Pierce, B. and Turner, D. Local Type Inference. in: Proceedings of the 25th Annual ACM SIGPLANSIGACT Symposium on Principles of Programming Languages, San Diego, January 19--21. 1998.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....If #, Y M : # and # # then #, # Y ]# # M : # Y ]# . Proof. By induction of the typing derivation and case distinction on the subtyping derivation. 4. 3 Type Checking Types of Mini ML expressions can be checked using a algorithm with uses only local type inference [30]. The idea is to avoid sophisticated concepts like unification variables or constraints in order to make type checking more predictable. We use the well known technique of bidirectional type checking (e.g. 13, 15] The rules are listed in Table 4. The rough idea is that the type of variables ....

B. C. Pierce and D. N. Turner. Local type inference. In POPL 98: The 25TH San Diego, California, 1998.

....initiated by Cardelli [2] is to start with an expressive but explicitly typed language, say F : and perform a sufficient amount of type inference, so that simple programs ideally including all ML programs would not need any type annotation at all. This lead to local type inference [24], recently improved to colored local type inference [21] These solutions are quite impressive. In particular, they include subtyping in combination with higher order polymorphism. However, they fail to type all ML programs. Moreover, they also fail to provide an intuitive and simple specification ....

....second order terms is unification under a mixed prefix [18] However, our notion of prefix and its role in abstracting polytypes is quite different. Actually, none of the above works did consider subtyping at all. This is a significant difference with proposals based on local type inference [2, 24, 21] where subtyping is a prerequisite. The addition of subtyping to our framework remains to be explored. Furthermore, beyond its treatment of subtyping, local type inference also brings the idea that explicit type annotations can be propagated up and down the source tree according to fixed ....

B. C. Pierce and D. N. Turner. Local type inference. In Proceedings of the 25th ACM Conference on Principles of Programming Languages, 1998.

....initiated by Cardelli [Car93] is to start with an expressive but explicitly typed language, say F : and perform a sufficient amount of type inference, so that simple programs ideally including all ML programs would not need any type annotation at all. This lead to local type inference [PT98] recently improved to colored local type inference [OZZ01] These solutions are quite impressive. In particular, they include subtyping in combination with higher order polymorphism. However, they fail to type all ML programs. Moreover, they also fail to provide an intuitive and simple ....

....terms is unification under a mixed prefix [Mil92] However, the notion of prefix in their work is rather different and does not express sharing . Actually, none of the above works did consider subtyping at all. This is a significant difference with proposals based on local type inference [Car93, PT98, OZZ01] where subtyping is a prerequisite. The addition of subtyping to our framework remains to be explored. Furthermore, beyond its treatment of subtyping, local type inference also brings the idea that explicit type annotations can be propagated up and down the source tree according to fixed ....

Benjamin C. Pierce and David N. Turner. Local type inference. In Proceedings of the 25th ACM Conference on Principles of Programming Languages, 1998. Full version in ACM Transactions on Programming Languages and Systems (TOPLAS), 22(1), January 2000, pp. 1--44.

....allows certain type annotations to be elided in actual programs. Local type inference determines types by a combination of type propagation and local constraint solving, rather than by global constraint solving. We refine the previously existing local type inference system of Pierce and Turner[PT98] by allowing partial type information to be propagated. This is expressed by coloring types to indicate propagation directions. Propagating partial type information allows us to omit type annotations for the visitor pattern, the analogue of pattern matching in languages without sum types. 1 ....

....type annotations for the visitor pattern, the analogue of pattern matching in languages without sum types. 1 Introduction Many modern programming languages are based on type systems which combine a notion of objects and subtyping with parametric polymorphism [Str91, Mey92, CDG 92, NC97, OW97, PT98, BOSW98] A popular basis for such type systems is F# , the second order lambda calculus with subtyping. While F# is an excellent basis for explaining the abstract type structure of programs, it is less suitable as a kernel language for concrete source programs, because of the excessive amount of ....

