From Jeffrey.Gosper@brunel.ac.uk Wed Aug 30 04:26:20 1995 Received: from monge.brunel.ac.uk for Jeffrey.Gosper@brunel.ac.uk by www.ccl.net (8.6.10/950822.1) id EAA13065; Wed, 30 Aug 1995 04:21:48 -0400 Received: from chem-pc-18.brunel.ac.uk by monge.brunel.ac.uk with SMTP (PP) id <07531-0@monge.brunel.ac.uk>; Wed, 30 Aug 1995 09:21:15 +0100 Date: Wed, 30 Aug 1995 09:21:07 BST From: Jeffrey J Gosper Reply-To: Jeffrey.Gosper@brunel.ac.uk Subject: SUMMARY: Cartesian to internal conversion To: chemistry@www.ccl.net Message-ID: Priority: Normal MIME-Version: 1.0 Content-Type: TEXT/PLAIN; CHARSET=US-ASCII Thanks to those who responsed to my question regarding algorithms and/or code for cartesian to internal coordinate conversion. They were all very useful. There were a number of references to the use of BABEL and MOPAC so I've only included one of each of these. Here are the responses I received. ******************************************************************** ***************************** the relevant references are : (1) E. Bright Wilson, Jr., J. C. Decius, Paul C. Cross, Molecular Vibrations, McGraw-Hill Book Company, New York 1955 (2) R. L. Hilderbrandt, J. Chem. Phys. 51(4) (1969) 1654-1659 (3) P. Pulay, Mol. Phys. 17(2) (1969) 197-204 (4) Peter Pulay, Geza Fogarasi, Frank Pang, James E. Boggs, J. Am. Chem. Soc. 101(10) (1979) 2550-2560 (5) Geza Fogarasi, Xuefeng Zhou, Patterson W. Taylor, Peter Pulay, J. Am. Chem. Soc 114(21) (1992) 8191-8201 (6) P. Pulay, G. Fogarasi, J. Chem. Phys. 96(4) (1992) 2856-2860 (7) Thomas H. Fischer, Jan Almloef, J. Phys. Chem. 96(24) (1992) 9768-9774 Esp. refs (3), (4), and (6) give a description of the corresponding algorithm. It is iterative because of the nonlinearity of this transformation and it goes as follows : x(i+1) = x(i) + A ( q - q(i) ) (G1) x(i) and x(i+1) are the old and new cartesian coordinates ( so, you need reasonable start cartesians x(0) ! ), q are the given internal coordinates and q(i) are the internal coordinates which correspond to x(i). A is the generalized right inverse of Wilsons B-matrix ( ref. 1 ) : A = m BT ( B m BT )**(-1) (G2) B is Wilsons B-matrix, BT the transpose of B, and m is any ( non-uniqueness ! ) 3N x 3N matrix such, that B m BT is not singular. Usually it is a unit matrix. The dimensions of the matrices involved are : m : 3N x 3N B : M x 3N BT : 3N x M A : 3N x M with M less or equal 3N-6 for nonlinear molecules. N is the number of atoms in the molecule. So, the individual steps are : (1) given some x(i) (2) compute q(i) from the given x(i) ( a trivial task ), compute B ( actually this is just a by-product of the computing of q(i) ! ) (3) make B m BT and invert it (4) compute A and the new cartesians x(i+1) from equations (G1) and (G2) above. Repeat steps (1) to (4) until q-q(i) is acceptable. The ab initio packages TEXAS ( Peter Pulay ), TURBOMOLE ( Reinhart Ahlrichs ), and DISCO ( Jan Almloef ) certainly have a module to do precisely this conversion from internal to cartesian coordinates. See esp. section IV of ref. (5). Because I am working with TURBOMOLE and this is not free for distribution you should probably contact one of the above mentioned authors to get the corresponding modul or even better, write your own code. Good luck Heinz Heinz Schiffer Hoechst AG Scientific Computing 65926 Frankfurt am Main Phone ++49-69-305-2330 Fax ++49-69-305-81162 Email schiffer@msmwia.hoechst.hoechst-ag.d400.de ******************************************************************** ***************************** You can get the MOPAC source code at www.ccl.net and look at the GMETRY.FOR code which has this convertion routines. Best regards, Edgardo Garcia Cristol Chem & Biochem University of Colorado BOULDER CO 80309 USA ******************************************************************** ***************************** Reply to: RE>CCL:Cartesian to Internal Conversion such algorithms are often part of standard MD programs (etc) which have the potential in internal coords but the hard work (integration, etc) in cartesian. for example try program VENUS, coauthored by Bill Hase and others (including myself). Bill Hase is at Wayne State Univ hase@sun.chem.wayne.edu VENUS has scheduled to be submitted to Quantum Chem Program Exchange this month. Kieran Lim xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx x x x Kieran F Lim ph: + [61] (3) 9344 7090 (office) x x School of Chemistry + [61] (3) 9344 6481 (School) x x University of Melbourne + [61] (3) 9435 0259 (home) x x Parkville _.-_|\ x x VIC 3052 / \ fax: + [61] (3) 9347 5180 (preferred) x x AUSTRALIA \_.--.o/ + [61] (3) 9344 6233 x x v x xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx x x x kieran_lim.chemistry@muwayf.unimelb.edu.au x x x xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx x x x Please note: All Melbourne phone and fax numbers have changed x x Numbers xxx-xxxx have become 9xxx-xxxx. x x eg: + [61] (3) 347 5180 becomes + [61] (3) 9347 5180 x x (03) 435 0259 becomes (03) 9435 0259 x x x xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx ******************************************************************** ***************************** Hello; Here is how to get Cart's --> Internal with Mopac. aigout=ab in. geom. output; noxyz =don't output any Cartesians Mopac will also go from internal--> Cartesians --------------- INPUT: am1 0scf aigout noxyz (CN)2 CARBENE, MP2 opt. Input: SPARTAN 6-31G* geom C -0.25275 1 -0.33068 1 -0.68783 1 C -0.60299 1 -0.53886 1 0.71870 1 C -0.70363 1 0.55964 1 1.55581 1 C 1.19821 1 -0.46268 1 -0.94150 1 O -1.05071 1 -0.16199 1 -1.54391 1 N -0.80356 1 1.41686 1 2.31356 1 N 2.31853 1 -0.54893 1 -1.09341 1 ------------------ OUTPUT: ******************************************************************** *********** ** MOPAC 93 (c) Fujitsu ** ******************************************************************** *********** AM1 CALCULATION RESULTS ******************************************************************** *********** * MOPAC 93.01 CALC'D. Wed Aug 23 10:58:47 1995 * AIGOUT - IN ARC FILE, INCLUDE AB-INITIO GEOMETRY * T= - A TIME OF 3600.000 SECONDS REQUESTED * DUMP=N - RESTART FILE WRITTEN EVERY 3600.000 SECONDS * AM1 - THE AM1 HAMILTONIAN TO BE USED * NOXYZ - CARTESIAN COORDINATES NOT TO BE PRINTED * 0SCF - AFTER READING AND PRINTING DATA, STOP ******************************************************************** ***060BY060 AM1 0SCF AIGOUT NOXYZ (CN)2 CARBENE, MP2 opt. Input: SPARTAN 6-31G* geom ATOM CHEMICAL BOND LENGTH BOND ANGLE TWIST ANGLE NUMBER SYMBOL (ANGSTROMS) (DEGREES) (DEGREES) (I) NA:I NB:NA:I NC:NB:NA:I NA NB NC 1 C 2 C 1.46435 * 1 3 C 1.38477 * 119.02875 * 2 1 4 C 1.47887 * 112.75119 * 94.13493 * 1 2 3 5 