NA Digest Monday, August 2, 1993 Volume 93 : Issue 28

Today's Editor:

Cleve Moler
The MathWorks, Inc.

Submissions for NA Digest:

Mail to

Information about NA-NET:

Mail to


From: Andreas Griewank <>
Date: Sun, 25 Jul 93 14:11:55 CDT
Subject: Change of Address for Andreas Griewank

Dear Colleagues,

I have taken up a postion at the Technical University
in Dresden, Eastern Germany, and will leave Argonne on August 1.
Please ciculate my new address:

Prof. Andreas Griewank
Technical University Dresden
Institute of Scientific Computing
Mommsenstr. 13
D-O1062 Dresden

Phone: 49-351-463-4187
Fax: 49-351-463-7114



From: tmanteuf@copper.Denver.Colorado.EDU (Manteuffel Tom)
Date: Tue, 27 Jul 93 07:28:57 -0600
Subject: Change of Address for Tom Manteuffel and Steve McCormick

Change of address for Tom Manteuffel and Steve McCormick


Tom Manteuffel
Program in Applied Mathematics
Campus Box 526
University of Colorado at Boulder
Boulder, CO 80309-0526 * (303)492-5199 office
* -4668 secretary * -4066 fax
* 444-0684 home

Steve McCormick
Program in Applied Mathematics
Campus Box 526
University of Colorado at Boulder
Boulder, CO 80309-0526 * (303)492-0662 office
* -4668 secretary * -4066 fax
* 442-0724 home
* 442-8191 home fax


From: Choi-Hong LAI <>
Date: Tue, 27 Jul 93 12:07:02 BST
Subject: A Report on Domain Decomposition

Centre for Numerical Modelling & Process Analysis Report:-

NMA001 (July 1993) An overview of domain decomposition methods. C - H Lai

Please mail with your mailing address to obtain
a copy. We are about to fix our ftp in the department, so that NMA series
can be obtained by anonymous ftp. I will let you know as soon as it is

Choi-Hong LAI
School of Math Stat & Comp
University of Greenwich
Wellington Street
London SE18 6PF
United Kingdom
Tel : +44-81-316 8712
internet :
na-net :


From: Trini Flores <>
Date: Wed, 28 Jul 93 14:45:58 EST
Subject: SIAM Conference on Linear Algebra in Signals, Systems, and Control

Third SIAM Conference on Linear Algebra in Signals, Systems,
and Control
University of Washington
Seattle, Washington
AUGUST 16-19, 1993


To register, please contact SIAM now.

Telephone: 800-447-SIAM
Fax : 215-382-9800


From: John Prentice <>
Date: Wed, 28 Jul 93 16:04:55 MDT
Subject: Looking for the MODFLO Groundwater Code

We are interested in doing some work with the MODFLO code, converting it
to 2 other languages (Fortran 90 and C++) and modifying some of the
methodology. It is a strictly IR&D project, though we will make the
code freely available as a part of the Quetzal Computational Toolbox
if things would out well. I have a version of MODFLO that had some
minor changes made to it by Tecsoft, Inc. and then resold as their
product. Whether this version is public domain or not is unclear to
me, but given that ambiguity, we would like to get a USGS version
that is up to date but completely public domain. Does anyone
know if MODFLO is available via anonymous ftp from any sites or
failing that, can anyone point me to who to contact at the USGS
about obtaining a copy?

Many thanks!


Dr. John K. Prentice
Quetzal Computational Associates
3200 Carlisle N.E., Albuquerque, NM 87110-1664 USA
Phone: 505-889-4543 Fax: 505-889-4598 E-mail:


From: John Prentice <>
Date: Wed, 28 Jul 93 16:17:54 MDT
Subject: Looking for Access to a KSR-1

A couple of us at Quetzal Computational Associates are building a state of
the art code for modeling the flow of contaminants in variably saturated
porous media. This is an area that has received comparably little
attention in the groundwater contaminant transport community and our
code should fill an important need. The model is fully three dimensional,
which means that it is most suitable for a supercomputer. We will shortly
submit an application to the ACL at Los Alamos for CM-5 time, but we are
actually a bit more interested in the developing the code on and for the
KSR system.