[Article contains additional citation context not shown here]

Benjamin C. Pierce and David N. Turner. Local type inference. In Proc. 25th ACM Symposium on Principles of Programming Languages, pages 252--265, January 1998. 12

....allows certain type annotations to be elided in actual programs. Local type inference determines types by a combination of type propagation and local constraint solving, rather than by global constraint solving. We refine the previously existing local type inference system of Pierce and Turner[PT98] by allowing partial type information to be propagated. This is expressed by coloring types to indicate propagation directions. Propagating partial type information allows us to omit type annotations for the visitor pattern, the analogue of pattern matching in languages without sum types. 1 ....

....type annotations for the visitor pattern, the analogue of pattern matching in languages without sum types. 1 Introduction Many modern programming languages are based on type systems which combine a notion of objects and subtyping with parametric polymorphism [Str91, Mey92, CDG 92, NC97, OW97, PT98, BOSW98] A popular basis for such type systems is F# , the second order lambda calculus with subtyping. While F# is an excellent basis for explaining the abstract type structure of programs, it is less suitable as a kernel language for concrete source programs, because of the excessive amount of ....

[Article contains additional citation context not shown here]

Benjamin C. Pierce and David N. Turner. Local type inference. In Proc. 25th ACM Symposium on Principles of Programming Languages, pages 252--265, January 1998.

....holds for our algorithm as well, and for that reason we have good hope that there actually exists a simple declarative speci cation for it. Finding such a speci cation is an area of ongoing research. It is also instructive to compare with Pierce Turner s work on local type inference here [PT98]. They have an algorithm that has a relatively simple declarative speci cation, although it is not a conservative extension of ML style type inference. Still, it is clear from the examples the authors provide that their algorithm sometimes will exhibit a quite surprising behaviour, simply 13 ....

Benjamin C. Pierce and David N. Turner. Local type inference. In Conference Record of POPL '98: the 25th ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages, 1998. 14

....definitions, the set SF had to be restricted since applications with recursively defined heads are potentially instable with respect to substitution at recursive parameter positions. The local erasure algorithm in this paper hides only arguments that can be elaborated by local methods [PT98], while our elaboration algorithm allows the global distribution of unification variables. Consider the polymorphic operation append on lists of type #T Prop . List T ) # (List T ) # (List T ) and a list l of type (List IN) The implicit term (append nil l) is elaborated into the explicit ....

B. Pierce, D. Turner. Local type inference. In Conf. Record of POPL'98, pp. 252--265. ACM Press, 1998.

....type inference [8] which requires no type annotations at all in programs. A step in this direction for a purely linear type system has been undertaken in [30] We plan to investigate to what extent these results can be extended to the present case. Alternatively, we might follow Pierce and Turner [28] who have argued convincingly that type annotations in definitions might be acceptable and even desirable as they provide some documentation. Based on these extra assumptions they were able to give a much simplified inference algorithm for ML. Another pragmatic aspect is that at present SLR is ....

Benjamin C. Pierce and David N. Turner. Local type inference. In Conference Record of POPL '98: the 25th ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages, 1998. Full version available as Indiana University CSCI technical report #493.

....it, which is one of our principal goals. Beyond this, there are many possible areas to look at. We hope to study the interaction of session types and polymorphism with a view to reintroducing polymorphism to the modi ed compiler. The Pict compiler uses a powerful partial type inference technique [11], and it will be interesting to see how it can be extended to handle session types. Because of the value of explicit session types as speci cations, we might not want to allow the programmer to omit them completely. However, automatic inference of, for example, polarities will probably be very ....

B. C. Pierce and D. N. Turner. Local type inference. In Proceedings, 25th ACM Symposium on Principles of Programming Languages, 1998. REFERENCES 26

....Our extension of Odersky and L aufer s system can be seen as a natural generalisation of the existing type inference algorithm to arbitrary ranked polymorphism. Another possibility would have been to abandon Hindley Damas Milnerstyle type inference in preference for local type inference [15, 14]. However, we felt that would have been too great a change for Haskell. Thirdly, we examined Russo s semantics for ML signatures and structures [17] in order to understand how the dot notation of ML modules interacts with ordinary existential and universal polymorphism. As a result, we re ned ....

