From chemistry-request@www.ccl.net Sun Apr 11 00:04:19 1999 Received: from paracelsus.fdm.uni-freiburg.de (paracelsus.fdm.uni-freiburg.de [132.230.97.1]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with SMTP id AAA02694 Sun, 11 Apr 1999 00:04:18 -0400 (EDT) Received: (from hubi@localhost) by paracelsus.fdm.uni-freiburg.de (SMI-8.6/8.6.9) id GAA13645; Sun, 11 Apr 1999 06:04:13 +0200 Date: Sun, 11 Apr 1999 06:04:13 +0200 (MET DST) From: "W. Huber" To: chemistry@www.ccl.net Subject: molecular similarity Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Dear CCLers, I am looking for a method, or a software packet, that compares a "drug-like" molecule against a database of other such molecules. The comparison should be based on 3D structure and take into account conformational flexibility of the molecules. Additionally, I would like to use this together with a virtual database, i.e. one that consists of fragments that are assembled according to certain rules. It seems that the combinatorics of all that make compute time be an important issue. I have found a number of references in the literature that offer partial solutions (SEAL, FLEXS, PLM, 3D-MCSS), so I would like to ask: - what is used by practioners in the field, - how they compare against each other (is there a relevant review?), - have I maybe missed the one method that does it all. Best regards... W. Huber hubi@fdm.uni-freiburg.de Freiburg Center for Data Analysis and Modelling Germany From chemistry-request@www.ccl.net Wed Apr 7 11:11:10 1999 Received: from web601.mail.yahoo.com (web601.mail.yahoo.com [128.11.68.165]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with SMTP id LAA05605 Wed, 7 Apr 1999 11:11:10 -0400 (EDT) Message-ID: <19990407151137.15824.rocketmail@web601.mail.yahoo.com> Received: from [148.234.75.80] by web601.mail.yahoo.com; Wed, 07 Apr 1999 08:11:37 PDT Date: Wed, 7 Apr 1999 08:11:37 -0700 (PDT) From: Victor Manuel "Rosas-García" Subject: Imaginary freqs in Spartan (thanks, but...) To: CCL mailing list MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Hello everybody, About a couple of weeks ago I asked a question about getting rid of imaginary frequencies in Spartan. First I want to thank: Stefan Fau Ned Charles Haubein Artem Masunov Edmund Ndip for their interest and suggestions. I'm thinking that I was not explicit enough regarding my problem, so here goes the situation: I've been doing calculations of solvation energies using semiempirical methods for about two years now. At first, I used COSMO in MOPAC93, and it was common to get "minima" with imaginary frequencies. Within MOPAC93, it was possible to tighten up some criteria, change the optimization method, or (last resort) run an IRC to get rid of those imaginary freqs. It was painful but it usually worked (although in some recalcitrant cases I was unable to eliminate all the imaginary freqs). Now I'm using the solvent models in Spartan and the problem with "minima with imaginary freqs" has reappeared. So my questions are: a) Is there a way to run an IRC within Spartan? b) Are solvent models bound to produce spurious imaginary frequencies? if so, why? Thanx Victor === Victor M. Rosas García Computational Chemistry Apprentice e-mail: quimico69@yahoo.com _________________________________________________________ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com From chemistry-request@www.ccl.net Wed Apr 7 15:26:25 1999 Received: from chris2.u-strasbg.fr (maerker@chris2.u-strasbg.fr [130.79.34.140]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id PAA07839 Wed, 7 Apr 1999 15:26:24 -0400 (EDT) Received: (from maerker@localhost) by chris2.u-strasbg.fr (8.8.8/8.8.8) id VAA31661 for chemistry@www.ccl.net; Wed, 7 Apr 1999 21:29:25 +0200 From: Christoph Maerker Message-Id: <199904071929.VAA31661@chris2.u-strasbg.fr> Subject: summary: object-oriented programming tools To: chemistry@www.ccl.net (ccl) Date: Wed, 7 Apr 1999 21:29:25 +0200 (MEST) Organization: Laboratoire de Chimie biophysique, Institut Le Bel, Universite Louis Pasteur Postal-Address: 4, rue Blaise Pascal, F- 67000 Strasbourg FRANCE Phone: +33/3-88-41-53-19 Fax: +33/3-88-60-63-83 X-Mailer: ELM [version 2.4ME+ PL37 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 2200 Dear netters, here is the summary about object-oriented programming tools. Thanks to Tim and James who had replied, see below. The problem with Fortran (77) is that it is not an object-oriented language per se. Nevertheless, it seems that some people have found have found a workaround to make life easier. Bien amicalement, Christoph ======================================= #1 Return-Path: Christoph, I don't know of any CASE tools for Fortran, but Rational Rose is considered to be the best for OO design and it's the best selling. TogetherJ works on Java, but they may also have a C++ version. Both packages are very expensive, running about $2500-3600 for node locked licenses. Tim ======================================= #2 From jamesx@bohr.chem.mun.ca Wed Mar 31 05:58:39 1999 Hello, I have a solution that is not quite as magical as you would hope, but it is a step in the right direction. The current version of Mungauss (formerly MONSTORGAUSS) is writen using OSIPE. OSIPE is a set of low level FORTRAN object oriented tools. OSIPE manages the creation, cataloging and retrieving objects created by the Mungauss program. The programmer does not need to to know or care about the route - OSIPE will handle the building of any objectss you will need. You need only ask for a desired object, and OSIPE will determine whether it has been built: if it is, it will be retrieved, else it will