NA Digest Sunday, October 2, 1988 Volume 88 : Issue 39

Today's Editor: Cleve Moler

Today's Topics:


From: Arnold Neumaier <>
Date: Mon, 26 Sep 88 14:36:52 cdt
Subject: Basis for Linear Equalities

I have the following question to the Numerical Analysis community:

Let C be the set of n-dimensional real vectors satisfying the system
of linear inequalities Ax >= b. By a well-known theorem, every point x
in C is a convex combination of at most n+1 vertices of C (and of at
most n if x lies on the boundary).
Are there efficient algorithms to find such vertices and the corresponding

Arnold Neumaier
present address:
Mathematics Department
University of Wisconsin
Madison, WI 53706
email address:


From: Daniel Boley <>
Date: Mon, 26 Sep 88 19:03:12 CDT
Subject: Minnesota CS Department Gets New Address

We are happy to announce that the Computer Science Dept. at the University
of Minnesota has moved into a nice new big building. We share the building
with the E.E. Dept. So far, the building is not all finished, and the
computer network is incomplete. We are still suffering from moving pains,
but it should turn out to be a fine building, for which we must thank
the State of Minnesota.

Therefore, please change your records to reflect our new street address:

Computer Science Dept.
University of Minnesota
4-192 EE/CSci Building
200 Union Street S.E.
Minneapolis, MN 55455 USA

As you can see, we are still in need of a name for the building.
The phone numbers and electronic mail addresses should all be unchanged.

- Daniel Boley


From: Ken Jackson <>
Date: Tue, 27 Sep 88 09:58:23 EDT
Subject: Frontiers of Science Conference Announcement

Frontiers of Science Conference
Moscow, U.S.S.R.
8-10 December 1988.

The Frontiers of Science Conference to be held 8-10 December in Moscow,
U.S.S.R., is the sixth in a series of international conferences (two of
which were forced to be canceled) organized by the Moscow Scientific
Seminar, a group of Russian "Refusnik" scholars who hold regular
meetings in the private apartments of group members. The objectives of
these conferences are to bring together eminent scientists from diverse
backgrounds, to provide the members of the Moscow Scientific Seminar
with an opportunity to communicate with other scientists, and to
provide seminar members with access to current research results which
otherwise are denied them.

The seminar members would warmly welcome participants from outside the
U.S.S.R. If you are interested in attending, please call or write

Ms. Dorothy Hirsch,
Committee of Concerned Scientists,
330 Seventh Ave., Suite 608,
New York, NY, 10001, USA.
Telephone: 212-695-2560.

She can supply you with more information about the conference programme
as well as travel information and related advice.

(Do NOT contact me for further details about the conference. I do
NOT have access to the most up-to-date information.)


From: Bob Plemmons <>
Date: Tue, 27 Sep 88 11:22:50 EDT
Subject: Position of Department Head at North Carolina State



Applications are invited for the position of Head, Dept.
of Mathematics, starting Fall 1989. The department is in the College
of Physical and Applied Mathematics, has a faculty of 90 and offers
all levels of undergraduate and graduate programs in pure and in
applied mathematics. It administers, along with the Computer Science
Dept., the Center For Research in Scientific Computation.

North Carolina State University has an enrollment of 24,000,
and, along with Duke and UNC Chapel Hill, is one of the three research
universities adjacent to the Research Triangle Park. Computing facilities
include a variety of modern architectures, including an Alliant and a
Sequent, along with Sun personal workstations in the dept. A contract has been
signed with Cray to locate a Y-MP in the Triangle later this year.

The successful candidate should have proven research ability,
commitment to quality instruction and leadership ability.

Applicants should send a resume with names of four references prior
to January 1, 1989 to: Richard S. Patty, Search Committee Chairman, North
Carolina State University, Raleigh, NC 27695-8202. NCSU is an Equal
Opportunity/Affirmative Action Employer.


From: Tim Kelley <CTK%NCSUMATH.BITNET@Forsythe.Stanford.EDU>
Date: Tue, 27 Sep 88 10:32:06 EST
Subject: Junior Positions Available at North Carolina State

Two tenure track positions will become open July 1, 1989 at the junior level
in numerical analysis and scientific computing. Emphasis for one of these
positions will be given to finite element methods for partial differential
equations. For the other position the area is open, but expertise in
parallel computing is desirable.

Applicants must have a strong potential or record in research and instruction.
To apply, send a resume, any reprints or preprints, a thesis abstract, and
arrange to have three letters of reference sent to C. T. Kelley, Department
of Mathematics, Box 8205, North Carolina State University, Raleigh,
North Carolina, 27695-8205.
Information regarding US citizenship or residency status should be included
with the application.

