From benoit.clement@IReS.in2p3.fr Tue Nov 16 17:06:48 2004
Date: Tue, 16 Nov 2004 14:46:07 +0100
From: Benoit CLEMENT <benoit.clement@IReS.in2p3.fr>
To: Herbert Greenlee <greenlee@fnal.gov>
Subject: Re: Common Analysis Format Proposed Contents (fwd)

Hi Herb,

  I went through the recommendation on CAF paper. a few questions/remarks
arise regarding the b-tagging part :
Regarding positive and negative tag : these are not exclusive notions, ie a
jet can be positivly and negativly tagged
(a more correct wording would be tagged using tracks(2nd vertex) with
positive/negative IP(decay length) signifiance)
There should therefore be 2 separate words for negative and positive tag in
TMBBTag. Related to that, there should be for each jet both positive and
negative JLIP probability in the TMBBTagJLIP class.

  Another issue regarding JLIP is about specific track related variables.
JLIP classify tracks in categories, computes an IP error correction and
finally computes a probability per track.
These 3 values (or at least the last one) should be kept in CAF. I would
naively suggest a new class TMBBTagTrack containing :
      - Track          :  TRef to TMBTrack
      - JLIPprob       :  double
      - JLIPcategory   :  int (or method)
      - JLIPcorrection :  double
      + any other variables useful for other btag algorithms !
The TMBFilteredTracks array should be filled with such objects.

  Regarding secondary vertices, I suppose specialist will comment, but it
seems to me that vertices used for loose/medium/tight
are reconstructed with different cuts, so there will probably need to have
the SVT information duplicated for each working point.

  Finally, I'm not sure I clearly understand what is meant by MC TRF, Data
TRF, and ScaleFactor. Do you intend to have the jets TRF weights stored into
the tree ?
What about the errors associated to these weights ?

Hoping these comment will be helpful,

  Regards,
   Benoit

----- Original Message ----- 
From: "Herbert Greenlee" <greenlee@fnal.gov>
To: <d0_b_id@fnal.gov>; <d0_b_id_work@fnal.gov>
Sent: Monday, November 15, 2004 10:04 PM
Subject: Common Analysis Format Proposed Contents (fwd)


> To the bid group,
>
> Please read and comment on the Common Analysis Format content proposal
> document described below.  The content proposal contains a section on bid.
>
> Herb
>
>
>
> ---------- Forwarded message ----------
> Date: Fri, 12 Nov 2004 10:19:03 -0600 (CST)
> From: Herbert Greenlee <greenlee@fnal.gov>
> To: d0-conveners@fnal.gov
> Cc: D0 Data Format Working Group <d0dfwg@fnal.gov>
> Subject: Common Analysis Format Proposed Contents
>
> Hello Physics and Algorithm Conveners,
>
> A draft content proposal (version 1.0) for the proposed Common Analysis
> Format is available on the d0dfwg web page at the following url:
>
>
http://www-d0.fnal.gov/Run2Physics/working_group/data_format/content_v10.pdf
>
> By now, content proposals exist for most branches, although a few are
> still missing.  Please ask your groups to review the proposed contents and
> either give your approval or suggest changes.  Comments can be sent to the
> data foramt group mailing list d0dfwg@fnal.gov.
>
> Thanks,
>
> Herb and Reiner
>