call the necessary routines to build it. Objects in Mungauss can be scalars (e.g. total energy, dipole moment), vectors (e.g. dipole vectors, Z-matrix parameters), matrices (e.g. density matrix, Cartesian coordinates), or tensors. The primary author of OSIPE and Mungauss is Dr. Raymond Poirier (Memorial University of Newfoundland, Canada). If you are interested in this code, I would recommend you contact him. His e-mail address is rpoirier@morgan.ucs.mun.ca Good luck, James Xidos Graduate Student Memorial University of Newfoundland < "You have the right to free speech - as long as >< James Xidos > < you're not dumb enough to actually try it!" >< -Newfie Greek- > < -The Clash, "Know Your Rights" >< -Chemist- > From chemistry-request@www.ccl.net Wed Apr 7 18:10:03 1999 Received: from cornelius.INS.CWRU.Edu (root@cornelius.INS.CWRU.Edu [129.22.8.216]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id SAA09266 Wed, 7 Apr 1999 18:10:02 -0400 (EDT) Received: from raman2.cwru.edu (chem51374.CHEM.CWRU.Edu [129.22.200.174]) by cornelius.INS.CWRU.Edu with SMTP (8.8.8+cwru/CWRU-3.4) id SAA20306; Wed, 7 Apr 1999 18:10:05 -0400 (EDT) (from hxt10@po.cwru.edu for ) Message-ID: <003d01be814b$ffcce160$aec81681@cwru.edu> From: "Gavin Tsai" To: Subject: Jaguar Date: Wed, 7 Apr 1999 18:11:41 -0500 MIME-Version: 1.0 Content-Type: multipart/mixed; boundary="----=_NextPart_000_0039_01BE8122.168A8300" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.00.2014.211 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2014.211 This is a multi-part message in MIME format. ------=_NextPart_000_0039_01BE8122.168A8300 Content-Type: multipart/alternative; boundary="----=_NextPart_001_003A_01BE8122.168A8300" ------=_NextPart_001_003A_01BE8122.168A8300 Content-Type: text/plain; charset="big5" Content-Transfer-Encoding: quoted-printable Hi, Everyone: I just read this paper " Correlated ab initio electronic structure calculations for large molecule" JPC A V103,1913, 1999. The paper shows = the Jaguar program is about 10 times faster than G92. Can anyone share me = the performance of this program? Any comment is welcomed! Thanks! Gavin Tsai ------=_NextPart_001_003A_01BE8122.168A8300 Content-Type: text/html; charset="big5" Content-Transfer-Encoding: quoted-printable
Hi, Everyone:

    I = just read=20 this paper " Correlated ab initio electronic structure
calculations = for large=20 molecule" JPC A V103,1913, 1999.  The paper shows the
Jaguar = program is=20 about 10 times faster than G92. Can anyone share me the
performance = of this=20 program?

Any comment is welcomed!
Thanks!

Gavin=20 Tsai
------=_NextPart_001_003A_01BE8122.168A8300-- ------=_NextPart_000_0039_01BE8122.168A8300 Content-Type: text/x-vcard; name="Hui-Hsu (Gavin) Tsai.vcf" Content-Transfer-Encoding: quoted-printable Content-Disposition: attachment; filename="Hui-Hsu (Gavin) Tsai.vcf" BEGIN:VCARD VERSION:2.1 N:Tsai;Hui-Hsu (Gavin) FN:Hui-Hsu (Gavin) Tsai NICKNAME:Gavin ORG:CWRU, Chemistry Department TITLE:Ph.D. Candidate NOTE;ENCODING=3DQUOTED-PRINTABLE:Do molecules calculate the chemical = reaction laws before they decide=3D0D=3D0Aho=3D w to react to each others?=3D0D=3D0A TEL;HOME;VOICE:216-229-5913 ADR;HOME:;;2416 Derbyshire Road;Cleveland;OH;44106;USA LABEL;HOME;ENCODING=3DQUOTED-PRINTABLE:2416 Derbyshire = Road=3D0D=3D0ACleveland, OH 44106=3D0D=3D0AUSA X-WAB-GENDER:2 EMAIL;PREF;INTERNET:hxt10@po.cwru.edu REV:19990407T231141Z END:VCARD ------=_NextPart_000_0039_01BE8122.168A8300-- From chemistry-request@www.ccl.net Wed Apr 7 18:39:12 1999 Received: from post.queensu.ca (post.QueensU.CA [130.15.126.6]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id SAA09454 Wed, 7 Apr 1999 18:39:11 -0400 (EDT) Received: from quchem.chem.QueensU.CA (CHEM.QueensU.CA [130.15.98.1]) by post.queensu.ca (8.9.0/8.9.0) with SMTP id SAA29834 for ; Wed, 7 Apr 1999 18:39:13 -0400 (EDT) Received: from quchem by quchem.chem.QueensU.CA (4.1/SMI-4.0) id AA23825; Wed, 7 Apr 99 18:48:27 EDT Date: Wed, 7 Apr 1999 18:48:27 -0400 (EDT) From: Tek X-Sender: kims@quchem To: chemistry@www.ccl.net Subject: summary: Computational Chemistry vs Theoretical Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi all! Thanks to all those who have contributed. Here is a summary of all the responses. Stephen Kim Department of Chemistry Queen's University Kingston, Ontario K7L-3N6 -------------------------------------------------------------------------- From: Rene Fournier To: STK <3tksk@qlink.queensu.ca> Cc: renef@YorkU.CA Subject: Re: CCL:CCL: Computational Chemist vs Theoretical Chemist In-Reply-To: Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Status: RO Hello, There is no universally accepted answer to that question, but it's a very interesting question. There was a debate about 2 years ago on the CCL on the difference between "computational chemistry" and "molecular modeling" and it went into "theoretical chemistry" also. I collected some opinions voiced on CCL, edited them, and wrote a short essay in an effort to organize my own thoughts on that. If you're interested it's on my web page: http://www.science.yorku.ca/chem/profs/renef/whatiscc.html#descriptioncc In short, my definition would be: - a computational chemist is someone who turns theories into numbers (and usually explains, or compares to, experiments); or someone who devises approximations or better algorithms or better programs to make calculations faster. - a theoretical chemist is someone who makes new theories: there are very, very few persons in that category in my view. A good test of "computation" vs "theory" is to imagine what would happen if we suddenly had INFINITELY fast