Why Molecular Modeling?

In early 1990, Autodesk was evaluating HyperChem, a Microsoft Windows-based molecular modeling package developed by Hypercube, Inc. (Details on how Autodesk and Hypercube were introduced may be found on page .) Technical evaluation of the product quickly established that the product was scientifically correct, well-built, and provided unprecedented ease-of-use for a molecular modeler. The biggest question in the minds of Autodesk management was whether the molecular modeling market was big enough to justify the effort, and if it made sense to get involved in a product so technically complicated, computationally intense, and specialised. The following is an advocacy piece I wrote which argues that every single one of the present doubts about HyperChem could have been raised in 1982 against AutoCAD. On August 17, 1990 Autodesk made an investment in Hypercube and obtained exclusive distribution rights to HyperChem; the product shipped on March 30, 1992. In January 1994, the Scientific Modeling Division was disbanded and the distribution of the product terminated.

As we proceed with evaluation of a possible business relationship with Hypercube, I'd like to briefly discuss the most fundamental question, ``Why should Autodesk get involved in molecular modeling?'' Agreement with the reasons I'll describe below and with the conclusion that we should enter that market when an attractive opportunity to do so appears does not necessarily mean we should conclude a relationship with Hypercube; that specific decision must hinge on our evaluation of that company and its principals, products, and strategy. But conversely, if you conclude that molecular modeling is a hopelessly narrow, arcane, and hard-to-sell niche market, there's no reason to waste any time evaluating the particulars of Hypercube.

One of the major tasks I had set for myself in the early part of this year was to write a paper that laid out my view of the currently emerging technologies in which Autodesk could become involved with relatively little investment or risk, and which were consistent with the overall direction of the company and the businesses in which we do well. I had scheduled a Technology Forum for January 27th titled, ``What Next: The Coming Revolution In Manufacturing'' to discuss just these items. Unfortunately, the recent spate of alarums and diversions has pushed that task further and further back in the queue, necessitating the somewhat sketchy arguments presented below. In any case, things are happening in this area much, much faster than I or many of the most voluble visionaries in the field had expected only a few months ago. I have been accumulating a file of clippings to circulate with the paper, should I ever finish it, and it seems like every week's Science and Nature contain increasing quantities of relevant material.

I believe that molecular modeling passes the two key tests I look for in evaluating any potential new product area. These filters are the same I applied when screening initial product ideas for Autodesk, and the fact that AutoCAD met these criteria was one of the reasons I strongly supported it as a product for Autodesk.

• Filter 1: It stands to benefit from increasing compute power
• Filter 2: It is central to emerging technologies
• Conclusion