V. F. Turchin. Program transformation with metasystem transitions. Journal of Functional Programming, 3(3):283--313, 1993.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....passes the KMP test in TSG and in MP. 11.1. Supercompiler: an Equivalence Transformation Modifier We outline an algorithm for supercompilation in TSG which will be our algorithm for equivalence transformation. A more detailed discussion of supercompilation and its methods can be found in [3, 26, 45, 51, 52, 53]) The supercompiler for TSG was implemented in Gofer (330 lines of pretty printed source text) Algorithm 25 [SCP (outline) Given TSG program pgm and class cls, the algorithm starts by driving the initial configuration pgm cls and builds a process tree gr (initially one node labeled with pgm ....

....language extensions (e.g. transforming programs written in a higher order language by a transformer for a first order language [46] While substantial practical evidence exists for equivalence transformation modifiers, not much work has been done regarding inversion modifiers. Experiments [40, 53] show that specializing an interpreter for inverse computation can invert programs. Only a very small number of publications studies the problem of program inversion [27, 28, 41, 42, 54] The modifier projections suggest a new way of constructing program inverters, namely by transforming an ....

[Article contains additional citation context not shown here]

V. F. Turchin, "Program transformation with metasystem transitions," Journal of Functional Programming 3(3) (1993) 283--313.

....supercompiler using a partial evaluator. The transformation is guided by the need to binding time improve the interpreters. 1 Introduction Our aim is to study the interpretive approach to improve the transformation of source programs and to automatically generate stand alone transformers [Tur93, GJ94]. The essence of the interpretive approach is to insert an interpreter between a source program and a generic program specializer. As defined by the specializer projections one may generate stand alone specializers from interpreters by self application of a program specializer [Glu91, Glu94] We ....

....an Interpreter The essence of the interpretive approach is to insert an interpreter between a source program and a program specializer. It is a surprising fact that, given an appropriate interpreter, this method may drastically increase the power of the overall transformation, as demonstrated in [Tur93, GJ94]. The same approach works for other program transformers, such as optimizing compilers [Jr92] Let ff be a program specializer with two parameters: a source program and a list of static input values. The arguments of the source program are classified as either s (static, known) or d (dynamic, ....

[Article contains additional citation context not shown here]

Valentin F. Turchin. Program transformation with metasystem transitions. Journal of Functional Programming, 3(3):283--313, 1993.

.... supercompilation [29] for first order functional programs by combining them with a mechanical extraction of answers (cf. 1, 25] We know of two techniques for inverse computation in functional languages: the universal resolving algorithm (see [1, 4] and walk grammars for inverse interpretation [30, 23]. The universal resolving algorithm in this paper uses methods from supercompilation [29] in particular driving, and is based on perfect process trees [12] Connections between inverse computation and logic programming are discussed in [1, 4] partial deduction and driving were formally related ....

V. F. Turchin. Program transformation with metasystem transitions. Journal of Functional Programming, 3(3):283--313, 1993.

.... [12, 11] It was shown that program specialisation combined with abstract interpretation can vastly improve the power of both techniques (e.g. going beyond regular approximations or set based analysis) 12] Interpreters Language independence can be achieved through the interpretive approach [18, 7, 1]: an interpreter serves as mediator between a (domain specific) language and the language for which the program transformer is defined. E#cient implementations can be obtained by removing the interpretive overhead using program specialisation (a notable example are the Futamura projections) Work ....

V.F. Turchin. Program transformation with metasystem transitions. Journal of Functional Programming, 3(3):283--313, 1993.

....in this case. 6 Related Work The ideas present in this paper have been heavily influenced by three concepts present in Turchin s work [22,23] metacoding, metavariables, and metasystem transition. Subsequently, these concepts have been formalized [7] and studied in different contexts, e.g. [25]. The correct treatment of metacode was found essential in self application [6] and this concept was singled out as elevation index [27] The problem of self application was the driving force behind the work on partial evaluation in the early eighties. The offline approach was originally ....

Valentin F. Turchin. Program transformation with metasystem transitions. Journal of Functional Programming, 3(3):283--313, 1993.

.... driving and unfolding in logic programming can be found in [25] More similarities between supercompilation and other techniques used for transforming logic programs are based on the idea of performing symbolic computations by meta programs or, in Turchin s terminology, metasystem transitions [24, 54]. For instance, the transformation technique presented by Gallagher in [20] works by specializing a meta interpreter, that is, a logic program which works as an interpreter for logic programs, w.r.t. a particular input program. Also the symbolic trace tree for compiling control may be generated ....

