Re: Sheilding Software Inquiry

[mikeg@SLAC.Stanford.EDU (Michael P. Grissom)]

One of our long time accelerator HPs (who was in a former
life an accelerator physicist) and who is an expert in the
use of radiation transport codes (co-author of EGS4 in
fact, i.e., Dr. Ralph Nelson) has provided the following
in response to Craig's question:

>Date: Thu, 13 Jul 1995 22:07 -0800 (PST)
>Subject: Re: Sheilding Software Inquiry
>
>In-Reply-To: mikeg@SLAC.Stanford.EDU -- 07/13/95 15:46
>
>Pass on to Craig the fact that......good accelerator HPs 
>should already understand ALL of the fundamental 
>interactions and, therefore, should be able to use analytic 
>methods to estimate the problem/solution from the start !!!
>
>The bottom line is that anyone worth his/her salt in 
>accelerator health physics would NOT use a code like: EGS4, 
>KASKAD, CASIM-GEANT-FLUKA (the latter were developed from the 
>other two anyways), and any other MC code, including MCNP or 
>ITS.  What the good accelerator HP should do is to start from 
>the basics and do back-of-the-envelope calculations, followed 
>up by measurements.  Period!  MC codes can be beneficial, but 
>they can be "idiot boxes for idiots" too.  They allow one to 
>be sloppy in approach to fundamental problem solving because 
>they allow one to use a "black box" invented by someone 
>else....allowing for the coverage of one's lack of fundamental 
>knowledge about a subject.  If this is the "Accelerator Health
>Physics of the Future", it will be over my dead body.
>
>Regarding photoneutron production, the members of the SLAC 
>Radiation Physics Department are currently working on, and 
>publishing, the connection between photoproduction of neutrons 
>(and other particles) using their transport.  We photoproduce 
>with EGS4---because I wrote it and I understand it---and we 
>then transport these particles with MORSE or MCNP or FLUKA; we 
>use and support all of these wonderful MC programs (really!).  
>MCNP does do a great job of transporting neutrons/photons.  So 
>does MORSE and so does FLUKA.  ETRAN and ITS are super too.  
>To answer Craig's question about using GEANT instead of FLUKA,
>the answer is simple:  FLUKA and EGS, in my opinion, tend to 
>be more "open to user inspection and intervention".  I guess 
>it boils down to who you know in the GEANT-Development World 
>that can support you.  Mike....please pass this on to Craig 
>Harmon, who seems to have already identified the problem.  Ralph.

Having sat in on Ralph and Ken Kase's Radiation Dosimetry and
Shielding course at San Jose State University, I can assure you
that Ralph isn't kidding about expecting you to know the "basic
fundamentals" involved!  I am sure many of the graduates of SJSU's
master's program would concur.

Also, many of the codes mentioned, particularly EGS4, require
a very large amount of (appropriate) input data, which puts
certain demands on the individual to know what the problem
really entails.  This can in effect lead to "rewriting" some of 
the code to optimize for the problem at hand.  As Ralph states
above, this can be dangerous territory for the unwary!

All of these opinions are either Ralph's or mine and we share
responsibility for them, respectively!
-----------------------
Michael P. Grissom
mikeg@slac.stanford.edu
Phone:  (415) 926-2346
Fax:    (415) 926-3030