    error-free when commited to BRAT. in fact, commiting while still in
    iteration may help the developer, since it allows others to test,
    debug, comment, etc. the code.
    
    In summary, I did the changes 'cause they were needed, and because the
    multiplicity people hadn't delievered anything for so long. Again, I
    apologize for my rashness. 
    
    Yours, 
    
    Christian  -----------------------------------------------------------
    Holm Christensen                             Phone:  (+45) 35 35 96 91 
      Sankt Hansgade 23, 1. th.                  Office: (+45) 353  25 305 
      DK-2200 Copenhagen N                       Web:    www.nbi.dk/~cholm    
      Denmark                                    Email:       cholm@nbi.dk
     
    



    This archive was generated by hypermail 2b29 : Tue Jan 30 2001 - 11:03:47 EST