From gwatts@phys.washington.edu Thu Apr 29 17:13:53 2004
Date: Fri, 16 Apr 2004 17:46:04 -0700
From: Gordon Watts <gwatts@phys.washington.edu>
To: Gustaaf Brooijmans <gusbroo@fnal.gov>, Serban Protopopescu <serban@bnl.gov>
Cc: d0dfwg@fnal.gov
Subject: RE: Questions at the ADM

Hi,
  I agree completely with Gustaaf here! There are two factors that go
into this decision. As Serban mentioned, speed is one. The other is
making sure an algorithm is the same from person to person to person.
Having everyone run a common (or complex) algorithm is bound to lead to
bugs in our analysis.

  bID is a perfect example of this. It is complex. Requires 5-7
packages, and you need to have the right versions checked out to make it
work properly!

  On the other hand, that it is applied in the root maker or in
d0correct doesn't really matter. And having something like bID write out
a simple edm chunk with the results is just fine.

      Cheers,
          Gordon.


-----Original Message-----
From: Gustaaf Brooijmans [mailto:gusbroo@fnal.gov] 
Sent: Friday, April 16, 2004 9:53 AM
To: Serban Protopopescu
Cc: d0dfwg@fnal.gov
Subject: Re: Questions at the ADM

> I see no compelling reason for an algorithm applied after d0correct
> that is fast compared  to unpacking to migrate (except if the sum of
all
> is substantial compared to unpacking).

I do - the fact that it gets applied centrally and nobody makes a
mistake
when running it!  We wouldn't have d0correct if all correction could be
made at reco time...

Gustaaf