B. C. Pierce and D. N. Turner. Local type inference. In Proceedings of the 25th Annual ACM SIGPLANSIGACT Symposium on Principles of Programming Languages, San Diego, California, pages 252-265. ACM Press, Jan. 1998.

....for their object oriented ML variant OML, in such a way that only constraint free type schemes are inferred [RR96] However, they do not describe their algorithm, nor do they provide any technical results. Most related to our approach is Pierce and Turner s work on local type inference for Pict [PT98] They start with a variant of (the explicitly typed) System F# , and develop an inference technique that is able to fill in missing type arguments, as well as missing annotations, in many cases. Pierce and Turner s algorithm is not complete w.r.t. the Hindley Milner system, but they have the ....

....holds for our algorithm as well, and for that reason we have good hope that there actually exists a simple declarative specification for it. Finding such a specification is an area of ongoing research. A comparison with Pierce Turner s work on local type inference is instructive here as well [PT98] Their algorithm has a relatively simple declarative specification, although it is not a conservative extension of ML style type inference. Still, it is clear from the examples the authors provide that their algorithm sometimes will exhibit a quite surprising behavior, simply because it fails to ....

Benjamin C. Pierce and David N. Turner. Local type inference. In Conference Record of POPL '98: the 25th ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages, 1998.

....which high level programming idioms are most useful in practice. Because types and type checking are so important for realistic programming, Pict is a typed language, and has become a vehicle for experimenting not only with type systems but also with type checking and type inference algorithms [28]. Pict starts with Pierce and Sangiorgi s input output types with subtyping, and adds record types and some basic data types such as int and bool. Polymorphism has been included, and indeed extended to a system of higher order polymorphism. The linear types of Kobayashi et al. have been added as ....

B. C. Pierce and D. N. Turner. Local type inference. In Proceedings, 25th ACM Symposium on Principles of Programming Languages, 1998.

....in the context of the Pict programming language. Our next goal is to extend the Pict compiler to support the type system presented here, after first removing some of Pict s existing features such as higher order polymorphism. The Pict compiler uses a powerful partial type inference technique [12], and it will be interesting to see how it can be extended to handle session types. Because of the value of explicit session types as specifications, we might not want to allow the programmer to omit them completely; however, automatic inference of, for example, some usage annotations will ....

B. C. Pierce and D. N. Turner. Local type inference. In Proceedings, 25th ACM Symposium on Principles of Programming Languages, 1998.

....and present a new notation since we wanted to support both functions and objects in a way which was as simple as possible. The notation used here is statically typed. Its type theoretic foundation can be developed along the lines of system F 2 with subtyping [CMMS94] Local type inference [PT98, OZZ00] is used to reduce the amount of type annotations which needs to be given explicitly. Details of the type system are not covered in these notes. The rest of this paper is structured as follows. Section 1 introduces a purely functional subset of functional nets. Section 2 presents the full ....

Benjamin C. Pierce and David N. Turner. Local type inference. In Proc. 25th ACM Symposium on Principles of Programming Languages, 1998.

....with a polymorphic type system with F bounded quantification[CCH 89] As a first approximation, it can be understood as a variant of system F # # , extended with (nominal) records and mutable variables. We often elide redundant type information, assuming that a local type inference system [PT98, OZZ01] is available to supply the omitted annotations. A record type definition introduces a new record with given supertypes, type components and value components. Here is an example: type R = S type t = Int val f: t # t val z: t This definition introduces the type R as a subtype ....

Benjamin C. Pierce and David N. Turner. Local type inference. In Proc. 25th ACM Symposium on Principles of Programming Languages, 1998.

....allows certain type annotations to be elided in actual programs. Local type inference determines types by a combination of type propagation and local constraint solving, rather than by global constraint solving. We refine the previously existing local type inference system of Pierce and Turner[PT98] by allowing partial type information to be propagated. This is expressed by coloring types to indicate propagation directions. Propagating partial type information allows us to omit type annotations for the visitor pattern, the analogue of pattern matching in languages without sum types. 1 ....

