J. Eifrig, S. Smith, and V. Trifonov. Sound polymorphic type inference for objects. In OOPSLA '95 Conference Proceedings, volume 30(10) of ACM SIGPLAN Notices, pages 169--184, 1995.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....approach which can be seen as a simpli ed representation of constraints. Another form of simpli cation would be to enforce the set of solved forms to be in syntactically unique form (see De nition 24 in Section 4. 2) Simpli cation strategies in the context of subtyping were studied in [Pot98, EST95a]. I can state the following results. I omit to give proofs of the soundness and completeness results. Both proofs are almost identical to the soundness and completeness results of the class X in case of inference in term form. If C; e : then C; e : For the completeness result, ....

J. Eifrig, S. Smith, and V. Trifonov. Sound polymorphic type inference for objects. In OOPSLA '95 Conference Proceedings, volume 30(10) of ACM SIGPLAN Notices, pages 169-184, 1995.

....languages, the type of an object reflects only its public interface; it cannot convey implementation information. More about Bruce et al. s work can be found in Section 9.6. Smith and the Hopkins Object Group have designed a type safe class based OO language with a rich feature set called I Loop, [EST95]. Their type system is based on polymorphic recursively constrained types, for which they have a sound type inferencing algorithm. The main advantage of this approach is the extreme flexibility afforded by recursively constrained types. Currently, the main problem is that it returns large, ....

J. Eifrig, S. Smith, and V. Trifonov. Sound polymorphic type inference for objects. In Object-Oriented Programming: Systems, Languages and Applications, pages 169--184. ACM, 1995.

....the system of recursive object types with subtyping can be solved in O(n ) Unfortunately, Henglein s method cannot be applied to our type system, as the key ingredient for breaking through the n barrier in his algorithm is the invariant rule for object subtyping. In a series of papers [7, 6, 23], Eifrig, Smith and Trifonov study the inference problem for a polymorphic type system that includes both functions and objects, and develop powerful simplification methods for the constraint sets generated during the inference. Some of these methods have independently been studied (and improved) ....

J. Eifrig, S. Smith, and V. Trifonov. Sound polymorphic type inference for objects. In Proceedings OOPSLA '95, ACM SIGPLAN Notices, pp. 169--184, 1995.

....that will be used at each declaration (in our terminology, origin) in the translated GJ program. The algorithm works by computing which origins flow to each type variable. As a simple example, consider the class class Box (sometimes 1: class Stack 2: 3: private Object[ data = new Object[10]; 4: private int size = 0; 5: Object top( 6: return data[size 1] 7: 8: Object pop( 9: return data[ size] 10: 11: void push(Object o) 12: data[size ] o; 13: 14: void exchange( 15: Object o1 = pop( o2 = pop( 16: push(o1) 17: push(o2) 18: 19: Figure ....

....Object[ O4 Stack.anewarray1 Object O5 Stack.top: return Object O6 Stack.pop: return Object O7 Stack.push: o Object Figure 2: Origins for the Stack example of Figure 1. In origin names, the : operator denotes in the scope of . Origins O1 and O2 represent the expression new Object[10] that appears in the body of the (implicit) Stack constructor. Origin number have no semantic meaning, but are only used for presentation in this paper. Figure 3: Lattice for the parameterisation analysis. Between NOVARIABLE and NULL is P( O1 . On ) the power set of the n origins, ordered ....

[Article contains additional citation context not shown here]

J. Eifrig, S. Smith, and V. Trifonov. Sound polymorphic type inference for objects. In Proceedings of the tenth annual conference on Object-oriented programming systems, languages, and applications, pages 169--184. ACM Press, 1995.

....of subtype # # . Subtype constraints are systems of inequations t#t # that talk about the subtype relation. Terms t and t # in subtype constraints are built from type variables and type constructors. Two logical operations on subtype constraints were investigated: satisfiability and entailment [6, 12, 1, 4, 18, 20]. Subtype satisfiability can be checked in cubic time for many type languages [9, 16] A quadratic time algorithm for the variable free case is presented in [10] Interest in subtype entailment was first raised by practical questions on type inference engines with subtyping [5, 21, 19] The ....

J. Eifrig, S. Smith, and V. Trifonow. Sound polymorphic type inference for objects. In ACM Conference on Object-Oriented Programming: Systems, Languages, and Applications, 1995.

