Mercury is a pure logic programming language intended for the creation of large, fast, reliable programs. The syntax of Mercury is based on the syntax of Prolog, but semantically the two languages are very different due to Mercury's purity, its type, mode, determinism and module systems.
Some example programs written in Mercury can be found here.
The main features of Mercury are:
Mercury is purely declarative: predicates and functions in Mercury do not have non-logical side effects.
Mercury does I/O through built-in and library predicates that take an old state of the world and some other parameters, and return a new state of the world and possibly some other results. The language requires that the input argument representing the old state of the world be the last reference to the old state of the world, thus allowing the state of the world to be updated destructively. The language also requires that I/O take place only in parts of the program where backtracking will not be needed.
Mercury handles dynamic data structures by providing several abstract data types in the standard library that manage collections of items with different operations and tradeoffs. Programmers can also create their own abstract data types.
Mercury is a strongly typed language.
Mercury's type system is based on many-sorted logic with parametric polymorphism, very similar to the type systems of modern functional languages such as ML and Haskell. Programmers must declare the types they need using declarations such as
:- type list(T) --->  ; [T | list(T)]. :- type maybe(T) ---> yes(T) ; no.
They must also declare the type signatures of the predicates they define, for example
:- pred append(list(T), list(T), list(T)).
The compiler infers the types of all variables in the program. Type errors are reported at compile time.
Mercury is a strongly moded language.
The programmer must declare the instantiation state of the arguments of predicates at the time of the call to the predicate and at the time of the success of the predicate. Currently only a subset of the intended mode system is implemented. This subset effectively requires arguments to be either fully input (ground at the time of call and at the time of success) or fully output (free at the time of call and ground at the time of success).
A predicate may be usable in more than one mode. For example, append is usually used in at least these two modes:
:- mode append(in, in, out). :- mode append(out, out, in).
If a predicate has only one mode, the mode information can be given in the predicate declaration.
:- pred factorial(int::in, int::out).
The compiler will infer the mode of each call, unification and other builtin in the program. It will reorder the bodies of clauses as necessary to find a left to right execution order; if it cannot do so, it rejects the program. Like type-checking, this means that a large class of errors are detected at compile time.
Mercury has a strong determinism system.
For each mode of each predicate, the programmer should declare whether the predicate will succeed exactly once (det), at most once (semidet), at least once (multi) or an arbitrary number of times (nondet). These declarations are attached to mode declarations like this:
:- mode append(in, in, out) is det. :- mode append(out, out, in) is multi. :- pred factorial(int::in, int::out) is det.
The compiler will try to prove the programmer's determinism declaration using a simple, predictable set of rules that seems sufficient in practice (the problem in general is undecidable). If it cannot do so, it rejects the program.
As with types and modes, determinism checking catches many program errors at compile time. It is particularly useful if some deterministic (det) predicates each have a clause for each function symbol in the type of one of their input arguments, and this type changes; you will get determinism errors for all of these predicates, telling you to put in code to cover the case when the input argument is bound to the newly added function symbol.
Mercury has a module system.
Programs consist of one or more modules. Each module has an interface section that contains the declarations for the types, functions and predicates exported from the module, and an implementation section that contains the definitions of the exported entities and also definitions for types and predicates that are local to the module. A type whose name is exported but whose definition is not, can be manipulated only by predicates in the defining module; this is how Mercury implements abstract data types. For predicates and functions that are not exported, Mercury supports automatic type, mode, and determinism inference.
Mercury supports higher-order programming, with closures, currying, and lambda expressions.
Mercury is very efficient
Strong types, modes, and determinism provide the compiler with the information it needs to generate very efficient code.
The Mercury compiler is written in Mercury itself. It was bootstrapped using NU-Prolog and SICStus Prolog. This was possible because after stripping away the declarations of a Mercury program, the syntax of the remaining part of the program is mostly compatible with Prolog syntax.
The Mercury compiler compiles Mercury programs to C, which it uses as a portable assembler. The system can exploit some GNU C extensions to the C language, if they are available: the ability to declare global register variables, the ability to take the addresses of labels, and the ability to use inline assembler. Using these extensions, it generates code that is significantly better than all previous Prolog systems known to us. However, the system does not need these extensions, and will work in their absence.