Electronic inquiries should be sent to (arpa)

or to


North Carolina State University is an affirmative action and equal opportunity
employer. Applications from women and minorities are especially encouraged.


From: Gene Golub <>
Date: Fri, 30 Sep 88 17:52:34 PDT
Subject: NA-Net Introduction


There are several conveniencies associated with NAnet.

1. Sending messages to everyone on the net via the DIGEST.
2. Sending a message to an individual.
3. Sending a message to SIAM or the AMS.
4. Obtaining software through netlib.
5. Obtaining address information for an individual and/or organization.
6. Distributing and obtaining titles and abstracts from netlib.

% requests, comments or problems to
% submissions for broadcast to
% Note: Most people will be able to mail to host "".
% However, for those who can't, the mail can be routed through another
% machine: e.g.,, or
% The only people who may have problems getting to na-net are those
% on the Arpa-Internet who are using static host tables rather than
% name resolution software. I believe that all the major gateways
% to the other nets have good software and there should be no problem
% with mailing from bitnet (netnorth, earn) or csnet sites. All
% uucp sites that enter the Arpa-Internet through a "good" host
% will also have no problems e.g.,,,
% This list is not exhaustive.
% -mark kent

A. To send an announcement for the DIGEST, mail to
and everyone on the distribution list will be sent the message.
To obtain the address list, send the following msg to netlib:
send nalist from misc

B. To mail to an individual, send a message to
na.<last name>
and the message will automatically be forwarded. Try your own name.
Some addresses are quite complex and/or are not registered so it's
often easier to use this device.
If you send a message to that will automatically
be sent to me wherever I am. I DON'T GUARANTEE ANY ADDRESS IS CORRECT OR

C. You can send mail to SIAM simply as or and the AMS as

D. You can get some software as indicated below. This effort is due
to Jack Dongarra and Eric Grosse.

E. You can get address information for individuals as indicated below.
Note the SIAM database is contained in netlib so that you are able to
get the latest address of a SIAM member. The AMS has not co-operated
with us on this matter so we cannot provide addresses of AMS members.

F. Richard Bartels (na.bartels) maintains an irregular title and abstract
distribution service. He would appreciate electronic notices of
preprints and technical reports. Send
1. title
2. list of authors
3. authors' addresses
4. abstract
5. means of requesting a copy of the work
When enough (about 10-15) notices have come in, an announcement
will be made in the DIGEST and the abstracts can be obtained from netlib.

G. Other suggestions are welcome.

H. Please be careful in using the REPLY command. On some systems,
if "r" is used everyone on the mailing list receives the response;
a reply with "R" will go to the sender. Do not reply to the Digest!

I. Address any problems to

J. If you want to change your address or if a colleague wants to be on
the net, send a msg to

K. I keep a data base which is updated from time to time.
Unfortunately not all the changes have not been included in the
past. If you want to be included complete the attached form and
send it to Note, I've changed the
form to include FAX numbers.

< Last Name, First Name>
FNAME Full name with title.
TITLES Titles accorded to person.
OADDR Office address.
OPHONE Office phone number.
HADDR Home address.
HPHONE Home phone number.
PADDR Permanent address.
PPHONE Phone at permanent address.
INTERESTS Scientific interests.
FAX Local FAX number.
MISCELLANEOUS MISCELLANEOUS comments. No special format.


You can get mathematical software automatically by electronic mail.
Send mail of the form:
send qag from quadpack
to one of the three addresses below:

"send x from y" gets you the code for routine x in library y
nd all routines x calls.
To get just x itself, say "send only ...." You can ask for several
routines at once, and can make several requests, as in the mail:
send i1mach r1mach from core
send svd from eispack
send only ssvdc from linpack

Currently we offer routines from LINPACK, EISPACK, MINPACK, the special
function library FNLIB, the book by Forsythe, Malcolm, and Moler,
QUADPACK, golden oldies (like gaussq), Collected Algorithms of the ACM,
BIHAR, and many others including recent research codes.
To get a table of contents:
send index
To get an index for a library y, append "for y."

For advice on what routine to try, you still need a numerics expert;
we're just trying to simplify code distribution. Comments are welcome.

Also, it is possible to get address information on certain individuals.
To do this, use the mail to netlib with the whois command.
For example
whois meyer
will give you the address of more than one Meyer.

Jack Dongarra 312-972-7246 Argonne Nat Lab, Chicago, IL

Eric Grosse 201-582-5828
research!ehg ATT Bell Labs, Murray Hill NJ

it's a good idea to look at the index from time to time --gg)


End of NA Digest