HLisp : Its Concept, Implementation and Applications

概要

論文の詳細を見る
The concept, implementation and applications of HLisp (Hash Lisp) are presented. HLisp is a dialect of Lisp which has two features of monocopy and associative computing. These features make HLisp a time and space efficient Lisp system. The monocopy feature is to make Lisp data objects uniquely, being free of duplicating copies. The monocopy of CONS data, symbols and associators is realized by using hash coding effectively. These data types and short integers which are represented by addresses of non existing array elements are called H-type in honor of hashing. The H-type data are utilized in speeding up the execution of equivalence check by equal and property fist handling by put , get, etc. The extended capability in assigning properties to H-type data objects is utilized in a feature called assoccomp. By using this feature the wasteful recomputation of the same function for the same argument value(s) is replaced with a search of such a property. Two features of rewrite protection and monocopy are utilized in the software virtualization of HLisp functions written in H-type data objects. HLisp was coded in FORTRAN for the sake of portablility. HLisp-REDUCE, REDUCE system based on HLisp, has been successfully implemented on various machines with different architectures and widely used for symbolic and algebraic manipulations on many machines.
一般社団法人情報処理学会の論文
1990-11-10