I implemented basically all the changes that were discussed in order to be
able to carry the analysis as discussed on monday.
The recipe is in
http://www.slac.stanford.edu/BFROOT/www/Physics/CP/beta/doc/postCP/tags-anal11
(with the modified tags marked as new)

The new definition of the ntuple is in
http://www.slac.stanford.edu/BFROOT/www/Physics/Analysis/AWG/EHBDOC/SemiExcl/ntpinfo
together with some relevant notes from this mail. Note that the structure
changed in a substantial way, the main reason being that it was too tough
for me to do the modifications starting from a patch of sin2b stuff (I was
also afraid to screw up the sin2b side).
The majour change is that there is no more a tagging block, but the trk
and gam block, which are pointed at by the D(*),Ks,pi0 etc etc ...  lists.
Of course this means that almost all variables changed name...

This recipe runs and fills the ntuple with something. A lot needs
to be done to debug its content

Here are the improvements (including comments when checks need to be
done):
1) we dump all Pi0, including the merged pi0s. We need to figure out if
the link for the photons in this case (the merged ones) works (and how to
deal with them)
2) we dump all Ks (both pi+pi- and pi0 pi0) in the same block. You can
separate among them with the d1kslund variable (211 or 111). We will be
able to veto also Ks->pi0pi0 from now on (but cuts need to be studied)
3) we dump all Charged Tracks. Their mass is set to the pion one, if you
want another one you need to assign it explicitly (based on selector
outputs)
4) in the "gam" block we dump both the CalorNeutral (emc only neutrals)
and the NeutralHad (IFR only neutrals). You can distinguish
 among them with the ecalgam variable(ecalgam>0 for photons). The match
among them is not done
(i.e. if a photon deposits in the IFR it happears as two particles).
The candidates with ecalgam<0 are to be used for Klong studies but they
need to be removed from the total energy until we understand them.
5) words are added to indicate if a track/photon belongs to a reco
candidate (this was the case also in the previous ntuple, but it required
explicit work here)
6) 4-mom covariance matrix of the B candidate (before the mass constraint)
was added
7) momentum of the B candidate after mass constraint is saved (this should
be used instead of the regular one in order to improve M_v^2)
8) the thrust (abs value and direction) of the X system was added [tracks
only]
9)B counting info was removed
10) Ifr info (including the shape of the cluster layer by layer ) was
added for the neutrals
[I would like to stress again that all neutrals with ecalgam>0 do not have
IFR by construction: a matching algorithm needs to be implemented at
ntuple level]
11) only events containing at least one B are dumped (I guess this was
done by the IbuFilter, but the option to turn this on or off came for
free)
12) semiexcl mode label and a-priori purities were added
13) recoil vertex information was filled both for the lepton and the X
vertex. I am a bit nervous about the possibility that the identified
lepton at Beta level might not be the same at analysis level.
14) recoil tag info was dropped

A very small example, just to allow you to see if variables are filled is
in ~rfaccini/scra/test/hbook/test_bak.hbook

This is meant to be the most comprehensive ntuple for the studies we want
to do. Please scream and shout asap if there are missing info. Please
complaint gently if things do not work
	ciao
	ric