....annotations for the visitor pattern, the analogue of pattern matching in languages without sum types. 1 Introduction Many modern programming languages are based on type systems which combine a notion of objects and subtyping with parametric polymorphism [Str91, Mey92, CDG 92, NC97, OW97, PT98, BOSW98] A popular basis for such type systems is F# , the second order lambda calculus with subtyping. While F# is an excellent basis for explaining the abstract type structure of programs, it is less suitable as a kernel language for concrete source programs, because of the excessive amount of ....

[Article contains additional citation context not shown here]

Benjamin C. Pierce and David N. Turner. Local type inference. In Proc. 25th ACM Symposium on Principles of Programming Languages, pages 252--265, January 1998. 12

....inference system is based on HM(X) a theoretical framework described in [Lufer and Odersky, 1994] Two other proposals to add objects to ML, one by Dominic Duggan [Duggan, 1995] and the other named Object ML [Reppy and Riecke, 1996] provide weaker kinds of type inference. Local Type Inference [Pierce and Turner, 1997] is an attempt to provide partial type inference support for new languages such as Pizza [Odersky and Wadler, 1997] that include powerful type systems and higher order functions. Their system infers the types of local variables and anonymous blocks, while still requiring the programmer to provide ....

Benjamin C. Pierce and David N. Turner. Local Type Inference, Indiana University CSCI Technical Report #493, October 1997.

....to design a formalism for specifying such logics to the framework. The generic support will include a generalization of Nuprl s type annotation scheme [12] and support for user extensible type inference. For the latter, we will explore an approach based on Pierce and Turner s local type inference [17]. We designed a less principled version of this approach for Nuprl which has proved to be effective and is still in use. The implementation of the framework will not be from scratch; most of the code for it will be borrowed from the soon to be released version of Nuprl which features a highly ....

Benjamin C. Pierce and David N. Turner. Local type inference. In 25th ACM Symp. on Principles of Programming Languages, 1998.

....on syntactic support for the general pattern of mutually recursive F bounded type parameters, in hopes of achieving the same syntactic conciseness and programmer comprehensibility as well. 4.5. 6 Languages Offering Local Type Inference The work on local type inference in an extension of F [Pierce Turner 98] especially the local type argument synthesis, is very similar to inference of instantiating types in Cecil: they address a similar problem and use a similar inference algorithm. Their setting is different from Cecil s: they work within an impredicative type system whereas Cecil s is ....

Benjamin C. Pierce and David N. Turner. Local Type Inference. In Conference Record of the 25th ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages, pages 252-- 265, January 1998.

....and intersections [2, 15] Finally, we show how our type assignment system can be extended to include a value restricted form of parametric polymorphism. However, we do not show how to extend our source language, since a generalization of bi directional type checking to local type inference [13] in the presence of polymorphism, subtyping, intersections, and a value restriction does not appear straightforward. We close the introduction with a simple example that illustrates the unsoundness of intersection in the presence of mutable references. Similar counterexamples can be constructed ....

....infinite analogs of those in Section 2. In particular we include a value restriction on the introduction of polymorphism, and omit a rule for distributivity with the function type constructor. We do not consider type checking here, however we hope that some of the ideas from local type inference [13] for parametric polymorphism may be incorporated into the type checking algorithm of Section 3 to obtain a practical programming language. 4.1 Syntax We add type variables and universal quantification to the types of Section 2. The terms and values of the language are as before. Types A : A1 ....

[Article contains additional citation context not shown here]

B. C. Pierce and D. N. Turner. Local type inference. In The 25th ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages (POPL'98), pages 252--265, San Diego, California, 1998.

....or JoCaml[9] The notation we use here is statically typed, using a standard type system with recursive records, structural subtyping and polymorphism. A type theoretic foundation of the notation can be developed along the lines of system F 2 with subtyping [7] We use local type inference [24] to reduce the amount of type annotations which need to be given explicitly. Again, we omit the details here. The rest of this paper is structured as follows. Section 2 introduces a purely functional subset of functional nets. Section 3 presents the full formalism which supports concurrent ....

Benjamin C. Pierce and David N. Turner. Local type inference. In Proc. 25th ACM Symposium on Principles of Programming Languages, 1998.

