Why is modelling in chemistry much more difficult and challenging than in the technical sciences such as aeronautics or civil engineering?

Chemistry is rather old and so is thinking about the reasons for the many observations of chemical processes in daily life. With concepts of atoms and molecules, many chemical processes such as chemical reactions, phase separation, solvation, ligand-protein binding, and protein folding, are nowadays investigated using numerical chemical models and state-of-the-art computing devices. Yet, the accuracy and applicability of theoretical computational models in chemistry is rather limited due to a variety of factors:
1. The degrees of freedom governing chemical processes are electronic, nuclear, atomic and molecular, and the corresponding particles have very different masses and sizes.
2. The interactions between these particles are governed by quantum mechanics, that is, the Dirac or Schrödinger equations of motion, or under particular conditions of not too small mass or low temperature, by classical equations of motion, that is, those of Newton, Lagrange, Hamilton or Langevin.
3. At non-zero temperatures, the behaviour of the particles is in addition governed by
statistical mechanics, that is, Bose–Einstein, Fermi–Dirac, or Boltzmann ensembles of
configurations are to be considered, not single structures.
4. Compared to the strong interaction between nucleons, the Coulomb interaction between nuclei and electrons is spatially rather long-ranged, which induces many-body effects that make an accurate modelling of the resulting forces computationally expensive.
5. The energy or free energy changes of chemical processes can be very small compared to the total energy of the interacting particles involved in the process.
6. The time scale of different chemical processes may easily span 15 orders of magnitude. These features of chemical processes severely complicate the formulation of accurate predictive models in chemistry. Yet, computational modelling of chemical processes is practised, because it is needed for the interpretation of experimental observations and because it complements the experimental methodology of investigation, in particular for microscopic length and time scales.
The basic challenges and choices with respect to modelling in chemistry will be reviewed and illustrated.