Glossa is a language for specifying the semantics (``what the words mean'') of ideas, data, and computations. (``Glossa'' means ``tongue'', and ``language'', in Greek.)

With the diversity of databases and computational resources serving biology today, it would be wonderful to have these be able to call upon each other's data and computations reliably and accurately. Unfortunately, that's not yet possible, because every database, like every scientist, has its own notion of what the words mean. Just try getting ten people to agree on the definition of a gene! The problem this presents for computations is in knowing how to correctly use the information retrieved, assuming it can be retrieved. The goal of Glossa is to provide a common language for defining the semantics of each computational resource's notions. Once these are computably defined, then translating programs can translate the definitions. Because the definitions are computational constructs, not natural language, and obey some grammatical rules, writing such translation programs is much easier.

One defines ideas with Glossa by building them from axioms, called semiotes, using the grammatical rules. A group of semiotes is called a bundle. Each semiote and bundle have four components of its definition: formal and informal (natural language) definitions of its syntax and semantics. Semiotes: Prolog files defining the semiotes' formal semantics, informal semantics, formal syntax, and informal syntax. We use a set of regular expressions to quickly and conveniently specify syntax. Bundles: Prolog files defining bundles's formal semantics, informal semantics, formal syntax, and informal syntax. Glossa's Grammar: A Backus-Naur Form grammar is shown. This is a left-recursive, context-free grammar. Glossa Manual: The current complete list and definitions of semiotes and bundles (PostScript and PDF).