1. Introduction
---------------
All results and input files are available at 

  http://www.slac.stanford.edu/xorg/hfag/semi/summer04/summer04.shtml

In the  following (in  no particular order),  the updates for summer 2004
with respect to winter conferences 2004 are discussed.

As usual, some of the error breakdowns can be found in

  http://www.slac.stanford.edu/xorg/hfag/semi/summer04/summer04.xls
spreadsheet. This is relevant only for the branching fraction B -> X l
nu and |Vub|. 

Both Belle and BABAR have shown  many new measurements of Vub based on
inclusive semileptonic B decays; all  are evaluated based on the shape
function parameters as determined with the BELLE b-> s gamma spectrum.


2. Changes to common parameters
-------------------------------

We use the updated input parameters for mixing and lifetimes. For more
information, please see

  http://lepbosc.web.cern.ch/LEPBOSC/combined_results/mar_20b
  
and

  http://lepbosc.web.cern.ch/LEPBOSC/combined_results/mar_20b/changes.txt

Some of the D branching fractions have changed with the publication of
PDG 2004. All new values are consistent with the old values within the
errors.

For f00 = BF(Upsilon(4S)->B0B0) we now take the value computed in 

  http://www.slac.stanford.edu/xorg/hfag/osc/beijing_2004/index.html

Assuming  non-BBbar decays of  the Upsilon(4S)  to be  negligible, the
precision of this averaged  quantity is improved considerably compared
to previous versions. This is  most obviously reflected in the chi2 of
the average for B0->D*+lnu, both for the branching fraction and |Vcb|.


3. Average of |Vub| from inclusive measurements on Upsilon(4S)
--------------------------------------------------------------

A substantial  change has  occurred here with  the usage of  the BELLE
photon spectrum measurement in B ->  s gamma decays.  This is used for
the determination of  the central values and errors  of shape function
parameters, used  in virtually  all |Vub| determinations  in inclusive
charmless  semileptonic B  decays. The  new shape  function parameters
have  substantially smaller errors  (x2) than  the ones  determined by
CLEO and lead to an increase of Vub (because of their central values).

BELLE has a  new determination of |Vub| based on  events tagged with a
fully reconstructed hadronic decay of a B meson. This analyis is based
on a phasespace selection with mX < 1.7 GeV and Q2 > 8 GeV2. In
addition, it has re-evaluated its previous two measurements. 

BABAR has also re-evaluated its mX < 1.55GeV analysis; note that there
is  a typo  in the  conference paper,  it has  been corrected  for the
average (many thanks to Tadao Nozaki for calling our attention to this
point).   In addition, BABAR  has three  new |Vub|  determinations: 1)
electron endpoint spectrum p > 2.0 GeV; 2) mX < 1.7GeV and Q2 > 8GeV2;
3) (El,Q2) cuts.


More details of the averaging are provided in the writeup.


4. Average of B0 -> D* l nu
---------------------------

The DELPHI measurement has been published since the last update, and
the BABAR measurement is close to submission. 


5. |Vub| from exclusive measurements on Upsilon(4S)
---------------------------------------------------

This is  still only a table and  not an average. BABAR  has updates on
the branching fraction measurements with events measured on the recoil
of a fully reconstructed hadronic B decay (BRECO) and -- for the first
time -- on the recoil of a partially reconstructed B decay.

BELLE has a  measurement of B ->  pi l nu, differential in  Q2, on the
recoil of a semileptonic B decay  and extract Vub in the high-Q2 range
where  lattice QCD  (unquenched, preliminary)  formfactor calculations
are availble.


6. |Vcb| and HQE parameters from inclusive measurements
-------------------------------------------------------

Moments of  various distributions have been measured  by BABAR, BELLE,
CDF, and CLEO. 


7. Conclusion
-------------

Your critical comments will be much appreciated. Please send them to