Explicit written permission from Hanson, reporting decision of ACM Publications Board. From mhw Thu Apr 20 12:04:30 EDT 1989 The single precision subroutine and main program are brkf45.f and main.f. The double precision versions are dbrkf45.f and dmain.f. ADMINT was installed in the office of Terje Espelid. From cs2.cs.wsu.edu!acg%dec2.WSU.EDU Fri Dec 01 10:52:11 PST 1989 Received: from dec2.cs.wsu.edu by cs2.cs.wsu.edu (5.59/1.34) id AA00871; Fri, 1 Dec 89 10:21:50 PST Received: by dec2.cs.wsu.edu (5.57/Ultrix3.0-C) id AA20122; Fri, 1 Dec 89 10:52:12 PST Message-Id: <8912011852.AA20122@dec2.cs.wsu.edu> To: terje@eik.ii.uib.no Cc: acg%cs2.cs.WSU.EDU@cs2.cs.wsu.edu, ehg@research.att.com Subject: Re: ADMINT In-Reply-To: Your message of Wed, 29 Nov 89 11:41:59 +0100. <8911291043.AA11288@brake> Date: Fri, 01 Dec 89 10:52:11 PST From: acg%dec2.WSU.EDU@cs2.cs.wsu.edu I am sorry to have not replied completely to Jarle's earlier message. I have no objection to ADMINT being added to netlib and do not mind if the naming is changed to something like "DADMNT, SADMNT" (for double and single precision versions) or, as you report Jarle recently has suggested "DCUHRE, SCUHRE". Either choice will be fine with me. I intend to get my list of responses to the referee reports to you and Jarle before Christmas. Best wishes, Alan ------m dongarra Tue Jan 23 09:15:59 EST 1990 Subject: TOMS tapes Bob Renka just sent a letter to Gary Shepherd directing him to send me 665-678 and, in the future, duplicates of the tapes as they are being sent to IMSL. Let's keep our fingers crossed that this all works. From epsilon.bldr.nist.gov!jrd Tue May 29 06:15:09 MDT 1990 Received: by pyxis; Tue May 29 08:15 EDT 1990 Received: by inet.att.com; Tue May 29 08:15 EDT 1990 Received: from epsilon.bldr.nist.gov by alpha.bldr.nist.gov (4.1/SMI-DDN) id AA18683; Tue, 29 May 90 06:15:26 MDT Received: by epsilon.bldr.nist.gov (4.1/SMI-3.2) id AA13936; Tue, 29 May 90 06:15:09 MDT Date: Tue, 29 May 90 06:15:09 MDT From: jrd@epsilon.bldr.nist.gov (Janet R. Donaldson X5114) Message-Id: <9005291215.AA13936@epsilon.bldr.nist.gov> To: ehg@research.att.com Subject: ODRPACK Reference Guide Eric - The ODRPACK reference guide that you requested from Bobby Schnabel follows. Please let me know if I can be of further help. Janet Donaldson (303) 497-5114 jrd@alpha.bldr.nist.gov ====================== From NSFnet-Relay.AC.UK!jcash Tue Jul 31 17:30:20 BST 1990 Received: by pyxis; Tue Jul 31 12:52 EDT 1990 Received: by inet.att.com; Tue Jul 31 12:52 EDT 1990 Date: Tue, 31 Jul 90 17:30:20 BST From: JR Cash (Imperial) To: ehg@inet.att.com Subject: toms algorithm 669 Dear Eric, This is a message about algorithm 669 (TOMS) which appears in netlib. You may remember when I was there that you got an e-mail about a line being commented out when it should not be. The mistake is still there! Is it possible to do something about it???? All that needs to be done in both the single and double precision versions is to uncomment the line defining erc24 and erc25. I hope all is well. I expect to be at Bell in early September. Best wishes, jeff cash ------m jcash@nsfnet-relay.ac.uk mhw Sat Aug 25 16:00:23 EDT 1990 From-: Eric Grosse ehg@research.att.com 908-582-5828 To-: jcash@nsfnet-relay.ac.uk mhw Date-: Sat Aug 25 15:52:45 EDT 1990 Subject: tragicomedy of errors (was: toms algorithm 669) Well, I don't know how the bug got into ACM's copy in the first place, or why it crept back into netlib's copy. (Most likely, I installed a tape from ACM and didn't notice that it included 669; it's hard to remember all those numbers!) Anyway, when your mail arrived, I was about to go off on some trips, so Margaret kindly fixed the file here. However, it appears to me looking at this now that I'm back in town that she only fixed the single precision version. On the (perhaps optimistic) belief that I understand what's going on, I have made a similar change to the double precision code (in the same Unix file). To be specific, I made the change "g/, ERC25/s/^C/ /". I made the change to the netlib file here and at nac.no, and in the staging area at ornl.gov. (That means it will get installed when Jack gets back to town.) The other netlib sites will get the change at the next update. With luck, that may be in the next couple weeks. There have been so many screwups here, it might be wise for you to check in a couple months to see that all the files are correct everywhere. Sorry for the trouble. Eric From uiuc.edu!stankerr Thu Jan 17 13:14:51 0600 1991 Received: by pyxis; Thu Jan 17 14:14 EST 1991 Received: by inet.att.com; Thu Jan 17 14:14 EST 1991 Received: from opus.cso.uiuc.edu by ux1.cso.uiuc.edu with SMTP id AA11549 (5.65a/IDA-1.4.2.3 for ehg@research.att.com); Thu, 17 Jan 91 13:14:51 -0600 Date: Thu, 17 Jan 91 13:14:51 -0600 Message-Id: <9101171914.AA11549@ux1.cso.uiuc.edu> To: ehg@research.att.com From: stankerr@uiuc.edu Subject: ACM TOMS algorithms Sorry to bug you, but would you please remove the notice from the 'toms' library that says TOMS algorithms can be obtained from IMSL? They stopped distributing them several months ago, and even returned to me an order I had already sent. They are telling the world (as far I know; at least, they told me) that TOMS algorithms are available ONLY through netlib! (The notice I saw was delivered by netlib at ORNL; I haven't tried att yet.) -------------------------------------------------------- Stan Kerr Internet: stankerr@uiuc.edu University of Illinois BITNET: stankerr@uiucvmd Computing Services Office Phone: (217) 333-5217 1304 W. Springfield Urbana IL 61801 !