krj@na.toronto.edu (Ken Jackson) (03/24/89)
From: nagmag%ukc.ac.uk@NSS.Cs.Ucl.AC.UK
Date: Thu, 23 Mar 89 10:25:39 EST
Subject: NAgMAg Vol 89 #06
Sender: trh%ukc.ac.uk@NSS.Cs.Ucl.AC.UK
NAgMAg Thursday, March 23 1989 Volume 89 Issue 6
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% The official electronic digest of the NAG Users Association %%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Today's Topics
E01ADF
Assistance sought for Mark 3 Graphics
Replies to NAGMAG 5
C Software
%% Editorial
%%
%% This digest is going out a day early -- so it can sit in your
%% mailbox 'til you've recovered from gorging too many chocolate eggs!
%% I'm away next week so the next digest (if there's any input -- HINT)
%% will go out sometime during the week beginning April 3.
%%
%% NAG wish it to be known that this is NOT an official NAG digest but
%% it IS an official NAGUA digest -- hence the appearance of the
%% boxed comment in the header.
%%
%% As you will see I have been forced to insert my own articles in
%% an attempt to generate some discussion. I don't have an infinite
%% supply of these. I just don't believe nobody has anything to say
%% about NAG. One or two people have said they are only interested
%% in particular areas (e.g., statistics) -- don't be afraid of
%% submitting articles I'm sure other readers will be interested.
%%
%% Have a peaceful Easter.
%% Tim
---------------------------
Date: Fri,17 Mar 89 16:37:36 GMT
From: Bob_Hatfield@vme.ccc.nottingham.ac.uk
Subject: E01ADF
I have a copy of the documentation for E01ADF if you still haven't
found a copy.
Bob Hatfield
Cripps Computing Centre
University of Nottingham
---------------------------
Date: Sun, 19 MAR 89 17:28:50 GMT
From: DAVID@vax.nag.co.uk
Subject: Assistance sought for Mark 3 Graphics
Requests for assistance from NAG Graphics Users
-----------------------------------------------
1)Do any of our NAG Graphics users have any data cases and/or
programs which generate interesting pictures or illustrate
applications to which NAG Graphics has been put?
We would like to include examples of 'real usage' of our software in
such documentation as advertising material, newsletter articles,
displayed on our trade stand, used to illustrate talks about the NAG
Graphics etc.
We would welcome any contributions of: code (and data), data only
or pictoral output (e.g. 35mm slide, good quality hardcopy etc.),
preferably accompanied by a brief explanatory description.
The source of any contributions would of course be acknowledged if
we were to use them.
2) In the Mark 3 Graphics Library Manual we wish to include a
section describing common user mis-understandings and the corrective
action to take. Below please find a list which NAG have compiled. We
believe that many NAGUA members will perhaps have experienced other
problems which should be included. If so please let us know and we
will see that they appear in the list.
Corrective actions for commonly encountered errors.
--------------------------------------------------
The following Section has been derived directly as a result of questions
received from our user sites about problems encountered when running the
NAG Graphics Library. It is hoped that the following table will
assist you in identifying the cause of any problems you may
encounter when linking or running NAG Graphics.
To use this table you should first identify the category of your problem
(left hand side of table) and check each of the possible causes listed
(right hand side) in turn. Full details associated with each cause are
listed following the table (indexed alphabetically).
Should you still be unable to resolve your problem after having checked
the advice given, you are recommeneded to contact your local NAG site
representative for further assistance.
Problem Possible Cause
------- --------------
Problems encountered at Link time:
Messages are generated indicating Libraries being searched in
that there are unresolved external wrong order (a)
references (or multiple references)
Incorrect routine name used (b)
Problems encountered at Run time:
Program will not run without Incorrect Parameter(s)
generating run time messages supplied (c)
e.g. access violations. Wrong base precision used (d)
Program was copied from
Handbook (e)
Wrong compiler version (f)
Graphical output is incomplete NAG Graphics has not been
terminated correctly (g)
Graphical output generated was Wrong base precision used (d)
not what was expected
Wrong NAG Graphics routine
used (h)
Incorrect Parameter(s)
supplied (c)
Site specific modifications
made to NAG Graphical
Interface (j)
Example output plot differs from Incorrect Parameter(s)
that printed in the Handbook supplied (c)
Underlying plotting software
has
different attributes from one
used
by Handbook (Adobe PostScript)
(i)
Site specific modifications
made
to NAG Graphical Interface (j)
Possible causes - further guidance
----------------------------------
a) Libraries are being searched in wrong order
The correct order for linking NAG graphics routines is:
NAG Graphics, NAG Graphical Interface, NAG routines,
Plotting Software
Please check the instructions supplied by your site or contact your
site representative for further assistance.
b) Incorrect routine name used
Check that you have used the NAG naming convention applicable to
your site. NAG graphics routine names may end in either F
(e.g J06GAF) or E (e.g. J06GAE).
Check that you have not mistyped the NAG Graphics routine name. All
NAG Graphics routines start with the characters J06 (J-zero-6) not
JO6 (J-oh-6); it is easy to mistype this.
Please see section 2 of Background Notes for further details.
c) Incorrect Parameter(s) supplied
Check that you have not supplied an incorrect value for one of
the input parameters to your selected NAG Graphics routine; you
are advised to check the following:
- Your calling sequence(s) matches that given in the associated
routine documention.