....construction of the symbolic computation model, the generalization steps can be suggested by an analysis of the part of the computation model already constructed. This analysis makes use of so called walk grammars and metatransition systems to reason about computation histories in a given model [54]. The analysis may be used, in particular, for avoiding the risk of generalizing too early (see, for instance, 53] Related works are the ones concerned with program improvement based on the analysis of computation histories for which the reader may refer to [1] where a recursion removal ....

[Article contains additional citation context not shown here]

V. F. Turchin. Program transformation with metasystem transitions. Journal of Functional Programming, 3(3):283--313, 1993.

....81] and the main results concerning selfapplication, metasystem transition, in 1973. The book [96] defined all three Futamura projections in terms of metasystem transition. In the English language, the work on supercompilation was first described in [83, 84, 85, 86] and then developed further in [87, 89, 93, 99]. For more historical details, see Turchin s personal account [95] Despite these remarkable contributions, supercompilation has not found recognition outside a small circle of experts. This paper gives a gentle introduction to the principles of supercompilation in terms of a positive ....

....1972 by performing subtraction by inverse computation of binary addition [80] In 1973 S.A. Romanenko and later S.M. Abramov implemented an algorithm, Universal Resolving Algorithm (URA) in which driving was combined with a mechanical extraction of answers [1, 65] For program inversion see also [39, 41, 66, 93, 60]. In logic programming, one defines a predicate by a program P x y and solves the inversion problem for Z = True . Theorem proving and program transformation are indistinguishable in the approach outlined above; they are two applications of the same equivalence transformation. The definition ....

V.F. Turchin. Program transformation with metasystem transitions. Journal of Functional Programming, 3(3):283--313, 1993.

....eighties [22,24] From its very inception, supercompilation has been tied to a specific programming language, called Refal [24] Applications of supercom pilation include, among others, program specialization, program inversion and theorem proving. Other related aspects have been investigated in [1,7,8,13,14,17,18,26]. The notion of perfect process graphs and perfect driving were introduced in [22,23] The language S Graph is closely related to Turchin s Refal graphs [25] But due to SGraph s simpler data structure, untyped variables and only two elementary contractions, one may build rather clear and concise ....

Turchin V. F., Program transformation with metasystem transitions. In: Journal of Functional Programming , 3(3): 283-313, 1993.

....the Petri net is not 1 safe: unsafe(s(0) s(0) Observe that none of the above queries (universally) terminate under Prolog or SLDresolution. 6. 5 Inductive Theorem Proving The relation between program specialisation and theorem proving has already been raised several times in the literature [53, 12, 54]. In this subsection we show how ecce can be used to perform some basic inductive theorem proving. Take the following simple program. even(0) even(s(X) odd(X) odd(s(X) even(X) By specialising this program with ecce, combined with a more specific program construction phase [38] another, ....

V. F. Turchin. Program transformation with metasystem transitions. Journal of Functional Programming, 3(3):283--313, 1993.

....as terms) 59] has excellent mathematical properties and allows arbitrarily high reflective towers. The Refal supercompiler a program transformer also written in Refal [56] can compile away extra levels of interpretation and can support very powerful meta programming and compilation tasks [58]. The metacircular Refal interpreter is virtually a universal theory; however, given that in Refal the computational interpretation is favored over the logical one, and that certain language conventions for example, the order in which the equations are applied prevent a semantic account given ....

Valentin F. Turchin. Program transformation with metasystem transitions. Journal of Functional Programming, 3(3):283--313, 1993.

....Table 2. The reason is that the substitutions in [3, 18] are actually accumulating parameters which are first fully generated before they can be consumed Deforestation of accumulators is still an open research problem (for functional languages, first, not yet automatic, approaches can be found in [29]) Let us adapt Ex. 2.4 into the rev last example: Example 6.1 (reverse last) revlast(L,X) reverse(L, a] R) last(R,X) reverse( L,L) reverse( H T] Acc,Res) reverse(T, H Acc] Res) last( X] X) last( H T] X) last(T,X) In the above program reverse is written using an ....

