Torben Amtoft and Franklyn Turbak. Faithful translations between polyvariant 23 flows and polymorphic types. In Programming Languages & Systems, 9th European Symp. Programming, volume 1782 of LNCS, pages 26--40. SpringerVerlag, 2000.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....properties, but over the last twenty years the scope of research on intersection types has broadened. Intersection types have many unique advantages over types. They can characterize the behavior of # terms more precisely, and can be used to express exactly the results of many program analyses [13, 1, 25, 26]. Type polymorphism with intersection types is also more flexible. For example, Urzyczyn [20] proved the # term (#x.z(x(#fu.fu) x(#vg.gv) #y.yyy) to be untypable in the system F# , considered to be the most powerful type system with measured by the set of pure # terms it can type. In ....

....# . #z : # . xy, xz) Even this extension of Pierce s approach would still not meet our needs. First, this approach needs intersection types to be associative, commutative, and idempotent (ACI) Recent research suggests that non ACI intersection types are needed to faithfully encode flow analyses [1]. Second, this approach arranges the type information inconveniently because it must be found from enclosing type variable bindings by a tree walking process. This is bad for flow based transformations, which reference arbitrary subterms from distant locations. Third, reasoning about typed terms ....

T. Amtoft, F. Turbak. Faithful translations between polyvariant flows and polymorphic types. In Programming Languages & Systems, 9th European Symp. Programming, vol. 1782 of LNCS, pp. 26--40. Springer-Verlag, 2000.

....types has broadened. Van Bakel has written a good summary of the early research [vB95] Intersection types have many unique advantages over types. First, they can characterize the behavior of # terms more precisely, and can be used to express exactly the results of many program analyses [PP01, AT00, WDMT97, WDMT02]. In particular, intersection types can give more detailed information about all the contexts in which a function is used. Type polymorphism with intersection types is also more flexible. For example, Urzyczyn [Urz97] proved the # term (#x.z(x(#fu.fu) x(#vg.gv) #y.yyy) # This work was partly ....

....# . #z : # . xy, xz) Even this extension of Pierce s approach would still not meet our needs. First, this approach needs intersection types to be associative, commutative, and idempotent (ACI) Recent research suggests that non ACI intersection types are needed to faithfully encode flow analyses [AT00]. Second, this approach arranges the type information inconveniently because it must be found from enclosing type variable bindings by a tree walking process. This is bad for flow based transformations, which reference arbitrary subterms from distant locations. Third, reasoning about typed terms ....

Torben Amtoft and Franklyn Turbak. Faithful translations between polyvariant flows and polymorphic types. In Programming Languages & Systems, 9th European Symp. Programming, volume 1782 of LNCS, pages 26--40. SpringerVerlag, 2000.

No context found.

T. Amtoft and F. Turbak. Faithful translations between polyvariant flows and polymorphic types. In ESOP '00 [ESOP00], pp. 26--40.

No context found.

T. Amtoft and F. Turbak. Faithful translations between polyvariant flows and polymorphic types. Technical Report BUCS-TR-2000-01, Comp. Sci. Dept., Boston Univ., 2000.

No context found.

T. Amtoft and F. Turbak. Faithful translations between polyvariant flows and polymorphic types. In ESOP '00 [ESOP00], pp. 26--40.

No context found.

T. Amtoft and F. Turbak. Faithful translations between polyvariant flows and polymorphic types. Technical Report BUCS-TR-2000-01, Comp. Sci. Dept., Boston Univ., 2000.

....somehow contains multiple flow type annotations. For example, using the notation of [Pie91] # id could be expressed as something like: for # # int, real .#x # .x. Although this notation is more compact, it makes type information less accessible and can be tricky to adapt to more complex situations [WDMT0X] We have made preliminary investigations into other representations, e.g. one based on the skeletons and substitutions of [KW99] Based on the empirical results presented here, we believe that developing a non duplicating representation of CIL may be not critical (though it may still be ....

....to the precision of the approximation. In this paper, we present data from two of these: 1. The typed source split analysis is a variant of Banerjee s [Ban97] modified for shallow subtyping [WDMT0X] the use of shallow subtyping makes it slightly less precise than the combination of monomorphization and 0CFA analysis. It introduces virtual tuples and virtual projections but neither virtual variants nor virtual case forms. 2. The min type respecting analysis is the least precise flow analysis that is still type correct (cf. JWW97] It conflates the flow information on all values of the same flow ....