O 1.18240 * 123.66246 * -90.51055 * 1 2 3 6 N 1.14848 * 175.76608 * -177.86070 * 3 2 1 7 N 1.13386 * 177.68500 * -30.11819 * 4 1 2 C: (AM1): M.J.S. DEWAR ET AL, J. AM. CHEM. SOC. 107 3902-3909 (1985) N: (AM1): M.J.S. DEWAR ET AL, J. AM. CHEM. SOC. 107 3902-3909 (1985) O: (AM1): M.J.S. DEWAR ET AL, J. AM. CHEM. SOC. 107 3902-3909 (1985) MOLECULAR POINT GROUP : C1 RHF CALCULATION, NO. OF DOUBLY OCCUPIED LEVELS = 16 INTERATOMIC DISTANCES 0 C 1 C 2 C 3 C 4 O 5 N 6 -------------------------------------------------------------------- ---------- C 1 .000000 C 2 1.464354 .000000 C 3 2.455582 1.384769 .000000 C 4 1.478870 2.450794 3.301317 .000000 O 5 1.182400 2.337068 3.201481 2.347542 .000000 N 6 3.516480 2.531531 1.148476 4.258541 4.175395 .000000 N 7 2.612204 3.437895 4.169022 1.133857 3.421177 5.021874 0 N 7 ------------------ N 7 .000000 GEOMETRY IN MOPAC Z-MATRIX FORMAT AM1 0SCF AIGOUT NOXYZ (CN)2 CARBENE, MP2 opt. Input: SPARTAN 6-31G* geom C .00000000 0 .0000000 0 .0000000 0 0 0 0 4.0000 C 1.46435433 1 .0000000 0 .0000000 0 1 0 0 4.0000 C 1.38476850 1 119.0287496 1 .0000000 0 2 1 0 4.0000 C 1.47887031 1 112.7511903 1 94.1349333 1 1 2 3 4.0000 O 1.18239987 1 123.6624556 1 -90.5105473 1 1 2 3 6.0000 N 1.14847603 1 175.7660839 1 -177.8607024 1 3 2 1 5.0000 N 1.13385740 1 177.6850027 1 -30.1181886 1 4 1 2 5.0000 GEOMETRY IN GAUSSIAN Z-MATRIX FORMAT AM1 0SCF AIGOUT NOXYZ (CN)2 CARBENE, MP2 opt. Input: SPARTAN 6-31G* geom C C 1 r21 C 2 r32 1 a321 C 1 r41 2 a412 3 d4123 0 O 1 r51 2 a512 3 d5123 0 N 3 r63 2 a632 1 d6321 0 N 4 r74 1 a741 2 d7412 0 r21 1.464354 r32 1.384769 r41 1.478870 r51 1.182400 r63 1.148476 r74 1.133857 a321 119.028750 a412 112.751190 a512 123.662456 a632 175.766084 a741 177.685003 d4123 94.134933 d5123 269.489453 d6321 182.139298 d7412 329.881811 TOTAL CPU TIME: .03 SECONDS == MOPAC DONE == JOB FINISHED ******************************************************************** ************* The Gaussian utility newzmat can do the job. In general, I am rarely satisfied with automatically generated Z-matrices, whether from Gaussian newzmat, MOPAC or other software. At best they give a fair starting point. At worst you end up doing the Z-matrix by hand anyway, especially if you want to impose specific symmetric constraints or a define particular reaction coordinate. Alternative methods such Spartan's Cartessian coordinate optimization which supports "internal" constraints and Gaussian 94 use of redundant internal coordinates are overall more convenient and, in my opinion, superior. Regards, Karl ____________________________________________________________________ ___ / \ | Comments are those of the author and not Unilever Research U. S. | | | | Karl F. Moschner, Ph. D. | | | | Unilever Research U. S. e-mail: Karl.F.Moschner@urlus.sprint.com | | 45 River Road Phone: (201) 943-7100 x2629 | | Edgewater, NJ 07020 FAX: (201) 943-5653 | \___________________________________________________________________ ____/ ******************************************************************** Pat Walters' group at Tucson offers Babel, which goes "both ways," and runs on dos and unix. The address is babel@mercury.aichem.arizona.edu I believe it ftp's from there. John Reissner Pembroke State University Pembroke NC 28372 USA reissner@pembvax1.pembroke.edu vox: (910)521-6425 fax: (910)521-6649 ******************************************************************** ********************** Dear Jeff, on April 27 I posted the same question the other way round, i.e. internal to cartesian. I can send you my summary again if you wish. For your problem, you should first take a look at BABEL which is a nice tool for conversion between several chemical file formats. I think you can get the BABEL package from ftp, maybe from the Ohio chemistry software server In general I think the this conversion isn't quite hard. Bonds are computed as Euclidean distances. Angles by the equation (x0-x1) o (x2-x1) cos phi = ------------------- |x0-x1| |x2-x1| if x1 is the top of the angle and "o" the standard scalar product. Torsion angles are defined as the angle between the planes made up by (x0,x1,x2) on one hand and (x1,x2,x3) on the other. So compute these planes and their normal vectors and then calculate the angle between them by the equation above. Ask me if you need more help. By the way, I think I also read a solution to this problem in the following book : Tim Clark: A handbook of computational chemistry, Wiley, New York 1985, 4. ed. Best wishes, Thomas Wieland +---------------+ Dipl. Math. |+---- +----+| Lehrstuhl II f. Mathematik |\ \ | || Universitaet Bayreuth | \ \ | || | \ \ | || 95440 Bayreuth | \ \\ || Germany | \ \\ || Tel. +49 (921) 553386 | \ \\ || Fax +49 (921) 553385 | \-------|| +---------------+ P.S.: Take a look at MOLGEN: http://btm2xd.mat.uni-bayreuth.de/molgen ******************************************************************** *** In reponce to your request I have written a TCL code for this purpose. I am glad to sent it to you - but it is still a little developmental - in that I have not checked the patch I put on the four quadrant correector yet. A more elegant solution might be a developmental program called chemsh. This is being authored by Paul Sherwood (et all) at Daresbury. I cannot say if he would be willing to release a version to you, but if you were interested in testing it - it does have just about any z-matrix conversion you could want. Try psh@dl.ac.uk I hope this helps All the best Alex +--------------------------------------------------+ |Alternate E-mail A.J.Turner@Bath.ac.uk | |www home @ http://www.bath.ac.uk/~chpajt/home.html| +--------------------------------------------------+ ******************************************************************** ********************** provided, the cartesian co-ordinates of atoms are given in vectors XAT,YAT,ZAT then you can use the following routines (written by me using algorithms described by many people, so I do not remember). Good luck, Frank Eisenmenger +------------------------------------------------------------------- -----+ | Institute of Biochemistry | | Medical Faculty of the Humboldt-University (Charite) | | | | Hessische Str. 3-4 | | 10115 Berlin, Germany | | | | Phone: +49-30-28468-612 or -239 FAX: +49-30-28468-600 | | E-mail: eisenmen@orion.rz.mdc-berlin.de | +------------------------------------------------------------------- -----+ c ******************************* real function bndlen(i1,i2) c .................................................. c PURPOSE: return bond length between atoms i1 & i2 c c INPUT: i1,i2 - indices of 2 atoms c c .................................................. COMMON /COOR/ XAT(1000),YAT(1000),ZAT(1000) bndlen=sqrt( (xat(i1)-xat(i2))**2 # +(yat(i1)-yat(i2))**2+(zat(i1)-zat(i2))**2 ) return end c ********************************** real function valang(i1,i2,i3) c ......................................... c PURPOSE: return valence angle (i1,i2,i3) c [in rad.] with 'i2' as vertex c c INPUT: i1,i2,i3 - indices of 3 atoms c ......................................... COMMON /COOR/ XAT(1000),YAT(1000),ZAT(1000) h1=xat(i2) h2=yat(i2) h3=zat(i2) x1=xat(i1)-h1 x2=yat(i1)-h2 x3=zat(i1)-h3 y1=xat(i3)-h1 y2=yat(i3)-h2 y3=zat(i3)-h3 x=x1*x1+x2*x2+x3*x3 y=y1*y1+y2*y2+y3*y3 u=x*y if (u.ne.0.0) then a=(x1*y1+x2*y2+x3*y3)/sqrt(u) a=max(a,-1.0) a=min(a,1.0) valang=acos(a) return else write (*,'(a,3i5)')' valang> Error in coordinates of atoms #: ' # ,i1,i2,i3 stop endif end c ************************************* real function dihedr(i1,i2,i3,i4) c ............................................. c PURPOSE: return dihedral angle (i1,i2,i3,i4) c [in rad.] c c INPUT: i1,i2,i3,i4 - indices of four atoms c c CALLS: none c ............................................. COMMON /COOR/ XAT(1000),YAT(1000),ZAT(1000) x1=xat(i2)-xat(i1) y1=yat(i2)-yat(i1) z1=zat(i2)-zat(i1) x2=xat(i3)-xat(i2) y2=yat(i3)-yat(i2) z2=zat(i3)-zat(i2) ux1=y1*z2-z1*y2 uy1=z1*x2-x1*z2 uz1=x1*y2-y1*x2 x1=xat(i4)-xat(i3) y1=yat(i4)-yat(i3) z1=zat(i4)-zat(i3) ux2=z1*y2-y1*z2 uy2=x1*z2-z1*x2 uz2=y1*x2-x1*y2 u1=ux1*ux1+uy1*uy1+uz1*uz1 u2=ux2*ux2+uy2*uy2+uz2*uz2 u=u1*u2 if (u.ne.0.0) then a=(ux1*ux2+uy1*uy2+uz1*uz2)/sqrt(u) a=max(a,-1.0) a=min(a,1.0) dihedr=acos(a) if (ux1*(uy2*z2-uz2*y2)+uy1*(uz2*x2-ux2*z2)+ # uz1*(ux2*y2-uy2*x2).lt.0.0) dihedr =-dihedr return else write (*,'(a,4i5)')' dihedr> Error in coordinates of atoms #: ' # ,i1,i2,i3,i4 stop endif end ******************************************************************** Dear Jeff, Many years ago when I was young and writing my programs myself I wrote such a programm which converts cartesian coordinates into internal ones. It worked in that years so I hope it will work now... If you will have problems with it - ask me. Cheers, Vlad c c subroutine fmintc(cartes,cint,int,natoms) c c Generate internal coordinates from cartesian c INPUT: c cartes(3,*) - initial cartesian coordinates c int(3,*) - connectivity of a given atom with others (bond length -int(1,*), c valence angle -int(2,i), torsion- int(3,*) c natoms - number of atoms c c OUTPUT: c cint(3,*) - Z-matrix c implicit real*8 (a-h,o-z) dimension cint(3,*),int(3,*),cartes(3,*),ext(3,4),v1(3),v2(3) data rad_to_degree/57.2958/ c do i=1,3 do j=1,3 cint(j,i) = 0.0 enddo enddo c cint(1,2) = sqrt( (cartes(1,1)-cartes(1,2))**2 + & (cartes(2,1)-cartes(2,2))**2 + & (cartes(3,1)-cartes(3,2))**2 ) cint(1,3) = sqrt( (cartes(1,3)-cartes(1,2))**2 + & (cartes(2,3)-cartes(2,2))**2 + & (cartes(3,3)-cartes(3,2))**2 ) do 2 i=1,3 v1(i) = cartes(i,1) - cartes(i,2) 2 v2(i) = cartes(i,3) - cartes(i,2) cint(2,3) = acos( (v1(1)*v2(1) + v1(2)*v2(2) + v1(3)*v2(3))/ & (cint(1,2)*cint(1,3)) ) * rad_to_degree c do 1 i=4,natoms do 3 j=1,3 ext(j,4) = cartes(j,i) kk = 4 - j do 3 k=1,3 3 ext(k,j) = cartes(k,int(kk,i)) 1 call intrnl(ext,cint(1,i) ) c return end c c subroutine intrnl(ext,cint ) c implicit real*8 (a-h,o-z) dimension ext(3,*),cint(*),h1(3),h2(3),vnorm(3),unitar(3,3), & vmidl(3,2) data pi/3.14159265453/,pp/57.2958/ c do i=1,3 h1(i)=ext(i,1)-ext(i,2) h2(i)=ext(i,3)-ext(i,2) enddo c call vecmul(h2,h1,vnorm) call rulcos(vnorm,unitar(1,3) ) call rulcos(h2,unitar(1,1) ) call vecmul(unitar(1,3),unitar(1,1),unitar(1,2) ) c do 2 i=1,2 do 2 j=1,3 2 vmidl(j,i) = ext(1,i+2) * unitar(1,j) + & ext(2,i+2) * unitar(2,j) + & ext(3,i+2) * unitar(3,j) c cint(1) = sqrt( (ext(1,4)-ext(1,3))**2 + & (ext(2,4)-ext(2,3))**2 + & (ext(3,4)-ext(3,3))**2 ) cint(2) = acos( (vmidl(1,1)-vmidl(1,2))/cint(1) ) * pp subd=sqrt( (vmidl(2,2)-vmidl(2,1))**2 + & (vmidl(3,2)-vmidl(3,1))**2 ) arg = (vmidl(2,2)-vmidl(2,1))/subd if ( abs(arg).gt. 1.D00 ) arg = sign( 0.9999999D00, arg ) CC cint(3) = -acos( arg ) * pp cint(3) = acos( arg ) * pp if ( vmidl(3,2)-vmidl(3,1).lt.0.0 ) cint(3)= -cint(3) c return end c c subroutine rulcos( ventry,rcos ) c implicit real*8 (a-h,o-z) dimension ventry(*),rcos(*) c s=0.0 do 1 i=1,3 1 s=s+ventry(i)**2 s=sqrt(s) do 2 i=1,3 2 rcos(i)=ventry(i)/s c return end c c subroutine vecmul(v1,v2,vnorm) c implicit real*8 (a-h,o-z) dimension v1(*),v2(*),vnorm(*) c vnorm(1) = v1(2)*v2(3) - v1(3)*v2(2) vnorm(2) = v1(3)*v2(1) - v1(1)*v2(3) vnorm(3) = v1(1)*v2(2) - v1(2)*v2(1) c return end Received: from n32.rhea.cnusc.fr by monge.brunel.ac.uk with SMTP (PP) id <20604-0@monge.brunel.ac.uk>; Wed, 23 Aug 1995 12:54:21 +0100 Received: by n32.rhea.cnusc.fr (AIX 3.2/UCB 5.64/4.03) id AA30959; Wed, 23 Aug 1995 13:51:50 +0200 Date: Wed, 23 Aug 1995 13:51:50 +0200 From: lcaogto@n32.rhea.cnusc.fr (\(SP_6\)) Message-Id: <9508231151.AA30959@n32.rhea.cnusc.fr> ******************************************************************** *********** /\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\ Dr. Jeff Gosper Dept. of Chemistry BRUNEL University Uxbridge Middx UB8 3PH, UK voice: 01895 274000 x2187 facsim: 01895 256844 internet/email/work: Jeffrey.Gosper@brunel.ac.uk internet/WWW: http://http2.brunel.ac.uk:8080/~castjjg \/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/ From JAQ@XRAY.BMC.UU.SE Wed Aug 30 05:26:20 1995 Received: from XRAY.BMC.UU.SE for JAQ@XRAY.BMC.UU.SE by www.ccl.net (8.6.10/950822.1) id FAA13679; Wed, 30 Aug 1995 05:15:11 -0400 Date: Wed, 30 Aug 1995 11:07:01 +0300 (MET-DST) From: JAQ@XRAY.BMC.UU.SE Message-Id: <950830110701.2340053f@XRAY.BMC.UU.SE> Subject: a reaction rate To: chemistry@www.ccl.net X-Vmsmail-To: SMTP%"chemistry@www.ccl.net" Hi there, Does anyone know whether the rate constant for the hydrogen exchange reaction H2O + OH- = OH- + H2O (aq) is known experimentally ? What about gas-phase experiments and calculations ? Johan Aqvist Molecular Biology, Uppsala University, Sweden aqvist@xray.bmc.uu.se From toukie@zui.unizh.ch Wed Aug 30 07:26:33 1995 Received: from rzusuntk.unizh.ch for toukie@zui.unizh.ch by www.ccl.net (8.6.10/950822.1) id HAA15306; Wed, 30 Aug 1995 07:20:36 -0400 Received: by rzusuntk.unizh.ch (4.1/SMI-4.1.9) id AA00580; Wed, 30 Aug 95 13:20:33 +0200 X-Nupop-Charset: Swiss Date: Wed, 30 Aug 1995 13:20:39 +0100 (MET) From: "Hr Dr. S. Shapiro" Sender: toukie@zui.unizh.ch Reply-To: toukie@zui.unizh.ch Message-Id: <48039.toukie@zui.unizh.ch> To: chemistry@www.ccl.net Subject: Dielectric constant for n-octanol Dear Colleagues; For some calculations which I should like to perform, I need to have the value of the dielectric constant for n-octanol. To date I have found two rather different numbers in the literature: (i) according to the Landolt-Boernstein Physikalische-chemische Tabelen, V/2, p. 1036 (1923), the dielectric constant of n- octanol at 18oC is 3.4 [ref.: W. Rudolph, Diss., Leipzig, 1911]. (ii) according to Lange's Handbook of Chemistry, 12th Edn., p. 10-113 (1979), the dielectric constant of n-octanol at 20oC is 10.34 [ref.: NBS Circular from the 1950's]. Does anyone know of any more _recent_ experimental determinations of the value of the dielectric constant for n-octanol? If so, kindly advise me of the relevent references. By the way, if anyone out there has access to or is otherwise familiar with the doctoral dissertation of W. Rudolph, completed at Leipzig in 1911, I'd be very interested in learning more about this work. Thanks in advance to all responders. Sincerely, (Dr.) S. Shapiro Inst. f. orale Mkrobiol. u. allg. Immunol. Zent. f. Zahn-, Mund- u. Kieferheilkd. der Univ. ZH Plattenstr. 