This describes work in progress on the semantics of Fifth, a declarative programming language which obtains efficiency without relying on domain-specific assumptions by means of an adaptive evaluation strategy. The adaptive evaluation strategy demands a consistent way of measuring the progress of a computation in terms of the information yielded by different control paths explored in the course of attempting to evaluate it. The Fifth system’s semantics are situated at the intersection of Icon, Prolog, Radul’s and Sussman’s information propagation networks and Codd’s relational model as developed by Date et. al. which in turn places it somewhere between abstract execution systems for program analysis and existing logic programming tools, and promises to enrich declarative programming with the semantics of relations on answer sets, and may through the consistent use of relations achieve the combination of expressivity and efficiency characteristic of the APL family and its consistent reliance on arrays.
[1]
Tony A. Plate,et al.
Holographic reduced representations
,
1995,
IEEE Trans. Neural Networks.
[2]
Oleg Kiselyov.
Delimited control in OCaml, abstractly and concretely
,
2012,
Theor. Comput. Sci..
[3]
Todd A. Proebsting,et al.
A new implementation of the Icon language
,
2000,
Softw. Pract. Exp..
[4]
Ivan Bratko,et al.
Prolog Programming for Artificial Intelligence
,
1986
.
[5]
Laslo Hunhold,et al.
The Unum Number Format: Mathematical Foundations, Implementation and Comparison to IEEE 754 Floating-Point Numbers
,
2017,
ArXiv.
[6]
C. J. Date,et al.
The third manifesto
,
1995,
SGMD.
[7]
Yunqian Ma,et al.
Manifold Learning Theory and Applications
,
2011
.
[8]
C. J. Date,et al.
Temporal data and the relational model
,
2002
.