....free. Firstly, the type checker and the inference algorithm can now be integrated, considering type inference as a special 6 Alternatively, we could have let the contextual information flow in the other direction and extract the expected type of e 0 from the type inferred for e, as is done in [PT98] But this scheme would not make any good use of the parameter , and since we most often do not need to help the inference algorithm finding types for the anonymous lambda abstractions that might appear in e 0 (which is the motivation behind the choice in [PT98] it makes more sense to ....

.... inferred for e, as is done in [PT98] But this scheme would not make any good use of the parameter , and since we most often do not need to help the inference algorithm finding types for the anonymous lambda abstractions that might appear in e 0 (which is the motivation behind the choice in [PT98] it makes more sense to collect as much information as possible before instantiating the generally polymorphic type of e. case of checking where the expected type is a unique instantiation variable. Secondly, it becomes possible to let the programmer exploit the use of partial type signatures, ....

[Article contains additional citation context not shown here]

Benjamin C. Pierce and David N. Turner. Local type inference. In Conference Record of POPL '98: the 25th ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages, 1998.

....will work with a non polymorphic version of Pict; later, after more theoretical study of the interplay between session types and polymorphism, we will integrate session types with the full Pict type system including polymorphism. The Pict compiler uses a powerful partial type inference technique [17], and it will be interesting to see how it can be extended to handle session types. Because of the value of explicit session types as specifications, we might not want to allow the programmer to omit them completely; however, automatic inference of, for example, some usage annotations will ....

B. C. Pierce and D. N. Turner. Local type inference. In Proceedings, 25th ACM Symposium on Principles of Programming Languages, 1998.

....for virtual types [Torgersen 98] His code to express the typechecking situations in Section 3 looks quite similar to ours. He also notes the need for syntactic sugars to make code more concise in case of large class hierarchies. The recent work on local type inference in an extension of F [Pierce Turner 98] especially the local type argument synthesis, is very similar to our inference of instantiating types: they address a similar problem and use a similar inference algorithm. Their setting is different from ours: they work within an impredicative type system whereas ours is essentially ....

Benjamin C. Pierce and David N. Turner. Local Type Inference. In Conference Record of the 25th ACM SIGPLAN -SIGACT Symposium on Principles of Programming Languages, pages 252--265, January 1998.

....we write S :X :T to mean the value of X must lie somewhere between S and T. In such a scheme, a completely unconstrained parameter is bounded below by Bot and above by Top. This application to type inference was the one that actually motivated the present study. It is described in detail in [PT98, PT97]. Unfortunately, the addition of Bot significantly complicates the meta theoretic properties of the system. Most significantly, in a type of the form All(X :Bot)T, the variable X is actually a synonym for Bot inside T, since X is a subtype of Bot by assumption and Bot is a subtype of X by the ....

Benjamin C. Pierce and David N. Turner. Local type inference. In Conference Record of POPL '98: the 25th ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages, 1998. To appear. Full version available as Indiana University CSCI technical report #493.

....metatheory. A sketch of such an extension is in progress. We are experimenting with a prototype implementation of the datatype mechanism described here in the context of the Pict language [PT97] An important goal is to investigate the interactions of datatypes with local type inference [PT98] One of the advantages of Rec(X)T style presentations of recursive types over the explicitly named formulation we have chosen here is that it is not clear how the latter should behave in combination with a module system. For instance, what should be the minimal signature of a module ....

Benjamin C. Pierce and David N. Turner. Local type inference. In Conference Record of POPL '98: the 25th ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages, 1998. Full version available as Indiana University CSCI technical report #493.

No context found.

Pierce, B. and Turner, D. Local Type Inference. in: Proceedings of the 25th Annual ACM SIGPLANSIGACT Symposium on Principles of Programming Languages, San Diego, January 19--21. 1998.

No context found.

Benjamin C. Pierce and David N. Turner. Local type inference. In ACM Symp. Principles of Programming Languages, pages 252--265, 1998. Full version in ACM Trans. Prog. Lang. Sys., 22(1):1--44, 2000.

No context found.