11 Postfach CH-8028 Zuerich 7 Switzerland Internet: toukie@zui.unizh.ch FAX-nr: ( ... + 1) 261'56'83 From sschulz@chemie.fu-berlin.de Wed Aug 30 08:41:24 1995 Received: from weasel.chemie.fu-berlin.de for sschulz@chemie.fu-berlin.de by www.ccl.net (8.6.10/950822.1) id IAA16575; Wed, 30 Aug 1995 08:27:50 -0400 Message-Id: <199508301227.IAA16575@www.ccl.net> Received: by weasel.chemie.fu-berlin.de (1.37.109.16/16.2) id AA066715651; Wed, 30 Aug 1995 14:27:31 +0200 From: Stefan Schulz Subject: ECP in GAMESS-US & GAUSSIAN92 To: chemistry@www.ccl.net Date: Wed, 30 Aug 95 14:27:31 MESZ X-Hpvue$Revision: 1.8 $ Mime-Version: 1.0 Content-Type: Message/rfc822 X-Vue-Mime-Level: 4 Mailer: Elm [revision: 70.85] Hi CCLers, a few weeks back I posted a question concerning the use of effective core potentials when migrating between the ab initio programs GAMESS-US and GAUSSIAN92. I got quite a a lot of responses and would like to thank all those poeple who took the time to help me. Here is a summary of the answers I got: ***************************************************************** In the GAMESS input : $CONTRL EXETYP=CHECK FRIEND=GAUSSIAN $END Will give you an example gaussian input file in $JOB.dat, which (with a little modifications) will run perfectly on gaussian9x. Be sure to check the d5 <-> d6 thing. Check if they are d5 (or d6) for both programs, otherwise specify an extra keyword. ***************************************************************** Hi there I have been using the 2 programs quite a bit lately and yes there is a simple command which you can use in GamessUS to prepare input for Gausian92. In the $CONTRL group if you include the keyword FRIEND=GAUSSIAN, then your run will be turned into a CHECK run (ie EXETYP=CHECK), and the input for G92 will be put in your PUN file for extraction. This should give you the correct input structure for the ECPs in G92. ********************************************************************** I've got the same problems just few days ago. I did not use d functions and the results in GAMESS-UK were more stable energetically than in GUASSIAN92. I'd appreciate a forward of any message that you get about this topic. ********************************************************************** Hi Stefan, hast Du dir die explizite formel von den ECP angeschaut? Ich weis nichtmehr wie es in den einzelnen Programmen Gaussian und Gamess ist aber ich errinere mich an aehnliche probleme die ich vor ein paar jahren hatte. Ich denke der unteschied ist in dem Jacobi term vom ECP r^2, welcher in einem der programe explizit eingegeben werden mus nd in dem anderen nicht. Also schau ins Handbuch und eventuel muss du dan eine 2 won den n(i) Werten abziehen. ********************************************************************** Thanks again to all those who answered. A few days after I posted my question to the CCL I became aware of the program package MOLPRO94, which as distinct from other commonly used quantum chemistry packages specializes on highly accurate computations for small molecules, with extensive treatment of the electron correlation problem. Since I am interested in calculating ab initio potential surfaces for electronically excited states of medium sized molecules e.g. Na (NH3)_n n=1,2,... I found the program package MOLPRO94 to be superior in this respect to the programs GAMESS-US and GAUSSIAN92. With this statement I really dont want to start any fundamental discussion on which program performs better than the other, but rather use this opportunity to bring MOLPRO94 to the attention of all those who might benefit from MOLPRO94. Unfortunately MOLPRO94 is NOT FREE (No doubt that this is a huge advantage of GAMESS-US). Information on MOLPRO94 is available at http://tcibm.mols.sussex.ac.uk/molpro/molpro.html Regards Stefan =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= | | Stefan Schulz | | sschulz@chemie.fu-berlin.de | FU Berlin - Theoretical Chemistry | | Tel. ++49/30/838 5384 (2351) | Takustrasse 3 | | FAX. ++49/30/838 4792 | D-14195 Berlin | =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From hinsenk@ERE.UMontreal.CA Wed Aug 30 09:11:25 1995 Received: from condor.CC.UMontreal.CA for hinsenk@ERE.UMontreal.CA by www.ccl.net (8.6.10/950822.1) id JAA17077; Wed, 30 Aug 1995 09:01:45 -0400 Received: from eole.ERE.UMontreal.CA (eole.ERE.UMontreal.CA [132.204.10.20]) by condor.CC.UMontreal.CA with ESMTP id JAA09619 (8.6.11/IDA-1.6); Wed, 30 Aug 1995 09:00:08 -0400 Received: from cyclone.ERE.UMontreal.CA by eole.ERE.UMontreal.CA (950221.405.SGI.8.6.10/5.17) id JAA23563; Wed, 30 Aug 1995 09:00:06 -0400 Received: by cyclone.ERE.UMontreal.CA (950221.405.SGI.8.6.10/5.17) id JAA09080; Wed, 30 Aug 1995 09:00:04 -0400 Date: Wed, 30 Aug 1995 09:00:04 -0400 From: hinsenk@ERE.UMontreal.CA (Hinsen Konrad) Message-Id: <199508301300.JAA09080@cyclone.ERE.UMontreal.CA> To: eknight@wppost.depaul.edu CC: chemistry@www.ccl.net In-reply-to: (eknight@wppost.depaul.edu) Subject: Re: CCL:Parity checking RAM in PCs The bottom line is this: if you run a calculation on one of the new PCs, some number of bits _might_ have been randomly flipped. My question is this: Is it legitimate to report calculations done on such a computer in a scientific journal? It would be nice to have more accurate data on the probability of memory errors, but assuming that the numbers you quote are of the right order of magnitute, I'd probably be willing to take that risk, and also the other hardware-related stochastic risks, like errors in disk access. In most situations I can think of, several calculations of rather similar systems would be done and a large error due to a stochastic source would be visible (in contrast to errors caused by systematic hardware failures such as the infamous Pentium bug). However, one should consider making a statement about