The SciFinder Scholar (SFS) search system was developed by Chemical Abstracts Service (CAS) and has been marketed to academic institutions since 1998. 1 SFS is an elegant search interface to six core chemical related databases: CAplus, the main Chemical Abstracts literature database, CAS REGISTRY, the substance database , CASREACT, providing access to reactions from the journal literature and patents, CHEMLIST, the regulatory chemicals database, CHEMCATS, listing suppliers of commercially available chemicals worldwide, and the complete National Library of Medicine’s (NLM) MEDLINE database back to 1950 containing over 15 million references.The natural language query (NLQ) processing algorithms involved in the SciFinder Scholar (SFS) include prepositions, Boolean operators, abbreviations, and automatic truncation. The best results are obtained by liberal use of prepositions, for example, ‘use of antibiotics to prevent diseases in cows’. SFS recognizes prepositions, does not analyze their linguistic meaning, and uses them to break the query into component concepts.
No truncation symbols are recognized. Boolean operators “AND”, “OR”, and “NOT” are not treated as Boolean operators, but rather as conversational English. However, once the NLQ algorithm identifies query concepts and performs the search, Boolean logic is used to generate the pick list of the various search results presented to the user. Parentheses can only be used to specify up to three synonyms separated by commas after a given term.An example of this syntax would be ‘paints (coatings) which are fire retardant (flame resistant, intumescent)’. Case is ignored. Searching ‘AIDS’ or ‘aids’ generates identical results. Also ignored are single and double quotes.
Given the size and technical nature of the database, querywords are not run against a spelling checker, except for some of the most common misspellings. 1. Somerville, A. N. SciFinder Scholar.
J. Chem. Educ. 1998, 75, 959, 975-976.