NA Digest Sunday, April 30, 1989 Volume 89 : Issue 17

Today's Editor: Cleve Moler

Today's Topics:


From: Iain Duff <>
Date: Mon, 24 Apr 89 15:42:05 CDT
Subject: London Conference on Ordinary Differential Equations

IMA Conference on Ordinary Differential Equations

I attach the following message from Catherine Richards of the
IMA concerning the forthcoming conference on ordinary differential
equations to be held at Imperial College London from July 3rd to
7th, 1989. It should be viewed partly as a response to two recent
messages from Hans Stetter concerning the fees for the meeting. I
should also draw attention to the fact that there is no external
funding for this conference, it lasts a whole week, and lunches
are included. I personally am not involved in this meeting but
feel it important that readers of the digest are acquainted with
all facts before passing judgement.
I have been involved in many other meetings which sometimes have
included lunches in the registration fee and sometimes have not.
There is assuredly no hard and fast rule as to which is better. It
is difficult to run a conference on a tight schedule if lunches
are not provided and, for people wishing them, the costs of eating
out in London would normally be much higher than the cost charged
in the fees.
Some questions have also been raised concerning the exact nature
of the residential accommodation and the alternatives in the
Imperial College region. Further information on this will appear
shortly in a subsequent na-net message.
Iain Duff




In view of some adverse comments that have been made about the
fees charged for this conference some facts ought to be

1 The Institute simply aims to ensure that fees received for any
conference shall cover the direct costs.

2 For the conference in question the costs are very heavy.
They fall under 3 headings:

a) charges made by Imperial College for lecture theatres and
other facilities;
b) accommodation and meals for invited and highlighted
speakers, and organising committees;
c) travel costs of invited speakers.

The total of these costs is such that the fee charged to members,
whose membership subscriptions make a contribution to the
Institute's overhead conference expenses, are intended simply to
be the appropriate level to cover direct costs. The additional
charge made to non-members is a contribution towards the
overheads. Overall the Institute does not make any profit on

3 Recently a special effort has been made to reduce the
commitment for speakers' travel expenses, with some success.
As a result the fees can be reduced to a level which is the
very minimum at which the Institute can hope to avoid having to
subsidise the conference.

The new conference fees (in pounds) are as follows:


Conference fee 115 145

Bona fide postgraduate 85 100

The residential fee and other charges are unchanged.

Catherine Richards
Secretary and Registrar
Institute of Mathematics and its Applications


From: Gene Golub <>
Date: Tue, 25 Apr 1989 23:45:32 PDT
Subject: Changes and Additions to NA List






From: Rex Jaeschk <eaussie!>
Date: 25 Apr 89 21:59:09 GMT
Subject: Numerical C Extensions Group Report

A few weeks ago an announcement was posted re the formation of NCEG.
Here's a progress status report.

Numerical C Extensions Group (NCEG) Status Report

The meeting proposed for May 10-11 at Cray Research is definitely on.
The response to my paper mailing and postings to various newsfeeds
was encouraging with some 20+ people expressing their intent to be at
that meeting. Another 10-20 people expressed support and asked to be
kept informed of the group's progress. As of April 25, the attendee
list (in alpha order of company) is as follows:

Analog Devices - Kevin Leary
Apple - Jim Thomas
Argonne Labs - Cody
BRL (US Army) - Mike Muuss
Computer X -
Convex -
Cray - Tom MacDonald + one or more Cray customers
DEC - Randy Meyers
HP - Walter Murray
IBM - Larry Breed
Information and Graphic Systems - Mike Branigan
Intermetrics - P.J. Plauger
Journal of C Language Translation - Rex Jaeschke (convener)
MasPar Computer Corporation - Peter Christy
Polaroid Microelectronics - William Vetterling
(coauthor of Numerical Recipes [in C])
Pyramid - Wendy Thrash
Stellar - Peter Darnell
Sun -
Supercomputers - either Chuck Rasbold or Kelly O'hair
Unisys - Don Bixler
UofMN - Prof. Patrick Powell

AT&T (not confirmed)
LLNL (not confirmed)
Plum Hall - Tom Plum (not confirmed)

Those wanting to be kept informed include: Solbourne, TI, and the
Free Software Foundation.

Proposed Meeting Agenda

(Please bring a hard-copy of this to the meeting if you plan to

Item 2 will set the tone for the rest of the technical agenda. The
topics suggested in sessions 4-8 are simply my educated guesses at the
hottest issues. We will modify the agenda as agreed to at the meeting.
Tom MacDonald has requested time to present his latest revision for
adding complex arithmetic and this has been accommodated on the second
day. Tom will have copies at the meeting. If you wish to distribute
technical proposals please bring at least 25 copies to the meeting.
Call me before the meeting if you think you'll need agenda time for
specific issues.