....that the presented approach can be extended to the general case. Keywords: subtyping, constraints, entailment, automata. 1 Introduction Subtyping is a natural concept in programming. This observation has motivated the design of programming languages featuring a system for subtype inference [8, 11, 2, 6, 18]. Simplification of typings turned out to be the key issue in what concerns the complexity of subtype inference systems [7, 19, 17] Several authors proposed to simplify typings based on algorithms for subtype entailment, i.e. entailment of subtype constraints [22, 16] First approaches towards ....

J. Eifrig, S. Smith, and V. Trifonow. Sound polymorphic type inference for objects. In ACM Conference on Object-Oriented Programming: Systems, Languages, and Applications, 1995.

....variance of the mytype type variable. In these languages, the type of an object reflects only its public interface; it cannot convey implementation information. Smith and the Hopkins Object Group have designed a type safe class based object oriented language with a rich feature set called I Loop,[15]. Their type system is based on polymorphic recursively constrained types, for which they have a sound type inferencing algorithm. The main advantage of this approach is the extreme flexibility a#orded by recursively constrained types. Currently, the main problem is that it returns large, ....

J. Eifrig, S. Smith, and V. Trifonov. Sound polymorphic type inference for objects. In ACM Conf. Object-Oriented Programming: Systems, Languages and Applications, pages 169--184, October 1995. 483

....simple input expressions, appears excessively long and cumbersome to read . The problem has generated a significant amount of work which aims at simplifying constraints in the typings generated by subtype inference algorithms; works addressing the subtype simplification problem in clude [28, 17, 41, 67, 21, 59, 73, 22, 4, 25]. As is argued in [4] simplification is beneficial for at least three reasons: first, it may speed up type inference, second, it makes types more readable, and, third, it makes the information content of a typing more explicit. The point about the speed of type inference comes out most ....

....Ker(C) is the set of all inequalities entailed by C that are not logically valid in P and which have all their variables constrained in C. 1The reader should be awaxe that examples can be scaled up to become much more complex. For more examples consult works on subtype simplification, such as [28, 17, 41, 67, 21, 37, 59, 73, 4]. 2In [73] a notion of constraint kernels is also defined, but it is not the same as ours. Notice that, since P is finite, Ker(C) is always a finite set of inequalities. Any constraint set C is logically equivalent to its kernel: Lemma 3.2.2 C p Ker(C) PROOF It is obvious that we have C p ....

J. Eifrig, S. Smith, and V. Trifonov. Sound polymorphic type inference for objects. In Proceedings OOPSLA '95, 1995.

....of these algorithms. Trifonov and Smith [43] present constrained types that are similar to our constraint systems, and they describe an incomplete algorithm for deciding the subtyping relation between constrained types. They do not discuss constraint simplification. Eifrig, Smith and Trifonov [11] discuss constraint simplification in the context of type inference for objects. They present three algorithms for simplifying constraint systems, Personal communication: Daniel Weise (February 97) 95 two of which which are similar to the empty and ffl removal algorithms, and the third is a ....

Eifrig, J., Smith, S., and Trifonov, V. Sound polymorphic type inference for objects. In Conference on Object-Oriented Programming Systems, Languages, and Applications (1995).

....and others have given inference algorithms for idealised subtyped languages [11] 10] his language does not cover all the complexities of MLj. This work is extended by [7] and [1] Rather than looking at general subtyped languages, Eiefig, Smith and Trifonov look at type inference for objects in [5]. Constraints are also used in program analysis. Nielson, Nielson and Hankin describe this in [12] Francois Pottier s work [13] is closest to our own, in particular in the manner he treats upper and lower bounds. In particular our algorithm differs from previous work as it requires branching ....

J. Eifrig, S. Smith, and V. Trifonov. Sound Polymorphic Type Inference for Objects. In OOPSLA, 1995. 12

....inference algorithms for idealised subtyped languages [Mit96, Mit91] His language does not include all the complexities of MLj. This work is extended by [JM93] and [AW93] Rather than looking at general subtyped languages, Eiefig, Smith and Trifonov look at type inference for objects in [EST95] Constraints are also used in program analysis. Nielson, Nielson and Hankin describe this in [NNH98] Francois Pottier s work [Pot96] is closest to MLj s, in particular in the manner he treats upper and lower bounds. Duggan s work on type explanations is also based on constraint solving (though ....

J. Eifrig, S. Smith, and V. Trifonov. Sound Polymorphic Type Inference for Objects. In Conference on Object-Oriented Programming, Systems, Languages and Applications (OOPSLA). ACM Press, 1995.