We anticipate doing most of the development over the next 6 months, but
would like access for up to a year if possible. We do not anticipate
requiring an exceptional amount of computer time however, roughly
10-20 hours per month maximum and more likely a great deal less during
the first half of the project. Our needs for computer time will
be a bit greater in the second half of the project when we are performing
validation runs, the first half will mostly be code development, so we
will be doing compiling and some small tests. When this code is complete,
it will be added to the Quetzal Computational Toolbox, which is a
suite of scientific codes that we have either collected or developed and
which are available to any qualified users free of cost (we make our
money by consulting, either from the development of codes for organizations
or from using them to solve problems for clients. Where possible, we
therefore make our codes freely available to other users).

If anyone can offer us an account on a KSR or if you can point us to who
to talk to about an account, I would very much appreciate it.

Many thanks!


Dr. John K. Prentice
Quetzal Computational Associates
3200 Carlisle N.E., Albuquerque, NM 87110-1664 USA
Phone: 505-889-4543 Fax: 505-889-4598 E-mail:


From: John Prentice <>
Date: Fri, 30 Jul 93 12:43:57 MDT
Subject: Soliciting Views on the Scientific Method

I am writing up a little paper for secondary school science fair students,
discussing how scientific research is done by real scientists. There
is a shocking gap between what students are taught about how science is
done and how we really do it. Alot of the confusion revolves around the
teaching of the so-called scientific method. The scientific method that
kids learn is pretty far removed from how real scientists do research
and it leads to all sorts of curious misinterpretations. If you want
a wakeup call to the problems in our schools, for even the best of
students, go judge a science fair.

This little paper is an attempt to present a modern view of scientific
research, from the perspective of real scientists doing real science.
I would like to include in the paper paragraphs written by scientists
in various disciplines, giving their view of what the scientific method
is and how real research is done. Here is your chance to get your
2 cents in about how the real thing works, as opposed to how some
education major writing a textbook thinks it works. If you would be
interested in helping, what I need is a single short paragraph giving
your ideas. I will edit these paragraphs for spelling, grammer, etc...,
but I will not change their content. I will be happy to send the
edited version back to you for your approval, if I decide to use it.
Please keep the paragraph fairly short, I want this paper to be
something that kids and teachers will read and if it becomes a long
scientific dissertation, that won't happen. Please put your name at
the end of the paragraph, and add your title and organizational
affiliation (if appropriate). If you want, put your e-mail address

I very much appreciate any help from the scientific research community
in this venture.


John K. Prentice
Quetzal Computational Associates
3200 Carlisle N.E., Albuquerque, NM 87110-1664 USA
Phone: 505-889-4543 Fax: 505-889-4598 E-mail:


From: Horst D. Simon <>
Date: Thu, 29 Jul 93 12:20:06 -0700
Subject: Distributed Computing Workshop at NASA Ames

Workshop at NASA Ames Research Center, Moffett Field, California
October 18 - October 20, 1993
sponsored by NAS Systems Division


Distributed computing, in particular computing on networked workstations,
has become widely used in the last years for solving large scale scientific
problems. The rapid development of microprocessor performance and the
availability of new tools for distributed computing has resulted in many
exciting results, demonstrating supercomputer performance on networked

In this workshop we will assess the state of the art of distributed
computing, and evaluate its potential for solving the grand challenge
applications of interest to researchers in the computational aerosciences.
The three day workshop will feature invited lectures, contributed papers,
vendor sessions, a panel discussion, and ample opportunity for discussions
and information exchange.

The primary focus will be on actual experiences on clusters of
workstations in the computational aerosciences. Presentations
on hardware and software tools have been selected in order to provide a
good overview on recent developments for participants new to distributed
computing. The workshop will provide the computational aerosciences
community with an excellent opportunity to get acquainted with distributed
computing. There will be discussions on what are the requirements of the
aerospace community and descriptions of what NASA and others are doing to
meet those requirements.


Bill Kramer (NASA)
Horst Simon (CSC/NASA)
Rod Fatoohi (CSC/NASA)
Marcia Redmond (Sterling Software/NASA)


The conference includes invited talks by

H. T. Kung - Harvard University
Dennis Duke - Florida State University
Vaidy Sunderam - Emory University

and presentations on hardware developments by representatives from
IBM, HP, SGI, etc., programming tools (LINDA,PVM, EXPRESS, P4, etc.)
and job scheduling tools, as well as overviews of distributed
computing activities at major aerospace companies and NASA centers.


