NA Digest Tuesday, February 17, 1987 Volume 87 : Issue 3

This weeks Editor: Gene Golub

Today's Topics:


Mail-From: GOLUB created at 14-Feb-87 15:39:58
Date: Sat 14 Feb 87 15:39:57-PST
From: Gene H. Golub <GOLUB@Score.Stanford.EDU>
Subject: Stanford NA Reunion
To: NA@Score.Stanford.EDU

Dear Friend,

As you may know, the CS department of Stanford is having a reunion:
"Computer Science at Stanford: the impact of the first ten years" at
Stanford from March 26 to March 28. As part of this reunion, we are
having a reunion of the greater Stanford NA community (this includes
former or present students, faculty, visitors and friends): "Numerical
Analysis at Stanford: Rememberance of Things Past", to be held on the
afternoon of Saturday, March 28. The tentative program consists of a
tour of Serra House, the drafting of a "family tree", several informal
talks and a cocktail party at Gene's followed by a banquet in a
Chinese restaurant.

Several people have agreed to talk and here is their tentative titles:

Ake Bjorck (Linkoping/USC): On Being a Frequent NA Visitor at Stanford;

Alan George (Univ. of Tenn./ORNL): Life at the Top: On Being a Dean;

Eric Grosse (Bell Labs): The Greater World of Bell Labs: the Netlib Revolution;

Jack Herriot (Stanford Emeritus): In The Beginning ....;

Cleve Moler (Intel): Fond Memories of George Forsythe;

John Palmer (Ncube): Standards and Startups;

Beresford Parlett (Berkeley): The Pleasures and Work of Jim Wilkinson;

Jim Varah (UBC): All in the Family: the Stanford NA Family Tree.

The cost of the festivities (including the banquet) is $25 per person.
Please send a check to Gene (made out to him) by March 21.

We really do hope that you can come and join us in this unique
celebration! Feel free to bring your family and friends. And bring your
memorabilia too!

Tony Chan
Gene Golub

PS: If you are unable to come, please feel free to send a brief
reminiscence and/or photos.


From: prlb2!mswe.DECNET!
Date: Mon, 26 Jan 87 08:52:58 +0100
Subject: Networking paper in CACM

The following paper is worth-while reading for current developments
in short- and long-distance research networking; perhaps, it should
be mentioned on NANET?

John S. Quarterman & Josiah C. Hoskins, Notable Computer Networks.
Communications of the ACM 29(1986)10, 932-971.

Herman J. Woltring


Date: Thu 29 Jan 87 17:20:28-PST
From: "Gene H. Golub" <>
Subject: na.<last name>

We have just changed our system so that na.<last name> will be incorporated
into the system more quickly. You might want to try a test on your own name.
You can send a msg to na.<your last name> and you should get a msg from

Gene Golub


Date: 2 Feb 1987 13:15:09 GMT
To: NA <>
Subject: Lectureship at Bath
From: Alastair Spence <>




Applications are invited for the above post in the
Mathematics group of the School of Mathematical Sciences,
which is tenable from 1st. September 1987.

The successful candidate will be expected to have interests
which are compatible in broad terms with existing research
activities in continuum mechanics, which relate mainly to
solid mechanics and aspects of wave propagation, though
expertise in mechanics of suspensions or the numerical
solution of partial differential equations would be

Duties will include teaching to students within the School
of Mathematical Sciences and in other Schools where a
substantial service teaching commitment exists.

Particular emphasis will be placed however on the successful
candidate's achievements and potential in research.

Salary in the range 8,020 - 15,700 UK pounds.
All lectureships are subject to a probationary period
of three years.

Closing date : 25th. February 1987.

Further particulars from:

Professor J R Willis,
School of Mathematical Sciences,
University of Bath,

UK Telephone: 0225 826184 (School Office)
0225 826241 (Professor Willis)


Alastair Spence on NANET: na.spence@su-score


Date: Mon, 2 Feb 87 16:14:44 PST
From: Mark Kent <>
Subject: na.lastname status message II

(These messages will not become a regular thing.)

