For each of the following, we think we have permission for the original author to distribute. beatson binary-lu bitnet bsmp.f dcgc decblas dpmpar.c dpmpar.f eiginv fft.f fft.mail ibmblas iccg iccg-doc iccg-paper lsqr bug fix received from Saunders Oct 87 machar.c mfft nalist netlib netlib-old netlib-paper parallel-list perlman ppde rainbow.c randnum-cray rolex sgefac smchar.c spline.c stree symmlq treeqr treeqr.f varp2 varpro varpro-drv varpro-input varpro-output vaxblas The following are believed public domain and already widely distributed. underwood ===== hartley ===== From-: research!ehg (Eric Grosse, 201-582-5828) To-: na.buneman Subject: FHT legalities I noticed in last month's Stanford Observer that Bracewell patented the Fast Hartley Transform and the Trustees of Stanford are licensing it. What does this imply about netlib's (free) distribution of your code? Should I withdraw it? ------ From research!csnet!sierra.stanford.edu!BUNEMAN Thu 30 Jul 87 20:56:07-PDT From: Oscar Buneman Subject: FHT legalities To: ehg%research.att.com@csnet-relay cc: bracewell@star.stanford.edu Eric Grosse: Our reaction to Ron Bracewell's success over patenting an algorithm (not a machine) has been generally favorable. We do not think that it means the algorithm can only be sold (not freely distributed). I see no difference between patenting and copyrighting an idea. Both simply mean acknowledging authorship. I am sure Ron Bracewell would be the first to raise objections if you withdrew my FAST HARTLEY TRANSFORM from netlib distribution. I wonder how many algorithms there are in NETLIB which have been published under some copyright somewhere or other! Is the law going to come after us because we are freely copying such material into and out of NETLIB? Notice, by the way, that patents and copyrights provide financial benefits, if any, to the author's institution and very rarly or minimally to the author himself! In our case, the encouragement to try and patent the algorithm, and the initiative, came from the university. I must admit that sooner or later there will be some serious trouble over patenting or copyrighting software - but in the meantime all original authors will be the happier the more their ideas are copied and used - with only the acknowledgment of their authorship. Oscar Buneman ------- From research!csnet!star1.stanford.edu!BRACEWELL Mon 3 Aug 87 15:33:44-PDT From: "Ronald N. Bracewell" Subject: Hartley Legalities To: ehg%research.att.com@csnet-relay The Hartley algorithm has been published by the IEEE Proceedings and by Oxford University Press, with copyright statements attached. If you are concerned, telephone Lisa Kuuttila (pron. KOO-tila) of Stanford Office of Technology Licensing at (415) 723-0651. R.N. BRACEWELL ------- 13:30 4 Aug 1987 ehg Phoned Kuuttila; she was in a meeting and will call me back. 15:47 Kuuttila says that Stanford retains the copyright and patent and that a license agreement is probably needed for netlib. But if Prof. Bracewell wants his code distributed that way, she'll find a way to make it legal. She says that she will contact Bracewell and then get back to me. 29 Sep 87 ehg, following phone conversation with Kuttila m dongarra Lisa Kuutila of Stanford's Technology Licensing Office insists that distributing Buneman's fast Hartley transform program is in violation of Stanford's patent. I'm therefore removing that file from netlib and will replace it with a licensing statement she will send me. From arpa!SIERRA!BUNEMAN Wed Jul 6 21:31:31 PDT 1988 Received: by sierra.STANFORD.EDU (3.2/4.7); Wed, 6 Jul 88 21:31:32 PDT Date: Wed 6 Jul 88 21:31:31-PDT From: Oscar Buneman Subject: [D5.N60@Forsythe.Stanford.EDU: ] To: ehg@research.att.com Message-Id: <584253091.0.BUNEMAN@SIERRA> Mail-System-Version: Eric G.: I am sneding you, attavhched, a new version of the Fast Hartelley Transformending you, attached, a new version of the Fast Hartley Transform which contains a preabmble required by Stanford"s's Patent Office. I meant to send you this sometime ago but was worried about how you would react! We have had a session with our patent attorneys with legalpeople forom the Law School as weell as the Philosophy Department - the latter concerned with the ethics of patenting ideas! The patent people turned out to be quite benevolent . They are genuinely concerned over the commercial exploitatinon of software with no benefir t for the oribginators. ANd theynd they encourage ed me to put the FHT on a network, provided their preamble goes with it. Incidentally, the simple versionis simple version of the FHT is for N= N = a power of 2 between 16 and 80192. I reckon for N up to 8 people would write their own transform and for N = 18384638144 and up they would extend the tables in obvious manner initial tables in obvious manner. I dont know of any FHT for other prime factors - someone will have to take the trouble of translating from corresponding FFT's. I have written FHT's in assembler for CRAY s (XMP and CRAY-2), suitable for myultidimensional taransforming which are available of the San Diego NSF system and on the Livermore system (NMFECC). I also have FHT-s in IBM PC Assembler for anyone who wants such. Oscar B. --------------- ------------end of Hartley correspondence--------- axxbc permission by telephone call (from Eric to Stephen, 25 Nov 1987) files copied to research from anl-mcs, March 1988: hssxev ov-bib iqpack C From arpa!antares.mcs.anl.gov!dongarra Sun Mar 12 10:26:26 CST 1989 Received: from homesun (homesun.mcs.anl.gov) by antares.mcs.anl.gov (4.0/SMI-DDN) id AA23928; Sun, 12 Mar 89 10:20:08 CST Received: by homesun (3.2/SMI-3.0DEV3) id AA17755; Sun, 12 Mar 89 10:26:26 CST Date: Sun, 12 Mar 89 10:26:26 CST From: dongarra@antares.mcs.anl.gov (jack dongarra) Message-Id: <8903121626.AA17755@homesun> To: pipe!ehg@research.att.com Subject: Re: SONEST I'll put it into netlib under misc, try: send sonest from misc or for complex matrices send conest from misc We are using this approach in LAPACK. Jack lsqr bug fix received from David Gay, approved by Saunders, 12 May 89 From cs.utk.edu!dongarra Sun Apr 28 10:40:15 0400 1991 Date: Sun, 28 Apr 91 10:40:15 -0400 From: dongarra@cs.utk.edu To: agw@cs.columbia.edu Subject: Re: Want to make a NA-NET-like clone Cc: na.grosse@surfer.EPM.ORNL.GOV Good idea, I have put the software required to run the na-net in netlib. You can get it by sending mail to netlib@ornl.gov and in the mail message type: send nanet-shar from misc There is a README file with details on the install. Good luck, Jack .