Benjamin C. Pierce and David N. Turner. Local type inference. In Proc. 25th ACM Symp. Principles of Programming Languages, 1998.

No context found.

B. C. Pierce and D. N. Turner. Local type inference. In The 25th ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages (POPL'98), pages 252--265, San Diego, California, 1998.

No context found.

Benjamin C. Pierce and David N. Turner. Local Type Inference. In Symposium on Principles of Programming Languages, pages 252--265, San Diego CA, USA, 1998. ACM.

No context found.

Pierce, B. and D. Turner (1998). Local type inference. In Proceedings of the 25th Annual ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages, pp. 252--265.

No context found.

Benjamin C. Pierce and David N. Turner. Local type inference. In ACM Symp. Principles of Programming Languages, pages 252--265, 1998. Full version in ACM Trans. Prog. Lang. Sys., 22(1):1--44, 2000.

No context found.

Benjamin C. Pierce and David N. Turner. Local type inference. In Conference Record of POPL 98: The 25TH ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages, San Diego, California, pages 252--265, New York, NY, 1998.

No context found.

Benjamin C. Pierce and David N. Turner. Local type inference. In ACM Symp. Principles of Programming Languages, pages 252--265, 1998.

No context found.

B. C. Pierce and D. N. Turner. Local type inference. In Symposium on Principles of Programming Languages, 1998.

No context found.

B. C. Pierce and D. N. Turner. Local type inference. In Symposium on Principles of Programming Languages, 1998.

No context found.

Benjamin C. Pierce and David N. Turner. Local type inference. In Proceedings of the 25th ACM Conference on Principles of Programming Languages, 1998. Full version available as Indiana University CSCI Technical Report 493.

No context found.

B. C. Pierce and D. N. Turner. Local Type Inference. In Proceedings of the 25th Annual ACM SIGPLANSIGACT Symposium on Principles of Programming Languages, pages 252--265, San Diego, California, Jan. 1998. 18 a, b, c

No context found.

Benjamin C. Pierce and David N. Turner. Local type inference. In 25th ACM Symposium on Principles of Programming Languages (POPL'98), pages 252--265, San Diego, January 1998. ACM.

No context found.

Benjamin C. Pierce and David N. Turner. Local type inference. In ACM Symp. Principles of Programming Languages, pages 252--265, 1998.

No context found.

Benjamin C. Pierce and David N. Turner. Local type inference. In Conference Record of the 25th Symposium on Principles of Programming Languages (POPL'98), 1998. Full version in ACM Transactions on Programming Languages and Systems (TOPLAS), 22(1), January 2000, pp. 1--44.

No context found.

Benjamin C. Pierce and David N. Turner. Local type inference. In Proc. 25th ACM Symp. Principles of Programming Languages, 1998.

No context found.

Benjamin C. Pierce and David N. Turner. Local type inference. In ACM Symp. Principles of Programming Languages, pages 252--265, 1998. Full version in ACM Trans. Prog. Lang. Sys., 22(1):1--44, 2000.

No context found.

Benjamin Pierce and David Turner. Local type inference. In 25th ACM Symposium on Principles of Programming Languages, pages 252--265, San Diego, CA, January 1998.

No context found.

Benjamin C. Pierce and David N. Turner. Local type inference. In ACM Symp. Principles of Programming Languages, pages 252--265, 1998.

No context found.

Benjamin C. Pierce and David N. Turner. Local type inference. In Proceedings of the 25th ACM Conference on Principles of Programming Languages, 1998. Full version in ACM Transactions on Programming Languages and Systems (TOPLAS), 22(1), January 2000, pp. 1--44.

No context found.

Benjamin C. Pierce and David N. Turner. Local type inference. In Proc. 25th ACM Symp. Principles of Programming Languages, 1998.

No context found.

B. C. Pierce and D. N. Turner. Local type inference. In POPL 98: The 25TH ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages, San Diego, California, 1998.

No context found.

Benjamin C. Pierce and David N. Turner. Local type inference. In Proceedings of the 25th ACM Conference on Principles of Programming Languages, 1998. Full version available as Indiana University CSCI Technical Report 493.

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