- The type of each variable
- That all EXTERNAL subroutines have been declared
- The base precision being used for interpreting real variables
- That you have supplied sufficient workspace
- The dimension of arrays
- The order of dimensions for 2D arrays
d) Wrong base precision used
NAG Graphics routines are available in both single and double
precisions. It is normal for a site to make available only one
precision version of the Library. If you call NAG Graphics
with single precision arguments and the implementation available
at your site is double precision your program may behave
unpredictably. A program will not necessary fail
and a plot may even be generated.
A common cause for selecting the wrong precision is copying code or
routine parameter specifications directly from the User Handbook
without considering the precision of the NAG Graphics Library you
are using. The examples and text in the NAG Graphics Library
Handbook are printed to reflect single precision implementations.
Emboldened terms are used to reflect those elements which need to
be converted to double precision
Please see section 2 of Background Notes for further details.
e) Program was copied from Handbook
Typing errors may be the source of your problem.
You are advised that your site has a copy of all example programs
printed in the Handbook which are specific to your local
implementation of the NAG Graphics Library.
Please contact your site representative for further details.
f) Wrong compiler version selected
When linking to NAG Graphics Libraries on some systems, especially
Personal computers you should ensure that your compiler is
compatible with the implementation NAG have supplied. The following
should be checked:
- Compiler version
- Whether a co-processor is required or not (PC machines)
- Whether G-Floating or D-Floating arithmetic is being used
You can determine the compiler version of the NAG Graphics
Implementation available at your site by calling NAG Graphics
routine J00AAF (zero-zero).
Please contact your site representative for further assistance.
g) NAG Graphics has not been terminated correctly
You have omitted to call NAG Graphical Interface routine J06WZF.
J06WZF ensures that the underlying plotting software and NAG
Graphics are terminated correctly and that all graphical
buffers are flushed.
h) Wrong NAG Graphics routine selected
Check that you have selected the appropriate routine to produce
your desired output plot.
A Pictoral Guide to using the NAG Graphics Library and detailed
guidance in each Chapter Introduction have been provided in the
Handbook to assist you in your selection of routines.
i) Underlying Plotting Software attributes vary
The form of graphical output is very much dependent upon your target
device and the NAG Graphical Interface in use. You are recommended
to use discretion when judging the acceptability of graphical
results. The graphical plots printed in the User Handbook are
supplied as examples of the style of output you would expect from
the routine.
The general form of output generated should match that supplied
i.e. the position of lines, textual content, relationship of
plot to the viewport etc. should be in a similar position,
however differences such as the
following can be expected:
- Representations of line styles (other than solid)
- Marker styles
- Text font representation.
- Text sizes (particularly if using medium quality (hardware) text.
- Colours (of lines or filled areas)
- Package supplied fill styles (most example programs supplied use
NAG generated fill styles; these should not vary significantly)
j) Site specific modifications
Your site may have made modifications to the default settings of the
NAG Graphical Interface you are using. Any modifications made should
be indicated in local documentation.
Please contact your site representative for further details.
LAC/16-Mar-1989
---------------------------
Date: Sun, 19 MAR 89 18:08:01 GMT
From: DAVID@vax.nag.co.uk
Subject: Replies to NAGMAG 5
RE: E01ADF
NAG keeps copies of the withdrawn routine documentation for these
purposes. In brief:
SUBROUTINE E01ADF(N,A,X,Y,W,D,IG,VAL)
INTEGER N,IG
real A,X(IG),Y(IG),W(IG),D(IG),VAL
N- the number of intervals ie there are N+1 datapoints
X,Y the datapoints (x,y)
W,D work arrays
A specified point where interpolated value is required
VAL on exit contains the answer.
Exactly the same answers should not be expected. In the newer
routines more natural end conditions were imposed ; earlier software
would not reproduce a cubic exactly - a feature that still exists
in other libraries I believe.
RE: GAMS
Ron Boisverts offer is very atractive to NAG. We would hope that
MARK 14 would be included. For completeness the Sparse Matrix
Library, the NAG Graphics, and the statistical packages GENSTAT and
GLIM should be seen alongside the other entries.
We would like to take the latest version when it becomes available
and make these additions.
---------------------------
Date: Thu, 23 MAR 89 15:15:01 GMT
From: trh@ukc.ac.uk
Subject: C software
I was recently assaulted by a colleague waving an advertisement
for NAG software which claimed to be able supply numerical s/ware
in Fortran/Pascal/ADA and C. What exactly is available in C?
Are there any plans to produce a C version of the main subroutine
library? If not, why not?
Anybody else out there perceive a need for C numerical s/ware?
Tim
---------------------------
%%
%% Replies or submissions to nagmag@uk.ac.ukc
%% Distribution changes to nagmag-request@uk.ac.ukc
%%
%% END OF ISSUE
Reposted by
--
Kenneth R. Jackson, krj@na.toronto.edu (on Internet, CSNet,
Computer Science Dept., ARPAnet, BITNET)
University of Toronto, krj@na.utoronto.ca (CDNnet and other
Toronto, Canada M5S 1A4 X.400 nets (Europe))
(Phone: 416-978-7075) ...!{uunet,pyramid,watmath,ubc-cs}!utai!krj