[Article contains additional citation context not shown here]

T. Amtoft and F. Turbak. Faithful translations between polyvariant flows and polymorphic types. In ESOP '00 [ESOP00], pp. 26--40.

.... Computer Science Department Wellesley College Wellesley, MA 02482, U.S.A. fturbak wellesley.edu http: cs.wellesley.edu fturbak Boston University Computer Science Department Draft of Technical Report BUCS TR 2001 XX July 31, 2001 This document is a draft of a technical report that expands upon [AT00] and is in preparation for submission to ACM Transactions on Programming Languages and Systems. # Supported by NSF grant EIA 9806745. Supported by NSF grant EIA 9806747. 1 Abstract Recent work has shown equivalences between various type systems and flow logics. Ideally, the translations ....

....are all finite, that u y # u u # y , and that for all q #dom(u xy ) it holds that either u xy .q =u x .q or u xy .q =u y .q. This example also illustrates that the I tags of a function reaching a call site need not match the sink label at that call site. For example, the app witness at application site 0 is # q g ## x, y # and the app witness at application site 2 within D #gy is q x ## 3 . The app witness e#ectively allows each copy of an application term in a derivation to have its own abstract sink labels that are independent of those from any other copy. 3.4 Semantic ....

[Article contains additional citation context not shown here]

T. Amtoft and F. Turbak. Faithful translations between polyvariant flows and polymorphic types. In ESOP '00 [ESOP00], pp. 26--40.

....somehow contains multiple flow type annotations. For example, using the notation of [Pie91] # id could be expressed as something like: for # # int, real .#x # .x. Although this notation is more compact, it makes type information less accessible and can be tricky to adapt to more complex situations [WDMT0X] We have made preliminary investigations into other representations, e.g. one based on the skeletons and substitutions of [KW99] Based on the empirical results presented here, we believe that developing a non duplicating representation of CIL may be not critical (though it may still be ....

....to the precision of the approximation. In this paper, we present data from two of these: 1. The typed source split analysis is a variant of Banerjee s [Ban97] modified for shallow subtyping [WDMT0X] the use of shallow subtyping makes it slightly less precise than the combination of monomorphization and 0CFA analysis. It introduces virtual tuples and virtual projections but neither virtual variants nor virtual case forms. 2. The min type respecting analysis is the least precise flow analysis that is still type correct (cf. JWW97] It conflates the flow information on all values of the same flow ....

[Article contains additional citation context not shown here]

T. Amtoft and F. Turbak. Faithful translations between polyvariant flows and polymorphic types. In ESOP '00 [ESOP00], pp. 26--40. 18

No context found.

Torben Amtoft and Franklyn Turbak. Faithful translations between polyvariant 23 flows and polymorphic types. In Programming Languages & Systems, 9th European Symp. Programming, volume 1782 of LNCS, pages 26--40. SpringerVerlag, 2000.

No context found.

T. Amtoft, F. Turbak. Faithful translations between polyvariant flows and polymorphic types. In Programming Languages & Systems, 9th European Symp. Programming, vol. 1782 of LNCS. Springer-Verlag, 2000.

No context found.

T. Amtoft, F. Turbak. Faithful translations between polyvariant flows and polymorphic types. In Programming Languages & Systems, 9th European Symp. Programming, vol. 1782 of LNCS. Springer-Verlag, 2000. 29

No context found.

T. Amtoft, F. Turbak. Faithful translations between polyvariant flows and polymorphic types. In Programming Languages & Systems, 9th European Symp. Programming, vol. 1782 of LNCS. Springer-Verlag, 2000.

No context found.

Torben Amtoft and Franklyn Turbak. Faithful translations between polyvariant 23 flows and polymorphic types. In Programming Languages & Systems, 9th European Symp. Programming, volume 1782 of LNCS, pages 26--40. SpringerVerlag, 2000.

No context found.

Torben Amtoft and Franklyn Turbak. Faithful translations between polyvariant flows and polymorphic types. In Proceedings of ESOP'00, European Symposium on Programming, pages 26--40. Springer-Verlag (LNCS 1782), 2000.

No context found.

Torben Amtoft and Franklyn Turbak. Faithful translations between polyvariant 23 flows and polymorphic types. In Programming Languages & Systems, 9th European Symp. Programming, volume 1782 of LNCS, pages 26--40. SpringerVerlag, 2000.

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