%0 Journal Article
%T The semantics of Chemical Markup Language (CML): dictionaries and conventions
%A Peter Murray-Rust
%A Joe A Townsend
%A Sam E Adams
%A Weerapong Phadungsukanan
%A Jens Thomas
%J Journal of Cheminformatics
%D 2011
%I BioMed Central
%R 10.1186/1758-2946-3-43
%X From an early stage, Chemical Markup Language (CML) was designed so that it could accommodate an indefinitely large amount of chemical and related concepts. This objective has been achieved by developing a dictionary mechanism where many of the semantics are added not through hard-coded elements and attributes but by linking to semantic dictionaries. CML has a number of objects and object containers which are abstract and which can be used to represent the structure and datatype of objects. The meaning of these, both for humans and machines, is then realised by linking an appropriate element in a dictionary.The dictionary approach was inspired by the CIF dictionaries [1] from the International Union of Crystallography (IUCr) and has a similar (in many places isomorphous) structure to that project. The design allows for an indefinitely large number of dictionaries created by communities within chemistry who recognise a common semantic approach and who are prepared to create the appropriate dictionaries. At an early stage, CML provided for this with the concept of "convention". This attribute is an indication that the current element and its descendants obey semantics defined by a group of scientists using a particularly unique label (Figure 1).During the evolution of CML we explored a number of syntactic approaches to representing and imposing semantics through dictionaries. These have ranged from a formally controlled ontology (ChemAxiom [2]) which is consistent with OWL2.0 [3] and the biosciences' Open Biological and Biomedical Ontologies (OBO)[4] framework, to uncontrolled folksonomy-like tagging. Although we have implemented ChemAxiom and it is part of the bioscientists' description of chemistry, we regard it as too challenging for the current practice of chemistry and unnecessary for its communication. This is because chemistry has a well-understood (albeit implicit) ontology and the last 15 years have confirmed that it is highly stable. The power of declaration
%U http://www.jcheminf.com/content/3/1/43