The Chemical Exchange Problem

Storing chemical information in a computer is not a trivial task. Many different approaches and formats have been used. Most that worked at all are still with us. For instance, there are between 30-40 important chemical formats—managing them is a formidable and expensive task. A small part of the problem is that arbitrary methods are used for encoding data, e.g., double bonds can be represented as the integer 2, the real bond order 2.0, the symbol "=", the enumeration DOUBLE, and as repeated connections. A bigger part of the problem is that most chemical file formats contain information which is meaningless except in terms of a specific program, e.g., "tautomeric", "ring-double", "exo-double " and "fragment double " bonds The biggest part of the problem is that different programs that process chemical information use different underlying models, e.g., in ab initio or M.O. programs, the idea of "bond " isn't a particularly useful concept. To be useful, we must provide an accurate method for representing the underlying data model. Approaches that don't work wel