Xavier Leroy. Typage polymorphe d'un langage algorithmique. These de doctorat, Universite Paris 7, 1992.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....simply typed terms, in particular terms without type variables. This means that our machine can not be used as basis for the implementation of a type polymorphic language. It is well known that typed functional programming without type polymorphism is not expressive enough to be of interest. See (Leroy 1992) for a 5 (M N [e] s) M [e] h N ; ei s) App) n m 1 M [e] h N n 1 ; fi s) m M [h N n 1 ; fi e] s) Lam) n[h N n m 0 ; ei f ] s) Nm [e] s) Access) n m 1[h N m 0 ; ei f ] s) n[f ] s) Skip) Fig. 3. The sc machine. ....

Leroy, X. (1992). Typage polymorphe d'un langage algorithmique. PhD thesis. Universite de Paris 7.

....presented here is not the type assignment algorithm itself but the study of the structure of pairs inferred by this algorithm, which is developed in the further sections. The type inference algorithm is presented in figure 3. For clarity, we do not formalize the notion of new type variable (see [16] for a precise definition) This algorithm is defined modulo the name of type variables, since we do not fix the choice of the new type variables. Moreover, one can notice that there is no notion of bound type variables. Remark: Two non overlapped calls to Infer give disjoint types and disjoint ....

Xavier Leroy. Typage polymorphe d'un langage algorithmique. PhD thesis, University of Paris 7, June 1992. 36

.... class include the p calculus [10] CML [15] FACILE [5] and Plain CHOCS [17] In these languages, channels are typically typed like locations in a store; processes are merely checked in an environment assigning types to channels and other variables, and are given a trivial process type (see e.g. [7]) Alternatively, 12] annotate functional types with behavioral informations from which various static analyses, including typing, can be performed. Calculi and languages in the latter class include CHOCS [16] TPL [11] and LCS [3] The channels as locations analogy would give here a poor ....

....type . a: p (assumed closed) is the type of processes that may communicate massages of type on label a. For other labels than a, the constraints are given by type p. The type inference rules for CCS expressions are shown in Table 3.1. A maps message and process identifiers to types. As in [7], Gen(A,p) is the type scheme obtained by universally quantifying the variables that are free in p but not in A. Input bound variables here are typed like lambda bound variables in [4] or [7] Table 3.1. Type inference rules for CCS let A P: p A(x) A(X) p p p ....

[Article contains additional citation context not shown here]

X. Leroy, Typage Polymorphe d'un Langage Algorithmique, PhD Thesis, Univ. of Paris VII, June 1992.

....call by value languages include imperative features like references, exceptions, and continuations that facilitate concise and efficient programs. Several solutions to integrating imperative features with HindleyMilner polymorphism in call by value languages have been devised [1] 3] 4] 8] [9], 11] 17] 18] 19] These solutions range in complexity from Tofte s relatively simple method that Standard ML adopted to Talpin s sophisticated system that infers types, regions, and effects. All of these solutions assign types to all purely functional expressions that are typable by the ....

....was supported in part by the United States Department of Defense under a National Defense Science and Engineering Graduate Fellowship. 2 more concise. But such an imperative version of sort has the following imperative types in Tofte s system [18] MacQueen s system [1] and Leroy s system [9], 11] 8 ff: ff ff bool) ff list) ff list) Tofte) 8ff 2 : ff 2 ff 2 bool ) ff 2 list) ff 2 list) MacQueen) 8ffLMNP: ff L ff M bool) N (ff list) P (ff list) Leroy) with fff . M; ff L ff M bool) Pg In each case, the extra ....

[Article contains additional citation context not shown here]

Leroy, X. Typage polymorphe d'un langage algorithmique, PhD thesis, L'Universite Paris 7 (1992).

....both evaluators define the same observational equivalence relation via the construction of a number of intermediate calculi and machines. It is far more complicated than our diamond and bisimulation techniques, possibly due to the inclusion of continuations. Independently, Reppy [21] and Leroy [17] define a formal operational semantics for an ML like language with first class synchronization operations. Reppy s language, Concurrent ML, can provide the future mechanism as an abstraction over the given primitives. The semantics is a two level rewriting system. The first level, also a program ....

Leroy, X. Typage polymorphe d'un langage algorithmique. PhD thesis, Universit'e Paris 7, 1992.

....application in parallel) and first class continuations. His primary goal is to design a semantics for the language that treats pcall as a pure annotation. His correctness proofs is far more complicated than our techniques, due to the inclusion of continuations. Independently, Reppy [28] and Leroy [24] define a formal operational semantics for an ML like language with first class synchronization operations. Reppy s language, Concurrent ML, can provide the future mechanism as an abstraction over the given primitives. The semantics is a two level rewriting system. Reppy uses his semantics to ....

Leroy, X. Typage polymorphe d'un langage algorithmique. PhD thesis, Universit'e Paris 7, 1992.