Contributed papers from participants are invited.
Papers about implementations and results of large scale applications,
on networks of workstations are especially welcomed.
We are also particularly interested in methods of managing work on networks.
Those interested in presenting a paper must submit an abstract of
approximately 300 words in length.
The deadline for submitting abstracts is August 20, 1993.
Abstracts will be reviewed and authors notified of their acceptance by
September 10, 1993. Please submit abstracts to:
Dr. Rod Fatoohi
Mail Stop T045-1
NASA Ames Research Center
Moffett Field, CA 94035-1000
Electronic submissions of abstracts are encouraged.


There is a $65 registration fee for all participants. A reduced registration
fee for NASA civil servants is $40. The registration fee includes
refreshments during the workshop, dinner on Monday night, a reception
on Tuesday night, and two lunches.

Registration deadline for US citizens and permanent residents
is October 1, 1993. The deadline for non US Citizens
is August 20, 1993. In order to register, please mail the attached registration
form with the registration fee to
Marcia Redmond
Mail Stop 258-6
NASA Ames Research Center
Moffett Field, CA 94035-1000
Make checks payable to Sterling Software/dcaa

Advance registration is required. The workshop will be held at NASA Ames
Research Center. Any questions can be directed to Marcia Redmond,
(415) 604-4373 or send e-mail to


From: Pierre Manneback <>
Date: Fri, 30 Jul 93 10:58:30 +0200
Subject: Workshop on Parallel Computing, Mons, Belgium

The Parallel Information Processing Laboratory of Faculte Polytechnique
de Mons, Belgium, is pleased to announce


September 23 & 24, 1993

The programme will consist in invited lectures only.
Each lecture will be split into a mini-tutorial (part 1) and an advanced topic
(part 2).

The list of Speakers is the following:

J.L. LARRIBA-PEY, UPC, Barcelona, E
R. LUELING, Univ. of Paderborn, D
S.G. PETITON, ETCA, Arcueil, F
F.A. RABHI, Univ. of Hull, U.K.
D. TRYSTRAM, IMAG, Grenoble, F.

This workshop is organized with the support of A.I.Ms
(Association des Ingenieurs de la Faculte Polytechnique de Mons) and of
the Ministere du Developpement Technologique pour la region Wallonne.

The participation fee is 1000 BEF (about 30 U.S. $).


Thuersday 23 September : "Parallel Architectures and Languages"

Optimal communications on network of processors D.TRYSTRAM
Parallel architecture and their efficient use R. LUELING.
Functional languages and parallelism F.A. RABHI.

Friday 24 September : "Parallel Numerical Linear Algebra"

Data parallel linear algebra methods S.G. PETITON.
Tridiagonal systems on parallel computers J.L. LARRIBA-PEY.
Load balancing in parallel combinatorial optimization R. LUELING.

Organization Committee

Prof. P. DUFOUR, Univ. Mons-Hainaut
Prof. G. LIBERT, PIP, Polytech. Mons
Dr P. MANNEBACK, PIP, Polytech. Mons

For hotel reservations, more information and registration,

please contact

Dr Pierre Manneback
Laboratoire P.I.P.
Faculte Polytechnique de MONS
rue de Houdain 9

tel: +32 65 374050
fax: +32 65 374500


From: SIAM <>
Date: Fri, 30 Jul 93 13:02:30 EST
Subject: Contents, SIAM Computing

Volume 22, Number 5

889 Finding a Smallest Augmentation to Biconnect a Graph
Tsan-Sheng Hsu and Vijaya Ramachandran

913 Different Modes of Communication
Bernd Halstenberg and Rudiger Reischuk

935 Suffix Arrays: A New Method for On-Line String Searches
Udi Manber and Gene Myers

949 Randomizing Reductions of Search Problems
Andreas Blass and Yuri Gurevich

976 An Optimal Algorithm for the Maximum Three-Chain Problem
Ruey-Der Lou and Majid Sarrafzadeh

994 Random-Self-Reducibility of Complete Sets
Joan Feigenbaum and Lance Fortnow

1006 Learning Binary Relations and Total Orders
Sally A. Goldman, Ronald L. Rivest, and Robert E. Schapire

1035 Drawing Graphs in the Plane with High Resolution
M. Formann, T. Hagerup, J. Haralambides, M. Kaufmann,
F. T. Leighton, A. Symvonis, E. Welzl, and G. Woeginger

1053 Probabilistic Analysis of Disjoint Set Union Algorithms
Bela Bollobas and Istvan Simon

1075 A Pseudorandom Oracle Characterization of BPP
Jack H. Lutz

1087 Polynomial-Time Approximation Algorithms for the Ising Model
Mark Jerrum and Alistair Sinclair


End of NA Digest