hardare (and software!) used for published work oligatory. It would only add one sentence to each article and at least give readers a chance to spot dubious results due to hardware/software problems, even if they are discovered only much later. ------------------------------------------------------------------------------- Konrad Hinsen | E-Mail: hinsenk@ere.umontreal.ca Departement de chimie | Tel.: +1-514-343-6111 ext. 3953 Universite de Montreal | Fax: +1-514-343-7586 C.P. 6128, succ. A | Deutsch/Esperanto/English/Nederlands/ Montreal (QC) H3C 3J7 | Francais (phase experimentale) ------------------------------------------------------------------------------- From ferenc@rchsg8.chemie.uni-regensburg.de Wed Aug 30 12:41:27 1995 Received: from comsun.rz.uni-regensburg.de for ferenc@rchsg8.chemie.uni-regensburg.de by www.ccl.net (8.6.10/950822.1) id MAA21498; Wed, 30 Aug 1995 12:25:34 -0400 Received: from rchsg8.chemie.uni-regensburg.de by comsun.rz.uni-regensburg.de with SMTP id AA00770 (5.65c/IDA-1.4.4 for ); Wed, 30 Aug 1995 18:25:07 +0200 Received: by rchsg8.chemie.uni-regensburg.de (931110.SGI/URRZ-Sub-1.5-irix) (for CHEMISTRY@www.ccl.net) id AA29954; Wed, 30 Aug 95 18:23:21 +0200 Date: Wed, 30 Aug 95 18:23:21 +0200 From: Ferenc.Molnar@chemie.uni-regensburg.de (Ferenc Molnar) Message-Id: <9508301623.AA29954@rchsg8.chemie.uni-regensburg.de> To: CHEMISTRY@www.ccl.net Subject: normal vibrations Dear Netters: I have a (maybe trivial) question concerning the calculation of normal modes in MOPAC. The diagonalization of the force matrix (in mass weighted coordinates) leaves you with a set of eigenvalues and normal coordinates. The eigenvalue number k correspond to (w_k)^2, where the w_ks are the vibrational frequencies, this means: (w_k)^2=K_k/M_k K_k: force constant of the kth normal mode M_k: reduced mass of the k_th normal mode Now my question is, if (w_k)^2 determines only the ratio of K_k and M_k, then how are the reduced masses, reported in the "vibrational analysis" section of the MOPAC output file, calculated? Is there a convention, which "fraction" of (w_k)^2 to use for K_k and which for M_k? Thank you very much in advance! Best regards, Ferenc Ferenc Molnar --------------------------------------------------------------------------- Institut fuer Physikalische und Theoretische Chemie - Lehrstuhl Prof. Dick - Tel.: (+49) 941 943-4466 /-4486 Universitaet Regensburg Fax.: (+49) 941 943-4488 Universitaetsstrasse 31 D-93053 Regensburg Deutschland / Germany --------------------------------------------------------------------------- EMail (SMTP): Ferenc.Molnar@chemie.uni-regensburg.de --------------------------------------------------------------------------- :-) ...THE PHYSICISTS HAVE MADE THEIR UNIVERSE, AND IF YOU DO NOT LIKE IT, YOU MUST MAKE YOUR OWN. -- E. R. HARRISON IN "COSMOLOGY" (1980) --------------------------------------------------------------------------- From markz@msi.com Wed Aug 30 13:41:27 1995 Received: from msi.com for markz@msi.com by www.ccl.net (8.6.10/950822.1) id NAA22648; Wed, 30 Aug 1995 13:32:10 -0400 Received: from gomez (gomez.msi.com) by msi.com (4.1/SMI-4.1) id AA16111; Wed, 30 Aug 95 13:27:32 EDT Received: by gomez (920330.SGI/911001.SGI) for CHEMISTRY@www.ccl.net id AA01399; Wed, 30 Aug 95 13:30:54 -0400 Date: Wed, 30 Aug 1995 13:30:54 -0400 (EDT) From: "Mark A. Zottola" Sender: "Mark A. Zottola" Reply-To: "Mark A. Zottola" Subject: Parity Checks Anon To: CHEMISTRY@www.ccl.net Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; CHARSET=US-ASCII This talk of parity checking seems specious. Assuming the error rates in PCs to be 1 bit a year (which is significantly higher than anyone has estimated on this thread), there are such *minimal* effects on compuational efforts as to make this a non issue. Unless there are people on this list who are involved in high level cryptography or launching thermonuclear devices while using PCs, there is nothing to worry about. IMHO. Mark Zottola From jkl@ccl.net Wed Aug 30 14:11:27 1995 Received: from bedrock.ccl.net for jkl@ccl.net by www.ccl.net (8.6.10/950822.1) id NAA23650; Wed, 30 Aug 1995 13:58:23 -0400 Received: from theory.tc.cornell.edu for jeanne@tc.cornell.edu by bedrock.ccl.net (8.6.10/950822.1) id NAA14383; Wed, 30 Aug 1995 13:58:21 -0400 Received: from [128.84.181.75] (JEANNE.TC.CORNELL.EDU [128.84.181.75]) by theory.tc.cornell.edu (8.6.9/8.6.6) with SMTP id NAA135879; Wed, 30 Aug 1995 13:51:57 -0400 Date: Wed, 30 Aug 1995 13:51:57 -0400 Message-Id: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" To: jeanne@TC.Cornell.EDU From: jeanne@TC.Cornell.EDU (Jeanne C. Butler) Subject: CTC Symposium: Protein Structure and Folding (free/open to public) Cc: donna@TC.Cornell.EDU CORNELL THEORY CENTER SYMPOSIUM "PROTEIN STRUCTURE AND FOLDING" OCTOBER 10 -11, 1995 700 CLARK HALL CORNELL UNIVERSITY CAMPUS For detailed announcement and/or more information: donna@tc.cornell.edu -or- http://www.tc.cornell.edu/Events/symp.html From elewars@alchemy.chem.utoronto.ca Wed Aug 30 14:11:30 1995 Received: from alchemy.chem.utoronto.ca for elewars@alchemy.chem.utoronto.ca by www.ccl.net (8.6.10/950822.1) id NAA23576; Wed, 30 Aug 1995 13:57:01 -0400 Received: (from elewars@localhost) by alchemy.chem.utoronto.ca (8.6.10/8.6.10) id NAA19025 for chemistry@www.ccl.net; Wed, 30 Aug 1995 13:57:00 -0400 Date: Wed, 30 Aug 1995 13:57:00 -0400 From: "E. Lewars" Message-Id: <199508301757.NAA19025@alchemy.chem.utoronto.ca> To: chemistry@www.ccl.net Subject: VIBRATIONAL FREQS Ferenc Molner asked a question about the calculation of vibrational freqs and the role of the reduced masses. It's my observation too that this isn't clearly explained in the usual handy refs--they gloss over it by saying that the freqs are the mass-weighted eigenvalues of the hessian matrix. I think several people out there would appreciate it if someone, in answering Ferenc's question, showed in some detail how an ab in calc on a small molecule like, say, H-O-Cl leads to the freqs, showing the actual matrices as the rectangular arrays they are. Thanks E. Lewars ===== From masayuki@hsc.usc.edu Wed Aug 30 14:16:33 1995 Received: from hsc.usc.edu for masayuki@hsc.usc.edu by www.ccl.net (8.6.10/950822.1) id OAA23730; Wed, 30 Aug 1995 14:00:31 -0400 Received: (masayuki@localhost) by hsc.usc.edu (8.6.12/8.6.4) id LAA20107 for chemistry@www.ccl.net; Wed, 30 Aug 1995 11:00:30 -0700 Date: Wed, 30 Aug 95 11:00:29 PDT From: Masayuki Yuki To: chemistry@www.ccl.net Subject: Summary: DNA model building software Message-ID: Dear CCL Subscribers: I would like to thank everyone who responded to my question regarding the availability of DNA model building software. I learned that there