....Semantics The static semantics is based on an extension of the type and effect discipline. This allows us to safely control the type generalization in the presence of polymorphic channels. Actually, types and effects are used to control the generalization of type and effect variables. As shown in [Ler92, TJ92], effectbased type systems outperform the other type systems when typing in the presence of mutable data. Furthermore, communication effects with principal types capture a valuable static information on the dynamic behaviors of a program. The latter information may be of great interest in the ....

X. Leroy. Typage polymorphe d'un langage algorithmique. PhD thesis, Universit'e de Paris VII, June 1992.

.... e 1 causes side effects, type variables involved in these side effects must not be generalized in assigning a type to x [14] The existing solutions attempt to determine when e 1 will cause undesirable side effects by recording information about potential side effects in an expression s type [1, 3, 4, 6, 13, 14, 15]. For example, Standard ML assigns the type 8 ff fi: ff fi) ff list fi list to the following imperative version of map: fun imperativeMap f l = let val a = ref l and b = ref nil in while not (null ( a) do (b : f (hd ( a) b) a : tl ( a) rev ( b) end The imperative type ....

....according to whether the computation is purely functional or uses imperative features. Suppose the bound expression involves purely functional computation, like the following example: 1 Leroy s original closure typing system [6] did not type all such functional expressions. His dissertation [4] corrects this oversight. let val f = fn x = x) fn y = y) in f 1; f true end With pure Hindley Milner typing, this expression would be typable by assigning to f the polymorphic type 8ff: ff ff. But with polymorphism restricted to values, f is not polymorphic, and this expression is ....

[Article contains additional citation context not shown here]

Leroy, X. Typage polymorphe d'un langage algorithmique. PhD thesis, L'Universit'e Paris 7, 1992.

....well suited within a programming environment that supports separate compilation features and modular programming paradigms. 11.1 Comparative Examples This section present a series of more or less sophisticated examples, adapted from a survey paper on this subject [O Toole, 1990] and from [Leroy, 1992], that establishes frontiers between the related type systems. Note that the results of the system presented in [Jouvelot Gifford, 1991] are given for programs with explicit polymorphic types. define id1 (let ( x (id 1) rid) define id2 (lambda (y) rid id) y) define id3 ( nop rid) id) ....

.... (f) let ( id5 (lambda (y) let ( r (new y) if true (lambda (z) if true r (new y) z) f) y) id5 id5) In addition to this comparison, the example of recursive typing presented in section 8 (the function eta ref below requires such a recursive typing) is well typed using the system of [Leroy, 1992]. Note that it would have been in ours, if we had chosen to make indirections between types and effects explicit in the static semantics, using constraint sets, as is done in the inference algorithm. define eta ref (lambda (f) let (r (new f) if true f (lambda (x) get (if true r (new f) ....

[Article contains additional citation context not shown here]

Leroy, X. Typage polymorphe d'un langage algorithmique. Doctoral dissertation, Universit'e Paris VII, 1992.

....our type system did not include polymorphic let expressions, the preceding changes would suffice to type Box Scheme. But as many authors have noted, naively combining Hindley Milner polymorphism and assignment leads to an unsound type system [62] Many solutions to this problem have been proposed [6, 14, 22, 29, 35, 38, 57, 60, 61, 62]. We adapt our own solution [62] which is the simplest of all. This solution restricts polymorphism to syntactic values. That is, we replace Pure Scheme s static type inference rule for let expressions (Figure 3.3) with the following two inference rules. 62 A e : 1 A (box e) box 1 ) ....

Leroy, X. Typage polymorphe d'un langage algorithmique. PhD thesis, L'Universit'e Paris 7, 1992.

....about identi ers bound to polymorphic objects are violated after the binding has changed, either because a reference was updated, or because a continuation was restarted. This problem with polymorphic continuations can be avoided by typing restrictions similar to those for polymorphic references [21, 12] with the same drawbacks: the corresponding type systems are either too restrictive or too complicated. Again, polymorphism by name provides an alternate solution. With by name semantics for generalization, it is impossible to capture a continuation that generalizes the type of the value ....

....proof is elementary: it proceeds only by structural induction over the terms representing values and evaluation derivations. In particular, there is no need for proofs by co induction [19, 14] The proofs also easily extend to other polymorphic type systems for references and continuations [12]. 4 Assessment Polymorphism by name supports references and continuations in a type safe way, while retaining the ML type algebra and typing rules, that are familiar and easy to understand. This is a strong advantage over the restricted type systems proposed for references and continuations in ....

[Article contains additional citation context not shown here]

X. Leroy. Typage polymorphe d'un langage algorithmique. Doctoral dissertation (in French), Universit e Paris 7, 1992.

No context found.

Xavier Leroy. Typage polymorphe d'un langage algorithmique. These de doctorat, Universite Paris 7, 1992.

No context found.

Leroy, X. Typage polymorphe d'un langage algorithmique. Doctoral dissertation, Universit'e Paris VII, 1992.

No context found.

Xavier Leroy. Typage polymorphe d'un langage algorithmique. These de doctorat, Universit e Paris 7, 1992.

No context found.

X. Leroy. Typage Polymorphe d'un Langage Algorithmique. Thesis, University Paris 7, 1992.

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