Signals Blog


Back when mainframe terminals were the hot new thing, entry of chemical structures was difficult at best. To solve this problem, a number of line notations were proposed, adopted by a few groups around the world, then abandoned with virtually no trace. In one of the best-known cases, Wiswesser Line Notation (WLN) rose to dominance in the early 1980s, only to fall into obscurity within a few years.


What replaced WLN and all of the other attempts to make it easy to work with chemical structures as text strings was a new system, Simplified Molecular-Input Line-Entry System (SMILES). One of the key advantages of SMILES was that it was far easier to encode/decode than the alternatives at the time such as WLN.


Despite increasing popularity and the appearance of a number of publications dedicated to the subject, one of the persistent problems with SMILES was its lack of precise documentation. An email discussion started by Craig James in 2007 brought a number people together to try to solve this problem:

Recently on the OpenBabel-development mailing list, it was concluded that a new SMILES specification is needed, created and maintained by the open-source community.

The OpenSMILES Specification

Through an ongoing, multi-year process of discussion and revision from many individuals, the OpenSMILES Specification distills everything written about SMILES in the public literature into a single document.

Anyone using or considering using SMILES would do well to refer to the OpenSMILES specification along with other public documentation.