Two new improvements have been installed that should improve

1) In the past the na.lastname messages were shipped out of Stanford
from a Vax called Navajo. Quite a few messages were going nowhere
because Navajo had trouble either opening connections with hosts
or maintaining a connection long enough to get the mail through.
This is probably caused by increased arpanet traffic and lack of
flexibility in certain parameters in the mail system. I have noticed
that other Unix systems have encountered the same problem recently.

Score, a Dec-20, has better luck with out-going messages. Maybe
this is because Score has a direct arpanet connection. Now, all
NAnet messages (both broadcast and individual) go through Score.
(Score also seems more forgiving with addresses.)

2) Previously, messages were shipped out as if the sender was
daemon@navajo. This caused some troubles as people could not
conveniently use the "reply" command to such messages. Also,
messages sent to would seem to get
no reply. As it turns out, the mailbox for daemon@navajo was
collecting replies from netlib. The daemon was also collecting
mail errors for things further down the line. So if you sent
a message to na.xxxx and the corresponding address was unreachable
for 5 days then you would never know it.

This has been fixed. You should now get error messages as if
you had typed the real address rather than the na.lastname address.
Hence, if some host is down and your message cannot get through
then you will know it.

Unfortunately, while I was installing these changes the database
for the na.lastname facility was somehow garbled. In fact it was empty.
So some of you may have got messages like
"na.someone not recognized"

The database has been fixed and everything should work from now on.



Date: Thu, 5 Feb 87 14:42:47 pst
From: "Alan J. Laub" <ucsbcsl!>
Posted-Date: Thu, 5 Feb 87 14:42:47 pst
Subject: notice of NATO ASI



SEPTEMBER 14-25, 1987

AIM: This ASI is intended to bring together both academic
and industrial control systems engineers who are active or
who wish to become active in the application of advanced
computing concepts and techniques in the design and imple-
mentation of control systems. The main areas to be covered
in the Institute lie at the interface between computer sci-
ence and control and include:

* expert systems in design and implementation of control

* control of distributed and discrete event processes

* parallel computing algorithms and architectures for
control and signal processing

A series of tutorial-level lectures will be presented by
invited lecturers in each of these main areas while a few
research-level contributions will be presented as short
papers by selected participants.

MAIN LECTURERS: The main lecturers for the Institute will
include: K.J. Astrom (Lund), G. Blankenship (Maryland), G.
Cybenko (Tufts), M.J. Denham (Kingston Polytechnic), A.J.
Laub (UC Santa Barbara), E.H. Mamdani (Queen Mary College),
J.P. Quadrat (INRIA), A.H. Sameh (Illinois), P. VanDooren
(Philips), and W.M. Wonham (Toronto)

FEES: The attendance fee of $600 includes accommodation and
all meals for the duration of the Institute (13 nights) and
transportation to and from Pisa airport.

DIRECTORS: The co-directors of the Institute are:
Professor Michael J. Denham Professor Alan J. Laub
Kingston Polytechnic Dep't. of Electrical and
Kingston upon Thames Computer Engineering
KT1 2EE University of California
England Santa Barbara, CA 93106
Tel. (44) (1) 549-1366 Tel. (1) (805) 961-3616

APPLICATIONS: Attendance is by personal invitation only.
Further details and application forms are obtainable from
the above. Please note that Professor Denham will be at the
Santa Barbara address until Apr. 1, 1987. To ensure full
consideration by the organizing committee, applications
should be submitted by May 15, 1987 at the latest.


Subject: MPS standard form for LPs?
Date: Mon, 16 Feb 87 14:14:40 -0500

Can someone help me? What is the MPS standard form
for Lp's and, more importantly, how to I translate the
MPS form to numerical form suitable for an LP program?
Of course, this is some sort of standard, but what is it?

charlie (


End of NA Digest