Refereeing Semiempirical Papers

[UDIM018%FRORS31.BITNET &$at$& phem3.acs.ohio-state.edu]

 E. M. EVLETH
 Dynamique des Interactions Moleculaires
 Universite Pierre et Marie Curie
 4 Place Jussieu, Tour 22, Paris 75005
 33-1-44-27-42-08 (work), 33 = France; 1 = Paris
 33-1-45-48-67-20 (home)
 FAX 33-1-44-27-41-17 (lab);44-27-38-66(University)
 e-mail UDIM018 at FRORS31.BITNET
 Henry Rzepa's comments on refereeing papers and reproducible results
 allows me to call attention to some of the earlier versions of AMPAC. I
 rejected an article several year ago which claimed to have done AM1
 calculations on sulfur containing structures using a version of AMPAC
 dated before that that parameterization had done. Moreover, the article
 containing the AM1 parameterizations had not yet been published. The
 authors had already published several previous papers using "AM1"
 parameter- zations. Neither the authors nor the referees could check the
 sulfur compound calculations with the literature. I told the authors to
 check with Dewar. They did and it turned out that earlier versions of
 AMPAC would grab MNDO sulfur parameters and do "the calculation".  The
 authors had to call attention to the fact that their earlier papers used
 "mixed AM1-MNDO" parameters. Later versions of AMPAC have been
 corrected
 and do not mix parameters.
 We have cited in some of own published work the inability of reproducing
 the semiempirical calculations of others who have programmed their own
 version of a particular semiempirical parameterization. Now that SAM1
 tables are published at least one can check one's results against those
 generated by AMPAC 4.5 and upcoming versions of that program. The debate
 of whether one can alternatively program SAM1 into one's own program
 based on published information will go on for while. But based on past
 experience this will generate some faulted programs.   Those who use
 them will have to continually check their results with the original
 paper.  Past experience also indicates that all will not and referees
 will have to run check calculations to verify results in submitted
 papers.  Not all referees will. So let the reader beware. The
 advantage of using a commercial program are net. If that program
 is wrong, one has somebody else to blame. That and the down time
 in writing one's own program is worth the price of a program.