....remarks are given in Section 5. Proofs are omitted from this paper due to space limitations. The full paper will be available as [23] 1. 1 Related Work Constraint based type inference systems have been designed for many languages, including ones that support subtyping [6, 11] objects [5], overloading [8, 22] and linear types [21] The inference issues in [21] are especially close to ours, and we received considerable inspiration from this work. One novel aspect of our system which does not arise in these other works is the need to associate constraints with each free identifier ....

J. Eifrig, S. Smith, and V. Trifonov. Sound polymorphic type inference for objects. In Object Oriented Prog. Syst., Lang. and Applications, pages 169--184. ACM, 1995.

....application to type inference for functional languages was proposed by Aiken et al. in [AW92, AW93, AWL94] This work has been extended by Aiken et al. in the BANE project [AFFS98] in order to provide a generic framework for developping ecient set based analysis. As advocated by Smith et al. in [EST95a, EST95b, TS96] set based type inference can be applied to sequential object oriented languages. The main purpose of this work is to extend set based typing system to concurrent object oriented languages. The aim of these type systems is the static detection of orphan messages in actor ....

J. Eifrig, S. Smith, and V. Trifonov. Sound polymorphic type inference for objects. In Proc. of the OOPSLA, 1995.

....may grow exponentially in the number of required iterations. We can drastically reduce the number of constraints if we can simplify the constraints to fewer but equivalent constraints. It is therefore no surprise that lots of work has been put into techniques for how to simplify constraints [FM89,Cur90,Kae92,Smi94,EST95,Pot96,TS96,FA96,AWP97,Reh97,FF97]. Another approach is to make the constraint language more powerful so that constraints can be generated by a simple linear time traversal over the program. This can be achieved by making substitution instantiation a syntactic construct in the constraint language. But when we make the constraint ....

....a b is in the constraint becomes just set membership in the appropriate set. 4 Related Work The motivation for this paper is to reduce the cost of the combination of subtyping and polymorphism and in this respect it is related to numerous papers on constraint simplification techniques [FM89,Cur90,Kae92,Smi94,EST95,Pot96,TS96,FA96,AWP97,Reh97,FF97]. Our work is particularly related to the work by Dussart, Henglein and Mossin on binding time analysis with binding time polymorphic recursion [DHM95] where they use constraint simplification techniques in combination with a clever fixedpoint iteration to obtain a polynomial time algorithm. In ....

J. Eifrig, S. Smith, and V. Trifonov. Sound polymorphic type inference for objects. In Proceedings OOPSLA'95, 1995.

....of these algorithms. Trifonov and Smith [43] present constrained types that are similar to our constraint systems, and they describe an incomplete algorithm for deciding the subtyping relation between constrained types. They do not discuss constraint simplification. Eifrig, Smith and Trifonov [11] discuss constraint simplification in the context of type inference for objects. They present three algorithms for simplifying constraint systems, Personal communication: Daniel Weise (February 97) 95 two of which which are similar to the empty and ffl removal algorithms, and the third is a ....

Eifrig, J., Smith, S., and Trifonov, V. Sound polymorphic type inference for objects. In Conference on Object-Oriented Programming Systems, Languages, and Applications (1995).

....to call a myClass parameter. This model is operationally quite expressive, but the type theory that it uses is also quite elaborate, and does not full our desires, of reconstruction of a principal type. The same remark actually applies to all the object models that use higher order types, e.g. [2, 9, 18, 20, 39]. In another approach, due to Kamin [28] and known as the self application semantics, an object is a record of pre methods, that are functions of the object itself. The object is bound to self only when a method is invoked, by applying the pre method to the object. In this way, the state of the ....

....behaviour of an object, to update the state (see [2] Section 5.2) As we have seen, the rst usage is not problematic in Wand s model, whereas the second is. Then an obvious idea is to abandon the ifunctional updatej approach in favor of a rather more natural imperative update approach, as in [2, 4, 18, 44]. This means that we are in a language with references (following ML s terminology) where a call by value strategy is assumed for evaluation. Now a new problem arises: to build objects as recursive records one must have the ability to build recursive non functional values, and this, in principle, ....

[Article contains additional citation context not shown here]

J. Eifrig, S. Smith, V. Trifonov, Sound polymorphic type inference for objects, OOPSLA '95, ACM SIGPLAN Notices Vol. 30 No. 10 (1995) 169-184.

