NA Digest Sunday, August 15, 1993 Volume 93 : Issue 30

Today's Editor:

Cleve Moler
The MathWorks, Inc.

Submissions for NA Digest:

Mail to

Information about NA-NET:

Mail to


Date: 10 Aug 1993 07:00:06 -0800 (PST)
Subject: Teaching Symbolic Computations

Teaching Symbolic Computations

I would appreciate feedback about using computer algebra software (CAS),
in teaching Numerical Analysis to undergraduate students.

Is there an interest in teaching CAS in Numerical Analysis?

Should there now be a component on using software such as MAPLE, Derive,
MACSYMA or Mathematica in the student computer laboratory exercises?

Should students be taught to use CAS to derive some of the formulas involved
in numerical analysis? (Such as interpolation, integration...)

Should any computations be done in rational arithmetic compared to decimals?

If you have had experience using CAS or are interested in it,
I would appreciate your correspondence.

Best Regards,

John Mathews
Dept. of Mathematics
Calif. State Univ. Fullerton
Fullerton, CA 92631


From: Kermit Sigmon <>
Date: Tue, 10 Aug 93 14:46:23 EDT
Subject: MATLAB Primer


This is to remind users of the MATLAB Primer that the latest
version is available via anonymous ftp from: in the directory pub/matlab

The following files are currently available:

primer.tex Plain TeX source of the second edition PostScript file for the second edition

primersp.tex Plain TeX source of the Spanish version of the 2nd ed. PostScript file for the Spanish version of the 2nd ed.

You are advised to download the appropriate file since it contains
a new copyright clearance statement as well as a few minor revisions
to the one you probably now have. In fact, you are advised to
download anew the latest version each term since minor improvements
and corrections may have been made.

While the second edition of the Primer is based primarily on version
3.5 of MATLAB, it is compatible with version 4.0 with the fundamental
differences noted. An edition of the Primer based on version 4.0 is,
however, under development. You will be informed here when is becomes

As many of you know, the MATLAB Primer was written to help students
begin to use MATLAB. It is intended to serve as an introduction to
and *not* a manual for MATLAB. While its primary purpose is for use
in courses which require use of MATLAB, it could, of course, serve
as an introduction to MATLAB for others. It is intended to be
distributed via a local copy center.

Kermit Sigmon Department of Mathematics University of Florida


From: Zahari Zlate v<>
Date: Wed, 11 Aug 93 15:23:13 +0200
Subject: Garcia's Request on Air Pollution

In the last issue of the NA Digest, Gustavo Montero Garcia
sent a request about finite element discretization of air
pollution models (together with some other finite element
applications). We are also interested in these topics.
Therefore we shall be very grateful if all answers sent
to Garcia are also forwarded to us. Our e-mail addreses
are (Beny Neta)

and (Zahari Zlatev)

Thank you very much in advance.


From: Jesse Barlow <>
Date: Fri, 13 Aug 1993 10:01:20 -0400
Subject: Address Changes for Jesse Barlow and Hongyuan Zha

Address changes for Jesse Barlow and Hongyuan Zha.

The computer science department at Penn State has been combined
with the computer engineering program and moved to a new building.
The new addresses are

Department of Computer Science and Enginnering
220 Pond Laboratory
The Pennsylvania State University
University Park, PA 16802-6106
FAX: 814-865-3176
Dept. Phone: 814-865-9505

Office addresses and phones

Jesse Barlow
312 Pond Lab

Hongyuan Zha
307 Pond Lab

The phone numbers are unchanged.

Jesse Barlow.


From: Dugald Duncan <>
Date: Wed, 11 Aug 93 09:21:04 +0100 (BST)
Subject: Scottish Computational Mathematics Symposium 1993

10am-5pm Monday 13th September
Hume Tower, University of Edinburgh, Scotland.


* V. Stanley (NAG Ltd.)
Introduction to the AXIOM computer algebra system.
(AXIOM will be demonstrated during the breaks.)

* J.C. Eilbeck (Heriot-Watt University)
Quantum mechanics, Mathematica and the eigenvalues of
tridiagonal matrices.

* A.J. Wathen (University of Bristol)
The convergence of preconditioned Krylov--subspace iterations
for the Stokes and Navier-Stokes equations.

* K.W. Morton (University of Oxford)
Cell vertex methods for steady systems of conservation laws.

* D.J. Higham (University of Dundee)
Global errors in numerical time-stepping.

POSTERS: Space for posters is limited so please contact the organisers
at the address below as soon as possible (and certainly before the end
of August) if you intend to submit one.

in the series organised by D.B. Duncan (Heriot-Watt University) and
D.M. Sloan (University of Strathclyde) with the aim of bringing
together mathematicians and others who develop and/or use computer
algorithms to solve mathematical problems. The meetings are open to
everyone interested and this one is supported financially by the London
Mathematical Society.

REGISTRATION & INFORMATION: To register, send a cheque for 15 pounds
payable to Heriot-Watt University and include your address so that we
can contact you. The fee includes tea & coffee breaks and lunch. If
possible, please register before the end of August so that we can
arrange the catering. To register or to obtain further information
please contact:
Dr. Dugald Duncan, Department of Mathematics,
Heriot-Watt University, Edinburgh, EH14 4AS.


From: Zahari Zlatev <>
Date: Wed, 11 Aug 93 14:37:35 +0200
Subject: Workshop on ODE's Arising in Chemistry

Here is the announcement for an informal workshop concerning some
important topics that are of interest for air pollution modellists,
chemists and numerical analysts working in the field of numerical
solution of stiff ODE's. The National Environmental Research
Institute, where the workshop will take place, is located
not far from Copenhagen.

Best regards,
Zahari Zlatev


Date: November 19 1993
Place: National Environmental Research Institute
Frederiksborgvej 399, DK-4000 Roskilde, Denmark

The chemical reactions are an important part of any air pollution model.
The reliability of the concentrations calculated by the model depends strongly
on the accuracy of the numerical methods used in the treatment of the chemical
reactions. The requirement for accurate methods is absolutely necessary, but
unfortunately not sufficient. The systems of ordinary differential equations
(ODE's) that appear when the chemical reactions are described mathematically
are normally very stiff and very badly scaled. Therefore not only should the
numerical methods be accurate, but it should also be possible to carry out the
integration by using large time-steps. Finally, the discretization of the air
pollution models leads to huge computational problems; it is not uncommon that
millions of equations have to be treated numerically during many time-steps.
This fact implies a third requirement: the numerical methods should be
suitable for high-speed computers. Thus, the modern large air pollution models
can be treated numerically only if accurate and fast numerical methods that
perform well on parallel and vector machines are used in the chemical part.

The main purpose with this workshop is to bring together physisists
working with large air pollution models and specialists in solving stiff
ODE's. The hope is that in the discussions some promising classes of
numerical methods for the chemical parts of the existing large air pollution
models will be selected and further experiments with these methods will be

A limited number of papers will be selected for the workshop. If you want
to present a paper and/or to attend the workshop contact (no later than
October 25 1993):

Zahari Zlatev
fax: +45 4630 1212,
phone: +45 4530 1149

There will be no conference fee, but the number of participants will be
limited (to no more than 30, and preferrably less than 30). We can assist you
in the reservation of hotels.

A tentative program with the talks at the workshop will be sent to the
participants in the beginning of November 1993.


End of NA Digest