and powerful computers (the starting point of a discussion in Physics Today some time back). If we had INFINITELY fast and powerful computers that were extremely easy to use, computational chemistry would quickly disappear. There would be no need at all for expert computational chemists; just click on the option "full CI, complete basis set, molecular dynamics with 10**50 time steps, with explicit treatment of solvent and impurities, etc..." and you instantly get all properties you want from the computer, with as much accuracy as you want. But there would still be a need for theoretical chemists. Maybe a more practical test of "computation" vs "theory" could be this: the work of a theoretical chemist would not be affected much by having exclusive access to a 256-node super-computer instead of an old standalone PC; but the work of a computational chemist would be very much affected by that. -- Rene Fournier. -------------------------------------------------------------------- | Rene Fournier | Bureau/Office 303 Petrie | | Chemistry Dpt, York University | (416) 736 2100 Ext. 30687 | | 4700 Keele Street, Toronto | FAX: (416) 736-5936 | | Ontario, CANADA M3J 1P3 | e-mail: renef@yorku.ca | -------------------------------------------------------------------- Date: Fri, 26 Mar 1999 09:43:43 +36000 From: Richard Wood To: STK <3tksk@qlink.queensu.ca> Cc: chemistry@www.ccl.net Subject: Re: CCL:CCL: Computational Chemist vs Theoretical Chemist In-Reply-To: Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Status: RO X-Status: Hi- I would say that theoretical chemistry involves either a) developing theories, or b) using theories that involve quantum chemistry. I would say that computational chemistry is the use of tools, be they quantum mechanical, or, mostly semi-empirical. Thus, IMO, computational chemistry is a sub-field of theoretical chemistry. In my "work" as a Ph. D. candidate, I do MD simulations of biologically important molecules. These simulations involve classical mechanics and other semi-empirical techniques. When people ask what I do, I say "I am a computational chemist", because i don't develop theory, I just use what theory has already been developed by others. Richard From: schrecke@t12.lanl.gov (Georg Schreckenbach) Subject: Re: CCL:CCL: Computational Chemist vs Theoretical Chemist Status: RO X-Status: Hi Stephen, unfortunately, I don't have any "authorative" comments. However, I'd like to point out that you can clearly do "theoretical chemistry" with just paper and pencil, i.e., without a computer. In that sense, "theoretical chemistry" would be the more general descriptor of the two, and it would comprise "computational chemistry" as a sub-area. Btw., I view myself as a "theoretical" AND a "computational" chemist ... Best regards, Georg P.S. please summarize ... -- ============================================================== Dr. Georg Schreckenbach Tel: (USA)-505-667 7605 Theoretical Chemistry T-12 FAX: (USA)-505-665 3909 M.S. B268, Los Alamos National E-mail: schrecke@t12.lanl.gov Laboratory, Los Alamos, New Mexico, 87545, USA Internet: http://www.t12.lanl.gov/home/schrecke/ ***new location!*** ============================================================== From: "Preston J. MacDougall" Subject: Re: Computational Chemist vs Theoretical Chemist In-Reply-To: X-Sender: pmacdougall@acad1.mtsu.edu To: STK <3tksk@qlink.queensu.ca> Message-Id: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Status: RO X-Status: Hello Stephen, I am sure that there is no consensus on what differentiates a computational chemist from a theoretical one. I am equally sure that many chemists wear both of these hats. I would suggest that computational chemistry is to theoretical chemistry what synthesis is to analysis in other branches of chemistry. Computational chemistry involves applied mathematics, computer programming and lots of number crunching to "construct" wavefunctions, potential energy surfaces et cetera, whereas theoretical chemistry involves much of the same "apparatus" but the goal is to provide answers to questions couched in the language of chemistry, usually extracted fromthe fruits of computational chemistry but sometimes with paper and pencil! Queen's own Vedene Smith is an example of a theoretical chemist of the finest calibre. Please say hello to him for me! Is Minhhuy Ho still there? If so please say "Hi" to him as well. Hartmut? ditto! Please summarize. I am sure many people are curious to see what other people say about this "word problem". Best wishes from Middle Tennessee, Preston J. MacDougall Assistant Professor Department of Chemistry, Box X101 Middle Tennessee State University Murfreesboro, TN 37132 http://www.mtsu.edu/~chem/ Date: Fri, 26 Mar 1999 15:29:29 +0100 (CET) From: Stefan Konietzny To: STK <3tksk@qlink.queensu.ca> Cc: chemistry@www.ccl.net Subject: Re: Computational Chemist vs Theoretical Chemist In-Reply-To: Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Status: RO X-Status: Hello everybody, If you take this question serious for yourself, stop reading and delete the message NOW. we are using a computer to do theoretical chemistry. I think there was a time when people did theoretical chemistry without a computer. They were theoretical chemists but not computational chemists. So what about Computational Theoretical Chemist. But Computation implements that it is theoretical; no? Or is it only theoretical, when it is really theoretical, i mean when you calculate things that man has never seen before. We can use the term Applied Theoretical Chemist, for those who calculate things the Lab-Chemist (or is it a Synthetical Chemist?) has synthesized. But there is organic, inorganic, physical... Oh my good, in the end its Physical Chemistry! I think i am falling more and more into an identity crisis. I am calculating (we are all calculators!) inorganic molecules (although there is carbon in it, they are inorganic). Until i know who or what i am, i call myself: Molecular Modeller in Applied Theoretical Inorganic Computational Quantum Chemistry. :-) Stefan. !!!!! reply to: konietz@chemie.uni-kl.de !!!!! finger or talk request: konietz@oktarin.chemie.uni-kl.de **************************************************************** * Stefan Konietzny * *--------------------------------------------------------------* *Fachbereich Chemie, AK Frank | Dept. of Chemistry * *Universitaet Kaiserslautern | University of Kaiserslautern * *Erwin-Schroedingerstr. | Erwin-Schroedingerstr. * *67663 Kaiserslautern | D-67663 Kaiserslautern * *0631/205-2964 | +49-631/205-2964 * **************************************************************** From: Jose Luis Carreon Macedo To: STK <3tksk@qlink.queensu.ca> Subject: Re: CCL:CCL: Computational Chemist vs Theoretical Chemist In-Reply-To: Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Status: RO X-Status: Hi! I had the same question and I found useful the next FAQ of Dr. Alan Shusterman. http://www.reed.edu/~alan/324/faq.html I hope this help you. --------------)(--------------- Jose Luis Carreon Macedo Mexico UNAM Facultad de Quimica Fisica y Quimica Teorica DEPg joseluis@eros.pquim.unam.mx From: Brian Salter-Duke To: STK <3tksk@qlink.queensu.ca> Subject: Re: CCL:CCL: Computational Chemist vs Theoretical Chemist References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.93.2i In-Reply-To: ; from STK on Thu, Mar 25, 1999 at 04:43:04PM -0500 Status: RO X-Status: I'm both. When I am devising or improving methods I'm a theoretical chemist. When I am using the methods I am a computational chemist. Also computational chemistry sounds more sexy than theoretical chemistry so it being more widely used. Cheers, Brian. -- Associate Professor Brian Salter-Duke (Brian Duke) School of Mathematical and Physical Sciences, Northern Territory University, Darwin, NT 0909, Australia. Phone (61) (0)8-89466702 e-mail: b_duke@lacebark.ntu.edu.au or b_duke@quoll.ntu.edu.au Date: Mon, 29 Mar 1999 11:41:29 -0500 (EST) From: "Prof. Robert Topper" X-Mailer: ELM [version 2.4 PL23] Sender: Computational Chemistry List Errors-To: ccl@www.ccl.net Precedence: bulk Status: RO X-Status: Hello all I think Stefan's got it just right (especially about the need to have a sense of humor!). I was trained as a theoretical chemist. To me, what this means is that I was trained to push a pencil with the goal of developing certain mathematical aspects of chemistry. This includes making theoretical improvements to transition-state theory, developing new quantum path-integral representations, inventing a new theory of the liquid state, working out better ways to set up the integrals for 5th-order perturbation theory, and so forth. Of course, I eventually use computers to implement all of the above. But experimentalists use computers too. IMHO, this may also justifiably be called "computational chemistry" (I know, not everyone agrees, this is an old debate, just my $0.02, sincere apologies to all who find my heresy distasteful). Novices tend to divide science into theory and experiment, and therefore "computation must be theory." As Stefan mentions, this is a historical bias. Although computer simulation work is not experiment, it isn't exactly theory either. When I was a student we used to call computer simulations "numerical experiments" (these days, maybe "virtual experiments" would be trendier), which could be tested either against theory or experiment. In Einstein's day things were different, but science now consists of theory, experiment and simulation. Some people work in one, two or even all three of these areas (I myself get into the lab once in a blue moon). So am I a theoretical physical chemist? A computational chemist? A chemical physicist? A computational molecular thermodynamicist? Personally, I don't think of myself as any of the above. I'm a chemist. ***************************************************************************** Robert Q. Topper email: topper@cooper.edu Department of Chemistry phone: (212) 353-4378 School of Engineering fax: (212) 353-4341 The Cooper Union subway: take the 6 to Astor Place 51 Astor Place or the N/R to 8th street New York, NY 10003 and you're there! ***************************************************************************** http://www.cooper.edu/engineering/chemechem/ ***************************************************************************** From: Stefan Konietzny To: chemistry@ccl.net Subject: CCL:Re: Computational Chemist vs Theoretical Chemist Sender: chemistry-request@www.ccl.net Errors-To: ccl@ccl.net Precedence: bulk Status: RO X-Status: Hello everybody, If you take this question serious for yourself, stop reading and delete the message NOW. we are using a computer to do theoretical chemistry. I think there was a time when people did theoretical chemistry without a computer. They were theoretical chemists but not computational chemists. So what about Computational Theoretical Chemist. But Computation implements that it is theoretical; no? Or is it only theoretical, when it is really theoretical, i mean when you calculate things that man has never seen before. We can use the term Applied Theoretical Chemist, for those who calculate things the Lab-Chemist (or is it a Synthetical Chemist?) has synthesized. But there is organic, inorganic, physical... Oh