From lestaw.k.bieniasz@uni-tuebingen.de Wed Apr 16 04:36:34 1997 Received: from outmail.zdv.uni-tuebingen.de for lestaw.k.bieniasz@uni-tuebingen.de by www.ccl.net (8.8.3/950822.1) id DAA09936; Wed, 16 Apr 1997 03:49:41 -0400 (EDT) Received: from by outmail.zdv.uni-tuebingen.de (4.1/ZDV-Uni-Tuebingen-1.0) id AB29857; Wed, 16 Apr 97 09:49:07 MES Received: from textserv.zdv.uni-tuebingen.de by mailserv.uni-tuebingen.de with SMTP (PP); Wed, 16 Apr 1997 09:48:36 +0200 Received: (from coibi01@localhost) by textserv.zdv.uni-tuebingen.de (8.6.12/8.6.12) id JAA21949; Wed, 16 Apr 1997 09:48:28 +0200 Date: Wed, 16 Apr 1997 09:48:27 +0200 (MSZ) From: "Lestaw K. Bieniasz" To: chemistry@www.ccl.net Subject: computational chemistry Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Tuebingen, 16.04.97 Dear Computational Chemists, I am new to this list, so please excuse me if questions similar to those asked below have already been discussed. I need very much your help and advice. Let me first introduce myself. For about 18 years my research has been associated with the computer modelling of the transient experiments in electrochemical kinetics. Briefly speaking this usually consists in solving sets of reaction-diffusion partial differential equations with various complications. I think my work has evolved from something one might classify as applications of various computational methods to very concrete electrochemical problems, to a more general approach, in which I am trying to: (a) Design universal computational strategies aplicable to a large class of electrochemical kinetic problems. (b) Pay more attention to the analysis of the computational methods themselves, having in mind particular requirements of the electrochemical kinetics. (c) Better understand various rational procedures and reasoning processes that are used in electrochemical kinetics, with the aim to enable automating these processes as much as possible, and represent them in the form of computer algorithms. (d) Elaborate a comprehensive problem-solving environment for electrochemical kinetic simulations, which apart from considering many example problems, involves also elements of algorithms/software design and pure programming. I have to say I have never thought of how to formally classify the above area of investigations, and I was quite satisfied with the fact that the reearch was interesting, seemingly modern and potentially very promising for the future, if one takes into account the enormous recent progress in computing technology. I also did not have much difficulties with publishing the results of my work, which additionally reinforced my good feeling that I have chosen the right direction. Unfortunately I am beginning to encounter criticism from the side of several chemists in my environment, who say that what I am doing is not electrochemistry, because there is so much numerical mathematics and computational science in my works. It seems that most electrochemists think that one needs to solve very specific electrochemical problems (like finding out what the mechanism of some reaction is, for example) in order to make progress in electrochemistry. They don't notice that one might also make a progress by making it possible that many tedious operations that the electrochemists had so far to do by hand, can now be done easily and rapidly by means of specially designed computer programs. They don't seem to notice that certain traditional research paradigms do not longer make much sense (for example, publishing long derivations for some kinetic models in the situation when a properly designed program might even create a whole paper automatically). These problems have drawn my attention to the issue of how to classify and better justify my work, and here go my questions: (1) My first idea was to say that I am doing computational chemistry. By analogy with computational physics which practically includes everything that has to do with physics and computers simultaneously, I would expect that my area of investigations might be called computational electrochemistry. Logically, if computational chemistry is now a part of chemistry, then I might keep saying that my area of interest belongs to electrochemistry. How much I was mistaken, though! Most of the computational chemists seem to put an equality sign between quantum mechanical calculations and computational chemistry, as if there existed no other possible uses of computers in chemistry. For example, the Journal of Computational Chemistry publishes almost exclusively quantum mechanical stuff. This fact was even reflected in one of the first definitions of computational chemistry cited in the preface to "Reviews of Computational Chemistry", vol. 1: "Use of molecular orbital theorems to determine structure and properties of molecules". Fortunately, the editors of this series officially accept a more elastic definition, and in vol. 2 they write: "The quantitative modelling of of chemical phenomena by computer-implemented techniques is how we view the scope of the field, so it includes practically all aspects of chemical research that are expedited or rendered practical by computers". However, in the entire set of reviews (7 volumes) I have managed to trace down only one article mentioning briefly some chemical kinetic simulation program. The entire huge area of kinetic modelling seems therefore not to exist in the consciousness of the computational chemists, as a part of their area. I would be glad to know what you, the CCL subscribers, think about this. (2) After thinking for a while about point (1) I have started to have some doubts, though. It might be that I do not precisely understand the meaning of the word "computational", but for me it seems to mean something close to "numerical calculations", whereas computers can do much more. For example, they can do logical operations, largely used in AI methods, e.g. in expert systems. Also, neural networks are becoming increasingly popular in solving chemical problems, not necessarily computational, but perhaps "intellectual", such as pattern recognition, signal processing, machine learning for chemical purposes. I feel a more general term like "computer-aided chemistry" would be more adequate to describe the current developments at the junction computers+chemistry. What do you think? (3) To further extend point (2), I'd like to quote Allen Newell, an AI expert [cf. D.G.Bobrow and P.J.Hayes, Artif. Intelligence 25(1985)3.]: "We should, by the way, be prepared for some radical, and perhaps surprising, transformations of the disciplinary structure of science (technology included) as information processing pervades it. In particular, as we become more aware of the detailed information processes that go on in doing science, the sciences will find themselves increasingly taking a metaposition, in which doing science (observing, experimenting, theorizing, testing, archiving,...) will involve understanding these processes, and building systems that do the object-level science." Isn't this what is happening now? Unfortunately, Allen Newell was not a chemist, so that I can't use this argument to convince chemists. Does anybody of you know about similar opinions being expressed by chemists? If possible, by prominent chemists? I'd be happy to know (references or addresses, please!) Or perhaps I am wrong, and even you, the declared computational chemists, do not share the above opinion, and you rather prefer to put clear dividing lines between all chemical and computational aspects of the computational chemistry, seeing only applications of computational science to chemistry, instead of a new emerging area that combines both aspects? Do you use to say, "before lunch we shall solve some chemical problem, and afternoon we shall work a litle on numerical methods" ? (4) By the way, do you know any papers, materials, internet resources, etc. devoted to the predictions of how the use of computers in chemistry (and in other natural sciences) may look like in the nearest decades? I'd like to find such predictions. (5) I wonder how those of you, who may possibly develop chemical expert systems, classify your research. Do you think that a research on what can be regarded as sound reasoning in chemistry is a subject belonging to chemistry? Or to computational science? What do the referees of your research proposals, and sponsors think about it? I hope I am not taking too much of your (computational) time. Yours sincerely L.Bieniasz *-------------------------------------------------------------------* | Dr. Leslaw Bieniasz | | temporary address: (3rd April 1997 - 31st August 1997) | | Institut fuer Organische Chemie, Universitaet Tuebingen | | Auf der Morgenstelle 18, 72076 Tuebingen, Germany. | | E-mail: lestaw.k.bieniasz@uni-tuebingen.de | *-------------------------------------------------------------------* | permanent address: | | Institute of Physical Chemistry of the Polish Academy of Sciences,| | Molten Salts Laboratory, ul. Zagrody 13, 30-318 Cracow, Poland. | | E-mail: nbbienia@cyf-kr.edu.pl | *-------------------------------------------------------------------*