V. Turchin. Program transformation with metasystem transitions. Journal of Functional Programming, 3(3):283--313, 1993.

....a principle which is similar to partial evaluation. In the present paper we use a supercompiler, which V.F. Turchin and we have described in [22] 23] The aim of our investigation has been to show, what deep changes ( w.r.t. run time ) in the programs can be achieved by supercompilation. In [21] V.F. Turchin presented a method to improve the transformational power of supercompilation without modifying the transformation system. We use this idea to show both the power of the method and abilities of our supercompiler. Our examples include a generation of both one step unfolding and ....

....has been to show, what deep changes ( w.r.t. run time ) of programs can be achieved by supercompilation when it terminates. We do not present here a general method of showing how our transformations improve efficiency. However, we show that our examples are clean in some natural sense. In [21] V.F. Turchin presented a method to improve the transformational power of supercompilation without modifying the transformation system. The main idea is to insert an interpreter between the object program and the transformation system. R. Gluck and J. J rgensen used this idea [8] for generating ....

[Article contains additional citation context not shown here]

Turchin V.F., Program Transformation with Metasystem Transitions, J. of Functional Programming, 3(3) 283-313, 1993.

....of constrained atoms: couples of the form c 2A consisting of a constraint c and an atom A. The richer possibilities conferred by the use of the constraints notably allows one to e.g. provide very precise control mechanisms, drive negative information (using the terminology of supercompilation [34, 35]) handle built ins (like =2; n= 2) much more precisely and even make use of type information or argument size relations. We elaborate on two of these motivations below. We will formally present the framework of constrained partial deduction and summarise correctness results in Section 3. 2.1 ....

.... one can provide a very refined treatment of the n= Prolog built in using the FT structure (this feature has actually been incorporated in the prototype of [22] The following example illustrates this, where a form of driving of negative information (using the terminology of supercompilation [34, 35]) is achieved by constrained partial deduction. Example 2.2 Take the following adaptation of the member program which only succeeds once. 1) member(X; X jT ] 2) member(X; Y jT ] Xn= Y; member(X; T ) Let us start specialisation with the goal member(X; a; Y; a] Using a determinate ....

[Article contains additional citation context not shown here]

V. F. Turchin. Program transformation with metasystem transitions. Journal of Functional Programming, 3(3):283--313, 1993.

....the point of view of partial evaluation. Trace terms seem to offer a refinement of these notions, allowing more flexible treatment and more precise control than previously. The idea of regular approximation of trace terms also appears to be related to more recent work by Turchin on walk grammars [Tur93, Tur96], though the exact connections have still to be studied. Trace terms may offer a uniform treatment of all these ideas. 2 Representing Computation Paths In this section we introduce a representation of (successful) computation paths. The idea is very similar in logic and functional programming, ....

V. Turchin. Program transformation with metasystem transitions. Journal of Functional Programming, 3(3):283--313, 1993.

....al. JGS93] presents the typical PE framework and addresses many pragmatic concerns for a wide variety of language constructs. The presentation of a taxonomy of program transformers given by Gluck and S rensen in [GS96a] is also relevant for concurrent programs. An interesting paper of Turchin [Tur93] is probably the only reference to the study of the PE of languages with nondeterministic constructs. It gives an example of a transformation that has as an intermediate result a program that is apparently nondeterministic; however, the nondeterminism can be further specialized and the final ....

V.F. Turchin. Program transformations with metasystem transition. Journal of Functional Programming, 3(3):283--313, 1993.

.... partial evaluator Pell Mell [70] Birkedal and Welinder created a generator of generating extensions [17, 18] Refal and supercompilation: Turchin created the Refal language and developed the program transformation techniques of driving and supercompilation, which generalize partial evaluation [95, 96, 97]. A number of recent surveys on driving and supercompilation exist [48, 89, 90, 91] Prolog partial evaluation was pioneered by Komorowski [64, 65] subsequent work on Prolog includes [13, 44, 45, 66, 93, 98, 99] Sahlin constructed a practical but non self applicable partial evaluator for full ....

V.F. Turchin. Program transformation with metasystem transitions. Journal of Functional Programming, 3(3):283--313, July 1993.

No context found.

V. F. Turchin. Program transformation with metasystem transitions. Journal of Functional Programming, 3(3):283--313, 1993.

No context found.

V. F. Turchin. Program transformation with metasystem transitions. Journal of Functional Programming, 3(3):283-313, 1993.

No context found.

V.F. Turchin. Program transformation with metasystem transitions. Journal of Functional Programming 3(3):283--313 (1993).

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC