Abstract
Lenses provide an abstract interface for manipulating data types through spatially separated
views. They are defined abstractly in terms of two functions, get, the return a value from the source type, and put that updates the value. We mechanise the underlying theory of lenses, in terms of an algebraic hierarchy of lenses, including well-behaved and very well-behaved lenses, each lens class being characterised by a set of lens laws. We also mechanise a lens algebra in Isabelle that enables their composition and comparison, so as to allow construction of complex lenses. This is accompanied by a large library of algebraic laws. Moreover we also show how the lens classes can be applied by instantiating them with a number of Isabelle data types. This theory development is based on our recent paper, which shows how lenses can be used to unify heterogeneous representations of state-spaces in formalised programs.
views. They are defined abstractly in terms of two functions, get, the return a value from the source type, and put that updates the value. We mechanise the underlying theory of lenses, in terms of an algebraic hierarchy of lenses, including well-behaved and very well-behaved lenses, each lens class being characterised by a set of lens laws. We also mechanise a lens algebra in Isabelle that enables their composition and comparison, so as to allow construction of complex lenses. This is accompanied by a large library of algebraic laws. Moreover we also show how the lens classes can be applied by instantiating them with a number of Isabelle data types. This theory development is based on our recent paper, which shows how lenses can be used to unify heterogeneous representations of state-spaces in formalised programs.
Original language | English |
---|---|
Pages (from-to) | 1-28 |
Number of pages | 28 |
Journal | Archive of Formal Proofs |
Publication status | Published - 1 Jun 2017 |