GNU APL

GNU APL is a free and (almost) complete implementation of Extended APL as specified in ISO/IEC 13751:2001 and is thus similar to APL2. It was initially written and is being maintained by Jürgen Sauermann. It includes extensions such as complex numbers and a shared variable interface.

In addition to a normal local APL session, GNU APL can used in a "scripting" language fashion by processing linear scripts passed to the interpreter at startup. It can also be compiled into a shared loadable library for embedding the interpreter into other projects.

Documents on GNU APL sometimes quote Richard Stallman, who both founded GNU and programmed in APL. However, Stallman is not directly associated with the project.

Encoding
GNU APL tends to prefer the Unicode Lozenge (U+25CA; ) and Element of (U+2208;  ) over Diamond operator (U+22C4;  ) and Small Element of (U+220A;  ) which are more commonly used by other dialects. This can lead to SYNTAX ERRORs if attempting to use code written for GNU APL in other implementations. However, GNU APL also accepts the more common code points.

Numeric types
GNU APL supports both 64-bit floats and 64-bit integers, as well as complex numbers whose components are floating-point. The integer type is used when a value can be conveniently determined to have an integer value, in particular when a primitive that is "integer by nature" is called on integer arguments and the result is not too large (GNU reserves a region at the boundary of the 64-bit range where results may be either integer or float). Monadic function  converts near-integer floating-point values to the integer type. Experimental support for rationals, which are ratios of 64-bit integers, can be enabled at compile time.