There are general problems representing delocalized bonding and organometallic compounds in cheminformatics (generally) and SMILES in specific.

To quote the Open SMILES standard:

This simple mental model [of a connection table between atoms using valence bonds] has little resemblance to the underlying quantum-mechanical reality of electrons, protons and neutrons, yet it has proved to be a remarkably useful approximation of how atoms behave in close proximity to one another. However, the valence model is an imperfect representation of molecular structure, and the SMILES language inherits these imperfections. Chemical bonds are often tautomeric, aromatic or otherwise fractional rather than neat integer multiples. Delocalized bonds, bond-centered bonds, hydrogen bonds and various other inter-atom forces that are well characterized by a quantum-mechanics description simply don’t fit into the valence model.

Other interesting cases are agostic interactions, 3-center 2-electron bonds (e.g., diborane).

There are several suggestions for advancing beyond the typical connection-table valence bond view, including:

• "Yet Another Representation of Molecular Structure" J. Chem. Inf. Comput. Sci., 1995, 35 (5), pp 787–802
• "Overcoming the Limitations of a Connection Table Description:  A Universal Representation of Chemical Species" J. Chem. Inf. Comput. Sci., 1997, 37 (4), pp 705–714

That both papers were published 17+ years ago should be some indication of:

• Chemists like the valence bond view of molecules, despite the limitations
• How complicated bonding can be in non-organic chemistry
• How far cheminformatics needs to go to properly handle organometallic and non-organic compounds

By the way... anyone interested in implementing such efforts should let me know.. Many people would be happy to help. :-)

There are general problems representing delocalized bonding and organometallic compounds in cheminformatics (generally) and SMILES in specific.

To quote the Open SMILES standard:

This simple mental model [of a connection table between atoms using valence bonds] has little resemblance to the underlying quantum-mechanical reality of electrons, protons and neutrons, yet it has proved to be a remarkably useful approximation of how atoms behave in close proximity to one another. However, the valence model is an imperfect representation of molecular structure, and the SMILES language inherits these imperfections. Chemical bonds are often tautomeric, aromatic or otherwise fractional rather than neat integer multiples. Delocalized bonds, bond-centered bonds, hydrogen bonds and various other inter-atom forces that are well characterized by a quantum-mechanics description simply don’t fit into the valence model.

Other interesting cases are agostic interactions, 3-center 2-electron bonds (e.g., diborane).

There are several suggestions for advancing beyond the typical connection-table valence bond view, including:

• "Yet Another Representation of Molecular Structure" J. Chem. Inf. Comput. Sci., 1995, 35 (5), pp 787–802
• "Overcoming the Limitations of a Connection Table Description:  A Universal Representation of Chemical Species" J. Chem. Inf. Comput. Sci., 1997, 37 (4), pp 705–714

That both papers were published 17+ years ago should be some indication of:

• Chemists like the valence bond view of molecules, despite the limitations
• How complicated bonding can be in non-organic chemistry
• How far cheminformatics needs to go to properly handle organometallic and non-organic compounds

By the way... anyone interested in implementing such efforts should let me know.. Many people would be happy to help. :-)

UPDATE

I should point out that there are some practical solutions to representing such things. For example Alex Clark suggested recording zero-order bonds (and sometimes dummy atoms):

"Accurate Specification of Molecular Structures: The Case for Zero-Order Bonds and Explicit Hydrogen Counting" J. Chem. Inf. Model., 2011, 51 (12), pp 3149–3157