....P e : # succeeds, fv(#) #, and # #P # # # . Detailed proofs of this and other results stated in this section can be found in the author s dissertation [Nor99] 28 7 Related work Algorithms for polymorphic subtype inference are described in [Mit84, FM90, FM89, Mit91, Kae92, Smi94, AW93, EST95, FA96, Hen96, Pot96, TS96, Reh97, MW97, Seq98, Pot98, MOW99, Pot00] Strategies for simplification of the inferred constraint sets are discussed in [FM89, AW93, EST95, Reh97, MW97, Pot98, Pot00] in particular. The choice to settle for a deliberately incomplete inference algorithm sets our work ....

.... 28 7 Related work Algorithms for polymorphic subtype inference are described in [Mit84, FM90, FM89, Mit91, Kae92, Smi94, AW93, EST95, FA96, Hen96, Pot96, TS96, Reh97, MW97, Seq98, Pot98, MOW99, Pot00] Strategies for simplification of the inferred constraint sets are discussed in [FM89, AW93, EST95, Reh97, MW97, Pot98, Pot00] in particular. The choice to settle for a deliberately incomplete inference algorithm sets our work apart from most of the cited systems, though. Smith and Trifonov [TS96] and Pottier [Pot96, Pot98] approximate entailment checking with decidable variants that result ....

[Article contains additional citation context not shown here]

J. Eifrig, S. Smith, and V. Trifonov. Sound Polymorphic Type Inference for Objects. In OOPSLA '95. ACM, 1995.

....in class encodings. In our calculus, instead of encapsulation at the object type level, we use subtyping to hide protected methods and binding to hide private elds. The Hopkins Object Group has designed a type safe class based objectoriented language with a rich feature set called I Loop [18]. Their type system is based on polymorphic recursively constrained types, for which they have a sound type inferencing algorithm. The main advantage of this approach is the extreme exibility a orded by recursively constrained types. However, inferred types are large and dicult to read. Bruce et ....

Eifrig, J., S. Smith and V. Trifonov, Sound polymorphic type inference for objects, in: Proc. OOPSLA '95, 1995, pp. 169-184.

....site has been identi ed as a major problem, making it very dicult to scale polymorphic subtyping to large programs. 2 In particular, the problem has generated a signi cant amount of research on constraint simpli cation, which aims at compacting constraint sets before they are copied [FM89, Cur90, Kae92, Smi94, EST95, Mos96, Pot96, TS96, FA96, AWP97, Reh97, FF97]. It is unlikely that constraint simpli cation techniques alone will solve this problem, and complete simpli cation is a hard problem itself [Reh98, FF97] 2.2 Problem 2: Demand driven ow computation Constraint copying methods for polymorphic subtyping systems are not demand driven [Mos96] For ....

J. Eifrig, S. Smith, and V. Trifonov. Sound polymorphic type inference for objects. In Proceedings OOPSLA '95, 1995.

....Science, The Johns Hopkins University, Baltimore, Maryland 21218; email: scott cs.jhu.edu. 1 When combined with universal quantification, such types enable efficient type inference for object oriented languages with polymorphism and subtyping, as demonstrated by Eifrig, Smith, and Trifonov [5, 4]. Until now, it has been unclear how expressive constrained types are. In this paper we characterize constrained types without universal quantification, that is, types of the form t n C where t is a simple type and C is a constraint system. Our example language is a calculus generated by the ....

J. Eifrig, S. Smith, and V. Trifonov. Sound polymorphic type inference for objects. In Proc. OOPSLA'95, ACM SIGPLAN Tenth Annual Conference on Object-Oriented Programming Systems, Languages and Applications, 1995.

No context found.

J. Eifrig, S. Smith, and V. Trifonov. Sound polymorphic type inference for objects. In OOPSLA '95 Conference Proceedings, volume 30(10) of ACM SIGPLAN Notices, pages 169--184, 1995.

No context found.

J. Eifrig, S. Smith, and V. Trifonov. Sound polymorphic type inference for objects. In OOPSLA, pages 169--184, Oct. 1995.

No context found.

J. Eifrig, S. Smith, and V. Trifonov. Sound polymorphic type inference for objects. In ACM Conference on Object-Oriented Programming: Systems, Languages, and Applications, pages 169--184. ACM Press, 1995.

No context found.

Eifrig, J., Smith, S., Trifonov, V.: Sound polymorphic type inference for objects. ACM SIGPLAN Notices 30 (1995)

No context found.

J. Eifrig, S. Smith, V. Trifonov, Sound polymorphic type inference for objects, in: ACM Conference on Object-Oriented Programming: Systems, Languages, and Applications, ACM Press, 1995, pp. 169--184.

First 50 documents  Next 50

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