are many commercial/academic softwares that I can either purchase of obtain through ftp. Here are the responses that I received so far: Date: Thu, 24 Aug 95 16:15:45 PDT From: macke@scripps.edu (Tom Macke) I have an experimental code that has been used to build all sorts of helices, hairpins, pknots, etc. It is structure decription language that allows you to build these molecules by a combination of rigid body transforms, distance geometry and if you have an AMBER license, molecular mechanics. Version 1.0 is on our anon ftp: ftp scripps.edu login as anon, etc cd /pub/macke binary get nab.XXXXX.tar.Z Version 1.1 with several functional improvements will be posted to the ftp on Monday 28 Aug. The new documentation approaches useable. The code called nab, compiles structure description programs into C which the user executes to create the structures. Runs on sun4SunOS4, sun4SunOS5, IRIX5, IRIX6, hp-ux, titan3000 and Convex. Will not run on MS boxes. Requires a C compiler (GNU-C will do). nab can generate AVS modules allowing real time (depending on size) interactive structure creation, where the paramters of the model are attached to AVS dials and buttons. Open Inventor Version planned, but must graduate first. If this seems interesting, let me know. tom Date: Fri, 25 Aug 1995 10:32:24 +0100 (BST) From: James Crabbe To: Masayuki Yuki The Windows version of DeskTop Molecular Molecular Modeller will fit the bill - cost about $250. It has been well reviewed (e.g. JACS), and is available from Oxford University Press - see the back cover of J. Mol. Graphics latest version. Contact Mary Levine, OUP, Madison Ave., New York. email mhl@oup-usa.org, fax : 212-726-6450. James Crabbe From: Andreas Goeller Date: Fri, 25 Aug 95 12:42:03 METDST Just have a look at the home page (see below). There is the access to the Oxford Molecular home page. Bye, AHG... --------------------------------------------------------------- Andreas Goeller Computer Chemie Centrum der Dipl.-Chem. Universitaet Erlangen/Nuernberg Naegelsbachstr. 25 phone: +49(0)9131-856583 D-91052 Erlangen fax: +49(0)9131-856565 Germany email: goeller@organik.uni-erlangen.de http://www.organik.uni-erlangen.de --------------------------------------------------------------- Date: Fri, 25 Aug 95 17:03:12 -0400 From: polowin@hyper.hyper.com (Joel Polowin) As long as you can specify the torsional angle values, you can build it with HyperChem. The standard nucleotide database has the usual residues; it's easy to add your own. Standard conformations are A, B, and Z; or you can specify "other" and fill in the required angles. Single- and double- strand are standard options; for other strandedness you'd have to build the extra strand(s) and move them into place. DNA and RNA are standard in the builder. The software runs under Windows 3.1 and our tests so far under beta-test Win'95 have beem positive; we'll have a version for SGI IRIX 5.3 (if you're thinking of upgrading) by the middle of September. I can send you more information and/or demo disks if you want; info and demos are also available from our WWW server, http://www.hyper.com . Regards, Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" to hyperchem-request@hyper.com WWW: http://www.hyper.com/ From: amm@giris18.pfizer.com (Alan M. Mathiowetz) Date: Mon, 28 Aug 1995 07:39:23 -0500 (EDT) You should contact David Case of Scripps. He has a program Nucleic Acid Builder which does most of the things you want. I do not have his phone number, but you might be able to find it through the Web at http://www.scrips.edu/ Date: Mon, 28 Aug 1995 09:27:23 -0700 From: "Mark Hanning-Lee" Reply to: RE>CCL:DNA Modeling Software Just a thought: if you want to run software on SGI, you should upgrade your operating system. Irix 4.0.2 is years old; do yourself a favor and upgrade to the latest version (5.3 I think). Call 1 800 800 SGI1 (or perhaps 1 800 800 4SGI). You may find it worth your while to get a support contract which lets you get free operating system upgrades etc. Good luck with SGI! Mark From: esc@t10.Lanl.GOV (Gene Carter) Date: Tue, 29 Aug 1995 11:08:50 -0600 (MDT) Cc: cst@transposon.LANL.GOV The program NAMOT is capable of generating A,B. and Z forms of DNA as well as triplex and tetraplex. NAMOT is unable efficeintly model single stranded reigons which, of course, rules out generation of structures such as hairpins. The web site for this program is: http://synapse.lanl.gov/~namot Hope this helps. Gene --------------------------------------------------------------------------- - Group t-10 LANL |Los Alamos, NM 87545 | =========================================================================== Sincerely, Masayuki Yuki USC School of Pharmacy Graduate Program masayuki@hsc.usc.edu From haney@netcom.com Wed Aug 30 14:41:28 1995 Received: from netcom7.netcom.com for haney@netcom.com by www.ccl.net (8.6.10/950822.1) id OAA24427; Wed, 30 Aug 1995 14:28:49 -0400 Received: by netcom7.netcom.com (8.6.12/Netcom) id LAA25841; Wed, 30 Aug 1995 11:25:44 -0700 From: haney@netcom.com (Dr. David N. Haney) Message-Id: <199508301825.LAA25841@netcom7.netcom.com> Subject: CCL:Parity checking RAM in PCs To: CHEMISTRY@www.ccl.net Date: Wed, 30 Aug 1995 11:25:44 -0700 (PDT) X-Mailer: ELM [version 2.4 PL23] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 1593 On Tue, 29 Aug 1995 eknight@wppost.depaul.edu wrote: > The bottom line is this: if you run a calculation on one of the new PCs, > some number of bits _might_ have been randomly flipped. My question is > this: Is it legitimate to report calculations done on such a computer in a > scientific journal? Anyone that has run a long computation can sympathize with the concern that the computation may need to be replicated to show that no error has occured. However, I am not sure that anyone that has spent long hours in the lab to verify that their experiments are valid will sympathize. As an experimentalist turned computational, I find it odd that computational people generally believe that the computer hardware, software and the users's design are so perfect that one does not need to replicate an experiment to verify the conclusions. I know replicating computations can be expensive and boring (so are laboratory experiments), but this is the scientific method. The only answer that I see is replication, not only for parity errors, but software errors, user errors, Murphy's Law and cosmic rays. Then when you report the hardware, software and error found in the experiment, colleagues might have a clue how to try and duplicate this work. -- ************** David N. Haney, Ph.D. **************** * Haney Associates Phone - 619-566-1127 * * 12010 Medoc Ln. * * San Diego, CA 92131 