Wednesday May 10

9:00 -- 10:30 Session 1

1. Opening activities

1.1 Welcome
1.2 Goals and purpose of the meeting
1.3 Host facilities
1.4 Introduction of participants
1.5 Procedures for this meeting

2. Discussion of Agenda (this meeting and long-term)

Prioritize technical issues as: Now, later, sometime, too hard.

Identify primary and alternate focal point person
for each main issue.

Identify related standards and prior art.

10:30 -- 10:45 Morning break

10:45 -- 12:15 Session 2

3. Discussion of Agenda (continued)

12:15 -- 2:15 LUNCH

2:15 -- 3:45 Session 3

4. IEEE Issues

4:00 -- 5:30 Session 4

5. Vector/Parallel Issues

6. Living with Fortran - calling, shared libraries, etc.

Thursday May 11

9:00 -- 10:30 Session 5

7. A proposal to add complex arithmetic to C -- Tom MacDonald (Cray)

10:45 -- 12:15 Session 6

8. General library issues

11:15 -- 12:15

9. Administration

9.1 Summary of decisions reached
9.2 Action items committed to
9.3 Future of NCEG as a group and/or long-term affiliation
9.4 Future meeting schedule - date, location, host
9.5 Mailings and submission procedures
9.6 Other business
9.7 Adjournment

Host for NCEG Meeting #2

Being optimistic I have tentatively arranged for the next NCEG meeting
to be held in conjunction with ANSI C X3J11 in Salt Lake City, Utah.
X3J11 meets on September 21-22 with NCEG on September 19-20 for one
and a half or two days. This eliminates the need for separate travel
expenses for those planning to attend both meetings.

DEC's user group DECUS is hosting that X3J11 meeting and since NCEG is
not currently affiliated with a recognized standards body, DECUS will
not host our meeting as well. The expected cost of hosting NCEG for
two days is about $900. This includes the meeting room, morning and
afternoon catering for 25 people, and A/V equipment.

Right now, the meeting room reservation is under my name, however, I
can assure you I will not be paying for it - I'm only a 1-man
operation and my contributions are taking other forms. I need to
exercise my option within 1-2 weeks after the May 10-11 meeting to
guarantee the facility. Please check with your management to see if
you can offer to underwrite all or part of this cost. I can arrange
to split it two (or perhaps even three) ways. I would very much like
a commitment by the end of the May meeting.

Lack of formal affiliation problems

At least one major vendor has expressed concern about the lack of a
formal standards affiliation. It seems their legal department forbids
employees from participating in any forum that could conceivably be
construed as an attempt by one group of companies to do something
proprietary. Issues of antitrust and monopolies raise their heads. Now
while my intent was to make NCEG a truly public forum this problem
must be faced, and soon, since companies with such policies otherwise
will not be providing input we need. Their participation can only
allow us to do a better job.

Becoming affiliated is not a bad idea. Certainly, it's much easier to
adopt existing voting rules, etc., than it is to invent our own. Also,
if we can eventually publish a Technical Bulletin for the ANSI C
Standard (for example) specifying numerical extensions, it will carry
more weight with the industry.

Before coming to the meeting find out what your company's policy is in
this regard.

I look forward to meeting you in beautiful Mendota Heights, MN in May.

Rex Jaeschke, Convener - April 25, 1989

Rex Jaeschke | C Users Journal | Journal of C Language Translation
(703) 860-0091 | DEC PROFESSIONAL |1810 Michael Faraday Drive, Suite 101
uunet!aussie!rex | Programmers Journal | Reston, Virginia 22090, USA


From: Steve McCormick <smccormi@copper.Colorado.EDU>
Date: Sun, 30 Apr 89 08:05:26 mdt
Subject: 4th Copper Mountain Multigrid Conf. Proceedings

Anyone interested in obtaining the preliminary proceedings of the Fourth
Copper Mountain Conference on Multigrid Methods can do so by sending a
check for $25 (payable to the University of Colorado at Denver) to:

Ms. Debbie Beltz
Computational Math Group, Box 170
1200 Larimer Street
Denver, CO 80204


From: Steve McCormick <smccormi@copper.Colorado.EDU>
Date: Sun, 30 Apr 89 08:06:09 mdt
Subject: Multigrid Use Survey

We are currently conducting an informal survey on the use of multigrid or
multilevel methods for solving practical problems. We mean this loosely to
include any situation where these methods have been used as important
components in the solution of problems of interest in physics, chemistry,
biology, medicine, economics, engineering,... We are especially interested
in production application codes. Any information you could supply us with
in this direction would be greatly appreciated. A few sentences with perhaps
a reference or contact would be most helpful. Please e-mail your comments to:

If our survay is successful, we'll publish the results in the MG Newsletter
with copies sent to the respondents. (Free Newsletter subscriptions can be
obtained by sending us your regular mail address.)


End of NA Digest