* Sheila, anything you want to show concerning the new
  notebook (version 1.1)?

The latest news on the workbook is summarized on the webpage:

http://www.slac.stanford.edu/~penguin/VubStuff/EvtVubBLNP4/index.html

I have not yet updated the workbook to Version 1.1.  
However, I did look through the new notebook, and 
then I sent the following email 
to Mr. Neubert and Mr. Paz.  Mr. Paz emailed me back 
and said he would send a detailed reply next week.  
So until then, I'm working on other stuff.

sheila

---------- Forwarded message ----------
Date: Thu, 3 Nov 2005 05:12:51 -0500 (EST)
From: Sheila Mclachlin <[log in to unmask]>
To: [log in to unmask]
Cc: [log in to unmask]
Subject: help with update to version 1.1 of B->Xulnu notebook

Hello,

I'm a grad student at BABAR, and I am trying 
to write an event generator for inclusive B->Xulnu, 
based on your notebook.

I had just gotten a generator based on 
Version 1.0 to work, and then I learned 
that there is a new Version 1.1 now.

For the generator, the only part I need 
is rate3 for B->Xulnu.

I looked through the new notebook to try 
to determine how your changes would 
affect my generator.  Here is a summary 
of what I think.  Could you please 
let me know if it is correct?

The form of the shape functions is the 
same.  But now, instead of using Lambda and b 
as input parameters, you are effectively 
using mbSF and mupisqSF as input parameters, 
and getting Lambda and b from them.  Since 
mbSF and mupisqSF are not used anywhere 
else anyway, they are really just "Lambda 
and b by another name."  So my model, 
which uses Lambda and b as input parameters, 
and doesn't use mbSF and mupisqSF at all, 
is still OK.  However, if I wanted to, 
I could make mbSF and mupisqSF the input 
parameters, instead. 

The main effect of the change is that when 
Lambda and b are the input parameters, 
then mbSF and mupisqSF are different 
for different models (eg, exp vs gauss).  
Whereas when mbSF and mupisqSF are the input 
parameters, Lambda and b are different 
for different models.  I suppose this 
might become important when users are 
trying different models to see how the  
variation affects their results.  
I assume that for this purpose the 
Version 1.1 method -- using mbSF and mupisqSF 
as input parameters -- is better.

The other big change in Version 1.1 is the 
change in the subleading shape functions.  
In Version 1.0, you had two options 
for wS: wS = w*S and wS = DS.  
Now, since it is much harder to 
do derivatives in C++ than in Mathematica, 
for my old EvtGenModel I just used 
the wS = w*S option, and filed the 
other option away as a possible future 
addition.  But now I see you have only 
the wS = DS option in Version 1.1.  
This means that either I have to 
write a differentiation program 
(to make it easier to adapt the model 
to incorporate many different shape 
functions in the future) or differentiate 
each function myself, separately (which 
is easy enough for exp and gauss at least, 
but will make it harder to add other models 
in the future).  Either way, it's kind of 
a drag.  So I am wondering: are the 
subleading shape functions in Version 1.0 
still OK?  Can I use them instead?  
Because using wS instead of DS will 
certainly speed up the program.  (It 
is very slow already due to all the 
4 unavoidable integrals.)

Also, a note to Mr. Paz, about the 
maximum-rate thing: I ran some tests and 
found that settting ratemax to 3.0 
(units of GF^2 Vub^2/(pi^3 hbar) )
works pretty well.  So you don't need to 
send me the maximum rate after all.  

Thanks for your help,

sheila