my good, in the end its Physical Chemistry! I think i am falling more and more into an identity crisis. I am calculating (we are all calculators!) inorganic molecules (although there is carbon in it, they are inorganic). Until i know who or what i am, i call myself: Molecular Modeller in Applied Theoretical Inorganic Computational Quantum Chemistry. :-) Stefan. !!!!! reply to: konietz@chemie.uni-kl.de !!!!! finger or talk request: konietz@oktarin.chemie.uni-kl.de **************************************************************** * Stefan Konietzny * *--------------------------------------------------------------* *Fachbereich Chemie, AK Frank | Dept. of Chemistry * *Universitaet Kaiserslautern | University of Kaiserslautern * *Erwin-Schroedingerstr. | Erwin-Schroedingerstr. * *67663 Kaiserslautern | D-67663 Kaiserslautern * *0631/205-2964 | +49-631/205-2964 * **************************************************************** Date: Fri, 26 Mar 1999 09:43:43 +36000 From: Richard Wood To: chemistry@ccl.net Subject: CCL:Re: CCL:CCL: Computational Chemist vs Theoretical Chemist Sender: chemistry-request@www.ccl.net Errors-To: ccl@ccl.net Precedence: bulk Status: RO X-Status: Hi- I would say that theoretical chemistry involves either a) developing theories, or b) using theories that involve quantum chemistry. I would say that computational chemistry is the use of tools, be they quantum mechanical, or, mostly semi-empirical. Thus, IMO, computational chemistry is a sub-field of theoretical chemistry. In my "work" as a Ph. D. candidate, I do MD simulations of biologically important molecules. These simulations involve classical mechanics and other semi-empirical techniques. When people ask what I do, I say "I am a computational chemist", because i don't develop theory, I just use what theory has already been developed by others. Richard Date: Mon, 29 Mar 1999 09:43:15 +1000 To: STK <3tksk@qlink.queensu.ca> From: "Kieran F Lim (Lim Pak Kwan)" Subject: Re: CCL:CCL: Computational Chemist vs Theoretical Chemist Status: RO X-Status: A theoretical chemist works with theory -- devleoping theory, writing equations, developing new models etc. often, computation is required to prove/apply the theory. quite often theory itself is the thing being tested. A theoretical paper will have a section entitled "theory" which is new theory. (cf. a synthetic paper reports a new synthesis.) A computational chemist uses theoretical methods to compute/calculate something. the theory itself is usually not the thing being tested. My own test is a) if the user doing the calculation has in some way "value-added" to the calculation then that is theoretical chemistry. b) on the other hand, if the user is just using a "black box" to do a calculation, then that is computational chemistry. Kieran ------------------------------------------------------------ Dr Kieran F Lim Biol. and Chemical Sciences (Lim Pak Kwan) Deakin University ph: + [61] (3) 5227-2146 Geelong VIC 3217 fax: + [61] (3) 5227-1040 AUSTRALIA mailto:lim@deakin.edu.au http://www.deakin.edu.au/~lim Dear Stephen, there is no "hard" line between the two, but in general, a theoretical chemist is concerned more with the "pure theory" and the mathematical side of quantum chemistry to develop new or revise existing approaches to solving the Schroedinger equation; the focus sometimes is the computational implementation, i.e., the algorithms and programs. A computational chemist, on the other hand, mostly uses the codes provided by theoreticians to study chemical problems, with a strong emphasis on the chemical interpretation and not so much on the underlying theory. Does this help? Peter //// ___|--00___________________________________________________________________ C ^ Dr. Peter R. Schreiner, Ph.D. (Univ. Georgia) \ ~/ Institut fuer Organische Chemie http://www.gwdg.de/~pschrei <><> Georg-August Universitaet Goettingen Tammannstr. 2 Phone: +49-(0)551-393287 D-37077 Goettingen, Germany FAX: +49-(0)551-399475 ___________________________________________________________________________ From: "Helder Marques" Organization: CHEMISTRY - WITS UNIVERSITY To: STK <3tksk@qlink.queensu.ca> Date: Fri, 26 Mar 1999 11:46:54 GMT + 2:00 Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7BIT Subject: Re: CCL:CCL: Computational Chemist vs Theoretical Chemist X-Confirm-Reading-To: "Helder Marques" X-Pmrqc: 1 Priority: normal In-Reply-To: X-Mailer: Pegasus Mail for Windows (v3.01b) Message-Id: <66329B3E07@aurum.chem.wits.ac.za> Status: RO X-Status: This is a personal opinion and not to be confused with any authoritative definition, but I would ahve thought a theoretical chemist was one concerned with a theoretical description of some aspect of chemistry whereas a computational chemist is one who would use computational procedures - perhaps well-established - as a *tool* to answer specific questions. Hence the developers of MOPAC might consider themselves theoretical chemists, but users of the toolbox would be computational chemists - or is that too simplistic? Helder ____________________________________________________________________ Prof. Helder M. Marques Room C301, Department of Chemistry, University of the Witwatersrand, P.O. Wits, Johannesburg, South Africa, 2050 Fax: +27+11+339-7967; Tel: +27+11+716-2303 http://www.chem.wits.ac.za Date: Mon, 29 Mar 1999 14:46:30 +36000 From: Richard Wood To: "Prof. Robert Topper" Cc: CHEMISTRY@www.ccl.net Subject: CCL:Computational Chemist vs Theoretical Chemist In-Reply-To: <199903291641.LAA07414@robin.cooper.edu> Message-Id: Sender: Computational Chemistry List Errors-To: ccl@www.ccl.net Precedence: bulk Status: RO X-Status: Dr. Topper- While I agree with you when you say "I am a chemist", as I have several years of experience working in the IR/CD area, as well as three plus years of experience doing MD simulations, I must disagree with you when you imply that the distinction between being a "theoretical chemist" vs. a "computational chemist" doesn't matter. It matters to me (however, don't think that I lose sleep over it) because to me developing a theory and using it are two ENTIRELY different things. I think that most any chemist can USE a theory; only a special "bird" can develop theory. Let me give you an example. A few years back, I was measuring IR/CD spectra of some (chiral) tetrasubstituted biphenyls, and also attempting to calculate the observed IR/CD spectra by varying the angle between the two rings (in other words, trying to find the angle that gave me the closest match with my experimental results). We were using a simple coupled oscillator model to calculate the spectra, and my calculations of the spectra were too small, regardless of what angle I picked. I then came to the conclusion that the distance between the too dipoles was too small, and thus R+ and R- were cancelling out, decreasing their magnitude. I also concluded that my theory needed some modification. Lacking the necessary theoretical "ability", my project basically crashed and burned. I guess that you are probably aware being a physical chemist, that if you are going to try and publish experimental results, you better have very good agreement between experiment and calculations. Richard On Mon, 29 Mar 1999, Prof. Robert Topper wrote: Date: 29 Mar 1999 16:59:58 -0500 From: "Boyd" Return-Receipt-To: "Boyd" Subject: definitions To: "Kim, Stephen T." <3tksk@qlink.queensu.ca> X-Mailer: Mail*Link SMTP for Quarterdeck Mail; Version 4.1.0 Status: RO X-Status: Hi Stephen, On Friday you asked about the distinction between a computational chemist and a theoretical chemist. There is a third term also: molecular modeler. You have probably gotten a lot of opinions. Here are mine. The term theoretical chemist is the oldest term, and many academic quantum chemists are quite happy to call themselves theoretical chemist. This term is fine for academics, but few companies want to hire a "theoretical" person. Therefore most people in industry call themselves computational chemist. They do a lot of work in addition to just quantum chemistry. Organic chemists mainly interested in manipulating 3D structures of molecules call the computational chemists "molecular modelers". However, the term computational chemist is a more general term, and encompasses a lot of computed-based approaches, such as QSAR and bioinformatics, as you mention. These terms have been discussed in the book series I edit. For instance, take a look at the Preface of Vol. 1, Reviews in Computational Chemistry (1990). Your university library should have it. Let me know if you need further clarification. Sincerely, Don Donald B. Boyd, Ph.D. Editor, Reviews in Computational Chemistry Editor, Journal of Molecular Graphics and Modelling Research Professor of Chemistry Department of Chemistry Indiana University-Purdue University at Indianapolis 402 North Blackford Street Indianapolis, Indiana 46202-3274, U.S.A. Telephone 317-274-6891, FAX 317-274-4701 E-mail boyd@chem.iupui.edu Web http://chem.iupui.edu/rcc/rcc.html Date: Mon, 29 Mar 1999 18:16:32 -0500 (EST) From: "Prof. Robert Topper" X-Mailer: ELM [version 2.4 PL23] Sender: Computational Chemistry List Errors-To: ccl@www.ccl.net Precedence: bulk Status: RO X-Status: Hi again all, Sorry for the bandwidth. Just want to point out that I personally feel that there is a very big difference between theoretical chemistry and computational chemistry, and that was in fact what I was trying to say - perhaps too obliquely. My parting comment was merely put in to point out that I have no particular axe to grind on this issue, as I have done work in all three areas (although I have done most of my work in the "theoretical" and "simulation" ends of things). Any inferences of other, guessed-at meanings are simply incorrect. Best, Robert ***************************************************************************** Robert Q. Topper email: topper@cooper.edu Department of Chemistry phone: (212) 353-4378 School of Engineering fax: (212) 353-4341 The Cooper Union subway: take the 6 to Astor Place 51 Astor Place or the N/R to 8th street New York, NY 10003 and you're there! ***************************************************************************** http://www.cooper.edu/engineering/chemechem/ ***************************************************************************** > > Dr. Topper- > While I agree with you when you say "I am a chemist", as I have > several years of experience working in the IR/CD area, as well as three > plus years of experience doing MD simulations, I must disagree with you > when you imply that the distinction between being a "theoretical chemist" > vs. a "computational chemist" doesn't matter. It matters to me (however, > don't think that I lose sleep over it) because to me developing a theory > and using it are two ENTIRELY different things. I think that most any > chemist can USE a theory; only a special "bird" can develop theory. > > Let me give you an example. A few years back, I was measuring IR/CD > spectra of some (chiral) tetrasubstituted biphenyls, and also attempting to > calculate the observed IR/CD spectra by varying the angle between the two > rings (in other words, trying to find the angle that gave me the closest > match with my experimental results). We were