Fax - 619-586-1481 * ************** Email - haney@netcom.com *************** From schwegle@chemsun.chem.umn.edu Wed Aug 30 15:11:28 1995 Received: from chemsun.chem.umn.edu for schwegle@chemsun.chem.umn.edu by www.ccl.net (8.6.10/950822.1) id PAA25025; Wed, 30 Aug 1995 15:00:55 -0400 Received: (from schwegle@localhost) by chemsun.chem.umn.edu (8.6.11/8.6.6) id NAA06174 for CHEMISTRY@www.ccl.net; Wed, 30 Aug 1995 13:55:51 -0500 From: Eric Schwegler Message-Id: <199508301855.NAA06174@chemsun.chem.umn.edu> Subject: Request for multipole expansion codes To: CHEMISTRY@www.ccl.net Date: Wed, 30 Aug 1995 13:55:49 -0500 (CDT) X-Mailer: ELM [version 2.4 PL23] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 956 Dear CCLers, I have developed new algorithms for computing the Cartesian multipole tensor (CPL, 67 {\bf 241} 1995**) and now I would like to compare how fast (or slow) they are relative to methods based on spherical tensor theory. So I was wondering if anyone might have some spherical harmonic code lying around that I could use for benchmark calculations? In particular I am looking for something written in Fortran for either vector or RISC computers. I am mainly interested in the generation of the interaction tensor to arbitrary order but I am also interested in the contraction of the tensor with a moment. Does anyone have something of this nature that I could have, or would any of you be willing to send me benchmark results? Thanks in advance, Eric Schwegler schwegle@chemsun.chem.umn.edu ** Several typo's managed to find there way into this letter, if you send me a request by e-mail I will give a typo-free version. From chpajt@bath.ac.uk Wed Aug 30 16:26:30 1995 Received: from goggins.bath.ac.uk for chpajt@bath.ac.uk by www.ccl.net (8.6.10/950822.1) id QAA26664; Wed, 30 Aug 1995 16:14:22 -0400 Received: from bath.ac.uk (actually host mary.bath.ac.uk) by goggins.bath.ac.uk with SMTP (PP); Wed, 30 Aug 1995 21:14:07 +0100 Date: Wed, 30 Aug 1995 21:13:53 +0100 (BST) From: A J Turner To: "Dr. David N. Haney" cc: CHEMISTRY@www.ccl.net Subject: Re: CCL:Parity checking RAM in PCs In-Reply-To: <199508301825.LAA25841@netcom7.netcom.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi! Surely it is necessary to validate computational results via experimental results. It is a good idea to use computers to make new and exciting insights into the nature as chemical reactions etc. However, the insights will always be based on theory, this theory - however ab-initio or otherwise pure it may be - is still theory. My point is that chemicals are not theories - they are real (let's not get too deep on this issue) and so no calculation - no matter how well replicated can be considered chemically relevant untill validated against experiment. If we consider that there is error in a calculation via bit flipping - this may or may not give enough error to make the calculation invalid. It may well be the case that small errors could creap in that do not deflect the result enough to make in chemically irrelevant. It is only comparaion with experiment that will final come down one way or the other. All the best Alex I'm not sure I beleive all that myself - but it is a point of view! +--------------------------------------------------+ |Alternate E-mail A.J.Turner@Bath.ac.uk | |www home @ http://www.bath.ac.uk/~chpajt/home.html| +--------------------------------------------------+ From jkl@ccl.net Wed Aug 30 17:11:30 1995 Received: from bedrock.ccl.net for jkl@ccl.net by www.ccl.net (8.6.10/950822.1) id QAA27356; Wed, 30 Aug 1995 16:59:27 -0400 Received: from origin for srheller@origin.gig.usda.gov by bedrock.ccl.net (8.6.10/950822.1) id QAA16003; Wed, 30 Aug 1995 16:59:25 -0400 Received: by origin (5.x/SMI-SVR4) id AA12105; Wed, 30 Aug 1995 16:44:02 -0400 Date: Wed, 30 Aug 1995 16:43:16 -0400 (EDT) From: "Stephen R. Heller" To: jcicshelp , chemistry@ccl.net, chminf-l@iubvm.ucs.indiana.edu, orgchem@extreme.chem.rpi.edu Subject: Software for Review Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII 30 August, 1995 Subject: Computer Software for Review As the Software Review Editor for the ACS Journal of Chemical Information and Computer Science (JCICS) I often get software for review in the journal. I have six (6) new software products. I am looking for someone who is willing to review these software product. In return for the review which is published in JCICS you get to keep the software or database. The review should be completed in 1-3 months. The length of the review is 4-10 double spaced typed pages. Sample reviews can be found in most of the recent issues of JCICS. Please try to give me some (short) reason to choose you over another person. I have tried this approach for about the past four years and it is working reasonably well. (REMINDER: For those who haven't finished your reviews of software sent months and months ago, this last sentence does not apply to you!) As a result, I am continuing this new method to find reviewers using this e- mail/user group system. I reserve the right to abandon this if it is a problem, or inappropriate. I will not notify people if I have found a reviewer. If you don't hear from me within a few days I have chosen someone else to review the particular package. As I get many, many, (too many) replies to this message, please do not respond after 1 September 1995 (Friday), as I am sure the software will be gone by then. I can be reached on Internet (SRHELLER@NALUSDA.GOV). PLEASE BE SURE TO INCLUDE AN STREET ADDRESS, PHONE, and FAX NUMBER!!! (I send the software by Federal Express.) Without this information I WILL NOT consider your request. Steve Heller The packages I now have are: 1. The software program HINT from Haney Associates. HINT is a program that calculates hydrophobic and hydrophilic partitioning parameters based on empirical data. HINT is available for Sybyl, InsightII and Chem-X. This program runs ONLY on the SGI. In order to provide the reviewer with the correct software, please provide answers to the following questions when you respond to this message. hostname: CPU type (use hinv): SGI: /etc/sysinfo -s Modeling Host Software: Tape Media: SYBYL Release/Version: ______ QIC InsightII Release/Version: ______ 4mm DAT ChemX Release/Version: ______ 8mm Exabyte 2. ChemSite for Windows. This a 3-D molecular modeling program for drawing, displaying, and simulating dynamic behavior of organic molecules. 3.gNMR, a Macintosh program from SoftShell International and Cherwell Scientific. This program performs 1-D nmr spectral simulation. It requires a 68020 or higher cpu, a 68881 numeric coprocessor, and system 6.0.5 or higher. 4. SciProtein from SciVision is a PC program that is a toolkit for proteins and peptides. The five main features are homology, prediction, graphics editor, search and find, and databases and property maps. You should have HyperChem. It interfaces with HyperChem 3.0 or 4.1. The program and database needs 50-80 MB of storage! Excel 5.0 should be available for rapid searches of the PDB. This program is for protein chemists and molecular biologists. 5. SciPolymer from SciVision is a PC program that contains an informational database and predictive tools to calculate polymer properties. You should have HyperChem 30. or 4.1. 6. EndNote2 Plus and EndLink 2.0 for Windows from Niles & Associates. These program help one manage bibliographic references. Steve Heller, USDA, ARS, Plant Genome Program Bldg. 005, Room 337 Beltsville, MD 20705-2350 E-Mail: srheller@nalusda.gov Phone: 301-504-6055 FAX: 301-504-6231 From sxr224@anugpo.anu.edu.au Wed Aug 30 23:11:34 1995 Received: from anugpo.anu.edu.au for sxr224@anugpo.anu.edu.au by www.ccl.net (8.6.10/950822.1) id XAA00688; Wed, 30 Aug 1995 23:01:54 -0400 Received: from surya.anu.edu.au (surya.anu.edu.au [150.203.7.194]) by anugpo.anu.edu.au (8.6.12/8.6.12) with SMTP id NAA21722 for ; Thu, 31 Aug 1995 13:01:51 +1000 Message-Id: <199508310301.NAA21722@anugpo.anu.edu.au> Comments: Authenticated sender is From: "Shoba Ranganathan" To: CHEMISTRY@www.ccl.net Date: Thu, 31 Aug 1995 13:00:33 +0000 Subject: CCL: Density maps from protein structure Reply-to: Shoba.Ranganathan@anu.edu.au Priority: normal X-mailer: Pegasus Mail for Windows (v2.0-WB3) Dear CCLers, I have a model protein structure which I need to compare with electron density maps of a related protein. Something like the converse of what an x-ray crystallographer would do. Is there a program that can generate density plots similar to those obtained from EM data - projections onto planes that can be stacked? I can do this the hard way: sort out atoms from the z-coordinate, project onto plane, use some "surface" representation, but I was wondering if someone hasn't already done this before me. My protein has ~350 residues: hence calculation of electron density as in SPARTAN won't be feasible. Please e.mail your replies to me and I will summarize if required. Thank you. Shoba =========================================================== Dr. Shoba RANGANATHAN Computational Mol Biology & Drug Design Group Div. of Biochemistry & Mol. Biology John Curtin School of Medical Research Australian National University Tel: +616-279-8301 Canberra ACT 2002 Fax: +616-249-0415 Australia. email:Shoba.Ranganathan@anu.edu.au ---------------------------------------------------------- From Matthew.Harbowy@tjlus.sprint.com Wed Aug 30 23:26:36 1995 Received: from merit-mx1.telemail.net for Matthew.Harbowy@tjlus.sprint.com by www.ccl.net (8.6.10/950822.1) id XAA00855; Wed, 30 Aug 1995 23:19:33 -0400 From: X400-Received: by /PRMD=LANGATE/ADMD=TELEMAIL/C=GB/; Relayed; Wed, 30 Aug 1995 19:14:00 -0400 X400-Received: by /PRMD=LANGATE/ADMD=TELEMAIL/C=GB/; Relayed; Wed, 30 Aug 1995 19:14:00 -0400 X400-Received: by /PRMD=LANGATE/ADMD=TELEMAIL/C=GB/; Relayed; Wed, 30 Aug 1995 19:14:00 -0400 X400-Received: by /PRMD=LANGATE/ADMD=TELEMAIL/C=GB/; Relayed; Wed, 30 Aug 1995 19:14:00 -0400 Date: Wed, 30 Aug 1995 19:14:00 -0400 X400-Originator: Matthew.Harbowy@tjlus.sprint.com X400-Recipients: non-disclosure:; X400-MTS-Identifier: [/PRMD=LANGATE/ADMD=TELEMAIL/C=GB/;Wed Aug 30 19:14:14 199500] X400-Content-Type: P2-1984 (2) Content-Identifier: Re: CCL:Parity c Message-ID: <"Wed Aug 30 19:14:14 199500*/G=Matthew/S=Harbowy/OU=2812NJLP/O=TMUS.TJL/PRMD=LANGATE/ADMD=TELEMAIL/C=GB/"@MHS> To: chemistry@www.ccl.net Subject: Re: CCL:Parity checking RAM in PCs MIME-version: 1.0 Content-type: multipart/mixed; boundary="5WXxUtNNGzm9g8GN5sbzGBznp4iphGxe" --5WXxUtNNGzm9g8GN5sbzGBznp4iphGxe Content-type: text/plain; charset="us-ascii" ---------------------------------- Forwarded ---------------------------------- From: Matthew Harbowy at 2812NJLP Date: 8/30/95 6:49PM To: chemistry-request@www.ccl.net at INTERNET Subject: Re: CCL:Parity checking RAM in PCs ------------------------------------------------------------------------------- --5WXxUtNNGzm9g8GN5sbzGBznp4iphGxe Content-type: text/plain; charset="us-ascii" RFC-822-Headers: X-Mailer: ELM [version 2.4 PL23] Errors-To: ccl@www.ccl.net Precedence: bulk --5WXxUtNNGzm9g8GN5sbzGBznp4iphGxe Content-type: text/plain; charset="us-ascii" On Tue, 29 Aug 1995 eknight@wppost.depaul.edu wrote: > The bottom line is this: if you run a calculation on one of the new PCs, > some number of bits _might_ have been randomly flipped. My question is > this: Is it legitimate to report calculations done on such a computer in a > scientific journal? Anyone that has run a long computation can sympathize with the concern that the computation may need to be replicated to show that no error has occured. However, I am not sure that anyone that has spent long hours in the lab to verify that their experiments are valid will sympathize. As an experimentalist turned computational, I find it odd that computational people generally believe that the computer hardware, software and the users's design are so perfect that one does not need to replicate an experiment to verify the conclusions. I know replicating computations can be expensive and boring (so are laboratory experiments), but this is the scientific method. The only answer that I see is replication, not only for parity errors, but software errors, user errors, Murphy's Law and cosmic rays. Then when you report the hardware, software and error found in the experiment, colleagues might have a clue how to try and duplicate this work. ------------------------------------ Even if an experiment is performed in duplicate, it is not like experimental work where you might take an 'average' over replicates to determine variation. Sadly, computation work is 'supposed' to be exact- perhaps this says a lot for using monte-carlo approaches where a certain degree of randomness is necessary. An approach i've only toyed with is parallel monte-carlo where for iterations, you don't force parallel calcualtions running on separate processors to all finish to update a matrix: you update as the results come back and start iterations with partially updated matrices. A well-disciplined algorithm might then take advantage of the inherent randomness of processor assignments across a multi-user, massively parallel supercomputer as a way of bringing 'true' randomness into a calculation. I don't have the whereabouts to run with the idea: It's free to a good home. matt who doesn't speak for the thomas j. lipton company --5WXxUtNNGzm9g8GN5sbzGBznp4iphGxe--