krj@na.toronto.edu (Ken Jackson) (03/18/89)
From: nagmag%ukc.ac.uk@NSS.Cs.Ucl.AC.UK Date: Fri, 17 Mar 89 09:45:13 EST Subject: NAgMAg Vol 89 #05 Sender: trh%ukc.ac.uk@NSS.Cs.Ucl.AC.UK NAgMAg Friday, March 17 1989 Volume 89 Issue 5 Today's Topics NAG Responses to NAGMAG #4 Library Mark 13 under NOS/VE Information on E01ADF needed GAMS Classification Scheme General Comments %% Editorial! %% %% The subscription list is still growing steadily (the number %% of non-NAG staff now stands at 110 and this includes several %% mailing lists in the US) which is some relief to me after the %% very slow start. %% %% ************** PLUG PLUG PLUG PLUG ***************** %% I hope you are all paid-up members of NAGUA (the NAG Users %% Association)! If not and you'd like more details drop me %% a message and I'll make sure you get an application form. %% If you are make sure you reserve your place at the NAGUA %% meeting at York (UK) this September. %% %% And look NO outstanding replies :-) %% %% Keep sending the articles ....... Tim --------------------------- Date: Sun, 12 MAR 89 17:49:24 GMT From: DAVID@vax.nag.co.uk Subject: NAG Responses to NAGMAG #4 For the time being it has been agreed that I will co-ordinate replies to the many interesting issues that NAGMAG has raised. Firstly apologies from NAG for the delays/confusion over replies to NAGMAG #3; I never expected anyone to reply to the individual rather than through the NAGMAG and so failed to advise against this course of action. RE: Online Workstation Help System We anticipated a demand for this some time ago and prepared a version for use on PC's. Until now there has never been any incentive to complete the work since the expected requests never materialised - even during negotiations with CHEST for the British Universities. In reponse to NAGMAG I am pleased to say that an effort is now being made to complete the work. Please await an announcement in these columns. RE: GAMS Scheme We discussed this at some length and we will probably provide the index at Mark 14. (We think we have found a volunteer within the office!) Armin Friedli's latest message is of interest though and we are very grateful to him for this. Of the people I have spoken to, none of us were aware of this new development. It clearly makes sense to conform to this revised version and this may delay things somewhat. RE: Nested Calls to D01AJF I think the message is most helpful - if only to highlight the dangers of non-standard Fortran. We are looking to make our manuals better and easier to use, so examples like this of user difficulty help us immensely. (On the topic of manuals internal thoughts range from having a smaller vocabulary, to assist the non-native English speaker, to cutting down the level of information provided. With this still a topic for debate have any of you any thoughts or comments ?) RE: Convective Boundary Problems At the moment NAG does not have any software directly applicable to 2 D parabolic equations. Richard Brankin tells me that he has some hopes that NAG may acquire some solvers for future Marks, but I am afraid that for the immediate problem we have no software to hand. Sorry - perhaps someone else out there can help ? %% Come on you FEM guys -- there is a whole NAG FEM library out %% there -- somebody must know how to use it. Don't they:-) %% Tim RE: D02EBF Looping Thank you for reporting this. Richard Brankin came in today (Sunday) to investigate this problem, which puzzled us both last night. He tells me that he has found the 'bug' which is present in all versions and which will be corrected. In the meantime subroutine OUT1 provides a safe solution. --------------------------- Date: Tue, 14 Mar 89 09:52:03 MEZ From: ZZENDE@dhvrrzn1.earn Subject: Library Mark 13 under NOS/VE I didn't send the question via NAgMAg, but I guess some people might be interested in the answer I got from NAG concerning the NOS/VE implementation of Mark 13 of the Library: "The latest position is as follows. FTN1 compiler: CDC have a full-time programmer workin on this, and the Library is compiled, all examples have been run and verified, the stringent test programs have been run and verification of the results is about to begin. We (NAG) hope to receive the Library from CDC at the end of March. We will then certify it for release, finalise the support notes, and hope to distribute it to users during April. FTN2 compiler: CDC will work on this following the FTN1 version. We (NAG) estimate that if all goes well it will be released to users in early June." I hope things will work that way| Peter Endebrock --------------------------- Date: Wed, 15 MAR 89 19:34:24 GMT From: SUPPORT@leva.icf.leeds.ac.uk Subject: Information on E01ADF needed Hello, A colleague of mine has inherited a Fortran program that was last used when the current release of the Nag library was Mk 9. It calls a routine E01ADF that has since been superseded by (we believe) E01BAF. We no longer have the documentation for E01ADF. A minimum need is a description of each parameter (data type, significance) but a photocopy of the full description for this routine would be ideal. Can anyone help, please? Regards, Ted Allwood --- JANET support@uk.ac.leeds.icf.leva Interactive Computing Facility Department of Mechanical Engineering University of Leeds Leeds LS2 9JT Fax: 0532 424611 --------------------------- Date: Wed, 15 Mar 89 13:40:05 EST From: Ronald Boisvert <boisvert@rfb.cam.nbs.gov> Subject: GAMS Classification Scheme Response to "GAMS Scheme" in (NAgMAg V89 #3 and #4): We are busy working on a new edition of the "Guide to Available Mathematical Software (GAMS)" which we expect to be ready at the end of this year. The purpose of GAMS is to provide summary of the mathematical and statistical software available to users of our local systems. Among the software packages included in the new edition will be ACM Algorithms (1975-1986) BMDP (1987) CMLIB (Version 3) DATAPLOT (1988) IMSL (Edition 9.2) IMSL/MATH (Edition 1.0) IMSL/STAT (Edition 1.0) IMSL/SFUN (Edition 2.0) MINITAB (1982) NAG (Mark 12) PORT (Version 2) SAS (5.03) SPSS (2.1) STARPAC (2.07) All software is cataloged according to the GAMS Classification Scheme. The scheme is tree-structured, problem-oriented, and quite detailed (there are more than 700 nodes). In several months we plan to issue a new version of the classification scheme. The new version has a completely revised statistics section, as well as other minor changes. All our software will have been classified under this scheme when it is issued. At that time we would be happy to provide a listing of new classification scheme and a list of software modules in selected libraries with their classifications to anyone who needs them. Send requests to spring@vax.cam.nbs.gov Ron Boisvert Sally Howe David Kahaner Jeanne Springmann National Institute of Standards and Technology Gaithersburg, MD 20899 USA --------------------------- Date: Sun, 12 Mar 89 12:39:31 EMT From: FSCPG@nobergen.earn <Pat Gaffney> Subject: General Comments I have received the recent mags and I must say that this is a very worthwhile idea and you are doing a great job. A few comments: 1) A UNIRAS graphics interface would be very welcome. 2) I find it very hard to believe that anyone would regard the GAMS classifications scheme as an "indispensable tool for people who need help in choosing an adequate Library routine...". How anyone can make the CORRECT choice from such an index is beyond me. In my opinion NAG should spend its resources on other worthwhile projects like TOOLPACK. 3) Recursive calls to routines from within FORTRAN should be discouraged. Nobody expects FORTRAN to handle recursiveness so therefore the NAG routine should not really be amended. --------------------------- %% %% 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