using a simple coupled > oscillator model to calculate the spectra, and my calculations of the > spectra were too small, regardless of what angle I picked. I then came > to the conclusion that the distance between the too dipoles was too > small, and thus R+ and R- were cancelling out, decreasing their > magnitude. I also concluded that my theory needed some modification. > Lacking the necessary theoretical "ability", my project basically crashed > and burned. I guess that you are probably aware being a physical > chemist, that if you are going to try and publish experimental results, > you better have very good agreement between experiment and calculations. > > Richard From chemistry-request@www.ccl.net Thu Apr 8 01:30:54 1999 Received: from Raven.und.ac.za (Raven.und.ac.za [146.230.128.50]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id BAA10746 Thu, 8 Apr 1999 01:30:50 -0400 (EDT) Received: from gwise-smtp.und.ac.za (GWise-SMTP.und.ac.za [146.230.128.63]) by Raven.und.ac.za (8.9.3/8.7.3) with SMTP id HAA06181 for ; Thu, 8 Apr 1999 07:30:47 +0200 (SAT) Received: from Routing-Domain-Message_Server by gwise-smtp.und.ac.za with Novell_GroupWise; Thu, 08 Apr 1999 07:30:46 +0200 Message-Id: X-Mailer: Novell GroupWise 5.2 Date: Wed, 07 Apr 1999 15:59:38 +0200 From: "Gert Kruger" To: chemistry@www.ccl.net Subject: Re: Spartan (unix version) on Linux? Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Disposition: inline Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by www.ccl.net id BAA10747 Dear CCL, Has anyone tried to run Spartan (unix) on a PC with Linux as operating system? Any hints would be appreciated. Best wishes Gert Kruger ___________________________________ Dr HG Kruger, Chemistry, University of Natal P.O. Box 18091, Dalbridge 4014 Durban, South Africa. tel: +27-31-2603090, fax: +27-31-2603091 email: kruger@scifs1.und.ac.za ___________________________________ From chemistry-request@www.ccl.net Thu Apr 8 14:39:20 1999 Received: from mercury.chem.nwu.edu (rassolov@mercury.chem.nwu.edu [129.105.116.2]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id OAA15580 Thu, 8 Apr 1999 14:39:19 -0400 (EDT) Received: from localhost by mercury.chem.nwu.edu (8.9.1/8.9.1) with SMTP id NAA17249 for ; Thu, 8 Apr 1999 13:41:46 -0500 (CDT) Date: Thu, 8 Apr 1999 13:41:46 -0500 (CDT) From: Vitaly Rassolov Reply-To: Vitaly Rassolov To: chemistry@www.ccl.net Subject: SUMMARY: lin.alg.? In-Reply-To: <199904072123.QAA04109@harvey.caam.rice.edu> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Yesterday I posted a question about a search for the basis set in the orthogonal subspace. Many thanks to everyone who replied. There were three groups of responses: 1. Gram-Schmidt type procedure. This usually works, but its numerical properties are not very good. 2. Eigenvalue search (through SVD, or otherwise). 3. The most suitable for me: Householder QR decomposition with follow-up recovery of Q. The routines are available in LAPACK: DGEQPF followed by DOPGTR. This is probably the most efficient and balanced way. The question and responses follow: ***************** Does anyone know the efficient algorithm/subroutine for calculating the basis in the orthogonal subspace? Or, more precisely: Given N M-dimensional vectors, how to find (M-N) M-dimensional orthonormal vectors, each of them being orthogonal to the each of the N original vectors. This seems to be a very universal problem, yet I couldn't find anything in LAPACK/NETLIB/Numerical Recipies. Any help will be very much appreciated _______ Vitaly Rassolov rassolov@chem.nwu.edu Chemistry Department tel. (847) 491-3423 Northwestern University fax (847) 491-7713 From: Zdenko Tomasic To: rassolov@chem.nwu.edu try QR decomposition with pivoting like in LAPACK's dgeqpf Q is what you are looking for ZT From: Vitaly Rassolov To: Zdenko Tomasic Thanks for your reply. Is this (in your opinion) an efficient enough method for the problem (after all, I don't need R)? From: Zdenko Tomasic To: rassolov@chem.nwu.edu yes. assuming you are dealing with dense matrices. R is often not needed, but is inevitable in this decomposition besides it is only triangular. ZT From: Vitaly Rassolov To: Zdenko Tomasic Thanks again! Allow me to bug you one last time. Instead of Q, the LAPACK subroutines return those Householder reflectors, and reconstructing Q seems to be quite time consuming (M products of M*M matricies). Do you know any alternative QR decomposition subroutines that return Q explicitly, or is there a simpler way to recover Q in LAPACK? >From zdenko@harvey.caam.rice.edu Thu Apr 8 12:03 CDT 1999 hoseholder reflectors (H's) are rather special and their application is very efficient. DOPGTR routine should form Q for you out of H's. ZT From: uccatvm To: rassolov@chem.nwu.edu (Vitaly Rassolov) Hi Vitaly, If I understand you correctly, you are looking for an implementation of Schmidt orthogonalization. You might try a web search for Schmidt or Gram-Schmidt. It should also be described in any linear algebra text. Hope this helps, Tanja -- ==================================================================== Tanja van Mourik phone University College London work: +44 (0)171-504-4665 Christopher Ingold Laboratories home: +44 (0)1895-259-312 20 Gordon Street e-mail London WC1H 0AJ work: T.vanMourik@ucl.ac.uk United Kingdom home: tanja@netcomuk.co.uk ==================================================================== From: Jordi Villa To: Vitaly Rassolov Subject: Re: CCL:lin.alg.? (basis set in the orthog. subspace) Dear Vitaly, if I am not wrong you are asking for a Gram-Schmidt ortogonalization procedure. You can use: v_i = N_i [1 - SUM_(k=1)^(i-1) (1 - delta_(ik) ) v_k * v_k^T] * v_i^0 where v_i^0 is the old (non-ortogonalized) vector and v_i is the new (ortogonalized) one. I hope you understand the equation (it is a little bit odd in ASCII) Best regards -- Jordi Villa i Freixa Department of Chemistry University of Southern California Los Angeles, CA, USA, 90089-1062 Tlf: 1-(213)-740 7671 Fax: 1-(213)-740 2701 villa@mutant.usc.edu http://www1.imim.es/~jvilla >From macrae@mslpc1.riken.go.jp Wed Apr 7 21:39 CDT 1999 To: Vitaly Rassolov Hi It looks to me as if you're describing the Gram-Schmidt process for extending an orthogonal basis. (It's covered to some extent in Numerical Recipes.) There are some numerical instabilities in the method, and modifications of it exist (see e.g. Golub and van Loan, Matrix Computations, Johns Hopkins Press, 1989). Hope this helps. Rod --- R. M. Macrae Muon Science Laboratory Institute of Physical and Chemical Research (RIKEN) Wako, Saitama, Japan From: pedersen@quantix.u-strasbg.fr Hey Vitaly, Your problem could in principle be solved by taking M orthogonal test vectors of dimension M (for example the unit vectors:) (1,0,0,0,0,0,....,0) (0,1,0,0,0,0,....,0) (0,0,1,0,0,0,....,0) (0,0,0,1,0,0,....,0) (0,0,0,0,1,0,....,0) . . . (0,0,0,0,0,0,....,1) You take the first of these and orthonormalize it with respect to the N starting vectors. This is then a basis vector in the orthogonal space. You then take the second of your test vectors and orthonormalize it with respect to your N starting vectors and the first basis vector in the orthogonal space. Take the third test vector and orhthnormalize with respect to the N starting vectors and the two basis vectors in the orthogonal space and so on ... As orthonormalization procedure many alternatives are available, look for example for Gram-Schmidt orthonormalization. While you are doing the successive orthonormalizations, you might encounter the problem, that your current test vector doesn't span a dimension in the orthogonal space. This means, if you are using Gram-Schmidt orthogonalization at least, that the current test vector is a linear combination of the N starting vectors and the already found basis vectors in the orthogonal space. As you have M test vectors and only M-N free dimensions, actually this problem will arise N times among your M test vectors. This is why you need M test vectors and not M-N test vectors, to find the M-N basis vectors in the orthogonal space. Every time you encounter the above mentioned problem, you just throw the test vector away and take the next one, then the problem solves itself. If your are using real variables, all dimensions are reasonably populated and the symmetry is not interfearing, you will normally not have to throw any test vectors away during your algorithm, as it will automatically be the last N of your test vectors that are linearly dependant, where you have used only the first M-N test vectors to find the M-N sought-after basis vectors of the orthogonal space. I hope this was of some help. Good look with the work Thomas ************************ Thomas Bligaard Pedersen C.N.R.S. UMR7551 Lab. Chimie Quantique Universit Louis Pasteur Strasbourg FRANCE ************************ To: Vitaly Rassolov From: Darko Babic Subject: Re: CCL:lin.alg.? (basis set in the orthog. subspace) Hi! One thing you could do is to form a matrix from those N given M-dimensional vectors as the density matrix is formed from the vectors representing the occupied molecular orbitals, and then to diagonalize the matrix. You should get N eigenvalues different from (greater than) zero and M-N zero eigenvalues. The eigenvectors corresponding to zero eigenvalues are those you are looking for. Each of them will be ortho- gonal to the first M eigenvectors, but they might be non-orthogonal between themselves (depends which subroutine you use for the diagonali- zation). In a formal way: let V_i, i=1...N, denote N given eigenvectors and let [V_i]_j denote the j-th coordinate of the vector [V_i]_j. Form the matrix P: [P]_ij = SUM{k=1...N} [V_k]_i * [V_k]_j and diagonalize it. The eigenvectors corresponding to non-zero eigenvalues span the N-dimensional subspace containing the starting N M-dimensional vectors, and the remaining eigenvectors span the rest of the space. I'm sure there must be another way, too. With best wishes, Darko Babic Institute "Rudjer Boskovic" HR-10001 Zagreb, P.O.B. 1016 Croatia >From assfeld@host23.lctn.u-nancy.fr Thu Apr 8 10:14 CDT 1999 Subject: Canonical orthogonalization Hello, have a look at paper : X. Assfeld, and J.-L. Rivail, CHem. Phys. Lett. 263 (1996) 100. you will find what you are looking for. C.U. ...Xav Ast. Pr. Xavier Assfeld Xavier.Assfeld@lctn.u-nancy.fr Laboratoire de Chimie theorique (T) 33 3 83 91 21 49 Universite Henri Poincare (F) 33 3 83 91 25 30 F-54506 Nancy BP 239 http://www.lctn.u-nancy.fr From: Konrad Hinsen To: rassolov@chem.nwu.edu CC: chemistry@www.ccl.net Such a set of vectors can be obtained as a "side effect" of singular value decomposition (SVD). I use the SVD routines in LAPACK for this purposes. What you get is 1) A basis for the subspace spanned by the set of input vectors. 2) A basis for the complementary space, which is what you are looking for. -- ------------------------------------------------------------------------------- Konrad Hinsen | E-Mail: hinsen@cnrs-orleans.fr Centre de Biophysique Moleculaire (CNRS) | Tel.: +33-2.38.25.55.69 Rue Charles Sadron | Fax: +33-2.38.63.15.17 45071 Orleans Cedex 2 | Deutsch/Esperanto/English/ France | Nederlands/Francais -------------------------------------------------------------------------------