/bin/mail uiuc.edu!stankerr Wow, that's news to me! I'll call Renka and find out what's going on. Thanks for the notice. Best wishes Eric !D ------m uiuc.edu!stankerr Fri Feb 1 22:57:28 EST 1991 From-: Eric Grosse ehg@research.att.com 908-582-5828 To-: uiuc.edu!stankerr Date-: Fri Feb 1 22:55:04 EST 1991 Subject: toms disclaimer I've finally gotten around to removing the IMSL address from the toms disclaimer. The new file should get to Oak Ridge sometime next week. Thanks for pointing out the problem. My friend at IMSL said they will probably still handle request for a few more months, but were losing money on the operation and want out. From cayuga.cs.vt.edu!ltw Wed Jan 30 08:35:29 0500 1991 Date: Wed, 30 Jan 91 08:35:29 -0500 From: ltw@cayuga.cs.vt.edu (Layne Watson) Subject: interesting problem with TOMS alg 555 from netlib Eric, A fellow named K. M. Briggs from LaTrobe University in Australia recently sent me a letter, claiming that the ACM TOMS algorithm 555 from netlib was wrong. Specifically, he said the subroutine STEP (which I took verbatum from Shampine's book) was missing some lines, and interestingly the missing lines were exactly those shown on the book cover! Well, I checked my old files, and 1) my version of STEP in algorithm 555 is correct in my disk file here; 2) the hard copy I submitted to ACM TOMS was correct; 3) TOMS didn't publish a listing of STEP, so I can't verify whether TOMS has 555 correct themselves. Can you look into this, at least from the netlib side? Hypothesis: what Briggs said was missing is about one microfiche listing page, the way TOMS used to put the programs on microfiche. Somehow a microfiche page got lost going from TOMS to netlib???? Thanks. Layne ------m cayuga.cs.vt.edu!ltw Mon Feb 4 10:02:59 EST 1991 From-: Eric Grosse ehg@research.att.com 908-582-5828 To-: cayuga.cs.vt.edu!ltw Date-: Mon Feb 4 09:54:46 EST 1991 Subject: netlib/toms/555 After spending a large chunk of the weekend reading tapes and running the files through various filters, I'm close to assembling a good copy of the TOMS algorithms. I have two tapes from the "official" TOMS distribution service that contain 555. In most cases, tape1 is more recent or has fewer character-code errors than tape2, but in the particular instance of 555, tape1 is missing lines from STEP, so I've taken the 555 from tape2 and installed it here. When all the other TOMS checking is done, I'll distribute the new library to other netlib sites. I have no idea how the ACM could have lost those lines, but I doubt it has anything to do with microfiche. Please pass this along to Briggs and thank him for bringing the problem to our attention. I've known for a couple years that I needed to check these tapes for internal consistency, but the task was tedious enough that I'd postponed doing it until your mail. Best wishes Eric Tue Feb 5 13:28:57 EST 1991 installed files from netlib@ukc.ac.uk, after passing through seqstrip 380 400 404 407 410 414 420 433 473 475 478 483 485 488 386 403 406 408 413 419 432 458 474 476 479 484 487 490 From cs.purdue.edu!jrr Wed Oct 16 14:28:55 EST 1991 To: "Stan Kerr" Cc: jrr@cs.purdue.edu, "Eric Grosse" Subject: Re: Notifying netlib managers of TOMS distribution service Date: Wed, 16 Oct 91 14:28:55 EST From: jrr@cs.purdue.edu Stan: I don't know how "official" it is, but it is a fact that the ACM Algorithms distribution was taken over by C.Abaci at the start of 1991. John Rice From ponder.csci.unt.edu!renka Tue Feb 1 10:58:04 0600 1994 Date: Tue, 1 Feb 94 10:58:04 -0600 From: renka@ponder.csci.unt.edu (Robert Renka) To: ehg@research.att.com, renka@ponder.csci.unt.edu Subject: Re: December TOMS issue Eric, A Remark should replace the original algorithm. Most Remarks correct bugs in the original algorithm. A major revision of an algorithm is usually published as a new algorithm even when it renders the original one obsolete. I therefore see no need to retain algorithms for which there are Remarks. Robert From ehg Tue Feb 1 11:59:39 EST 1994 From: Eric Grosse 908-582-5828 To: renka@ponder.csci.unt.edu Date: Tue Feb 1 11:59:02 EST 1994 Subject: Re: December TOMS issue Robert, Thanks, I'll go through and make that happen. (Maybe not until next week, however.) Best wishes, Eric