## Labelled Reductions, Runtime Errors, and Operational Subsumption (1997)

Venue: | of Lecture Notes in Computer Science |

Citations: | 6 - 1 self |

### BibTeX

@INPROCEEDINGS{Dami97labelledreductions,,

author = {Laurent Dami},

title = {Labelled Reductions, Runtime Errors, and Operational Subsumption},

booktitle = {of Lecture Notes in Computer Science},

year = {1997},

pages = {782--793},

publisher = {Springer}

}

### OpenURL

### Abstract

Introduction Consider the "name-switching" function F def = x:fl 1 = x:l 2 ; l 2 = x:l 1 g in a - calculus with records. Most type systems would reject program (Ffl 1 = 3g):l 2 because the type of F is fl 1 : X; l 2 : Y g ! fl 2 : Y; l 1 : Xg and fl 1 : X; l 2 : Y g cannot be unified with fl 1 : Intg, the type of the record argument. However this program reduces to 3 without error. This shows that the common notion of "erroneous" terms, as implemented in most typed languages, is sometimes

### Citations

1700 |
Java Language Specification
- GOSLING, JOY, et al.
- 2005
(Show Context)
Citation Context ... contrast, exceptions correspond to other abnormal conditions such as failure to open a file, or indexing an array out of its declared bounds. This distinction is exactly the one of the Java language =-=[16]-=-, where the Error class characterizes uncatchable errors, while RuntimeException characterizes catchable exceptions. Example 2.9 A language containing constructs isnat, islam, ispr; : : : for identify... |

1163 | The Lambda Calculus Its Syntax and Semantics - Barendregt - 1984 |

923 | A Theory of Objects - Abadi, Cardelli - 1996 |

227 | Data types as lattices - Scott - 1976 |

220 | Call-by-name, call-by-value, and the *-calculus - Plotkin - 1975 |

149 | Operations on records - Cardelli, Mitchell - 1994 |

128 | An Ideal Model for Recursive Polymorphic Types - MacQueen, Plotkin, et al. - 1986 |

123 | Inheritance as implicit coercion - Breazu-Tannen, Coquand, et al. - 1991 |

117 | Type inference for record concatenation and multiple inheritance
- Wand
- 1991
(Show Context)
Citation Context ...ance requires an operation to extend or modify some fields. We write a( )l = b for the record a in which field l has been added or overwritten with value b; this is exactly like the with construct of =-=[29, 33]-=-. The set of record values does not change, but the set of head normal forms does: the extension construct introduces new head normal forms like x:x( )l = b. Interestingly, the status of the empty rec... |

100 | Equality in lazy computation systems - Howe - 1989 |

87 |
A modest model of records, inheritance and bounded quantification
- Bruce, Longo
- 1990
(Show Context)
Citation Context ...es equals the usual approximation ordering (Theorems 4.6, 5.6); its main interest is to directly interpret subtyping in a term model, which is simpler than the partial equivalence relations (PERs) of =-=[8] or the co-=-ercion functions of [7]. This is illustrated in Section 5 where the abstract framework is applied to a comparison of various -calculi. Our framework only applies to "uncatchable" errors, i.e... |

49 |
Two extensions of Curry’s type inference system
- Cardone, Coppo
(Show Context)
Citation Context ...pproximations are not bound to a particular language. Furthermore, we also propose in Section 6 an innovative application of such approximations as an alternative to the embedding-projection pairs of =-=[9]-=- for solving recursive type equations. Note. This is an expanded version of the extended abstract published in [13]. Remarks from Scott Smith, Manuel Serrano and from anonymous referees on a previous ... |

49 |
Tω as universal domain
- Plotkin
(Show Context)
Citation Context ...ave a simple structure which often eases mathematical treatement; nevertheless the majority of semantic studies now uses more complex models (various forms of cpo constructions) because of criticisms =-=[28, 6] related t-=-o the "over-defined" top element, which somehow implies the presence of multivalues (for example what is the meaning of an upper bound of the truth values tt and ff ?). Without entering the ... |

45 | Bisimilarity for a first-order calculus of objects with subtyping - Gordon, Rees |

37 | From Operational Semantics to Domain Theory - Mason, Smith, et al. - 1996 |

31 | A lambda-calculus for dynamic binding - Dami - 1998 |

27 | A Theory of Binding Structures and Applications to Rewriting - Talcott - 1993 |

21 | Full Abstraction and the Context Lemma - Jim, Meyer - 1996 |

16 | Faithful ideal models for recursive polymorphic types - Abadi, Pierce, et al. - 1991 |

11 | Theoretical aspects of object-oriented programming: Types, semantics, and language design - Gunter, Mitchell, et al. - 1994 |

9 |
Semantic Techniques for Deriving Coinductive Characterizations of Observational Equivalences for -calculi
- Lenisa
(Show Context)
Citation Context ...re the language to be fixed and cannot be done at an abstract level of presentation. However recent progress has been made towards generalizations of the context lemma; interested readers may consult =-=[21, 24]. This com-=-pletes the abstract presentation of the language family. In the following we will restrict our attention to languages satisfying most of the "good" properties defined above, namely stuck-fre... |

8 |
Records and variants as a natural extension of ML
- emy, D
- 1989
(Show Context)
Citation Context ...ance requires an operation to extend or modify some fields. We write a( )l = b for the record a in which field l has been added or overwritten with value b; this is exactly like the with construct of =-=[29, 33]-=-. The set of record values does not change, but the set of head normal forms does: the extension construct introduces new head normal forms like x:x( )l = b. Interestingly, the status of the empty rec... |

6 |
Can LCF be topped? Flat lattice models of typed -calculus
- Bloom
- 1990
(Show Context)
Citation Context ...ement, " is subsumed by any term and therefore is the top element in the subsumption ordering. In consequence the semantic structure is a lattice, like in the original work of Scott [31]. As note=-=d in [6]-=-, lattices have a simple structure which often eases mathematical treatement; nevertheless the majority of semantic studies now uses more complex models (various forms of cpo constructions) because of... |

5 | the *-calculus. Theoretical Computer Science - Call-by-name - 1975 |

5 |
et al. The Haskell Report and Haskell Tutorial
- Hudak
- 1992
(Show Context)
Citation Context ...heory [1] x:\Omega is a value different from\Omega . This theoretical distinction has important practical consequences, since a family of modern programming languages are now based on the lazy theory =-=[18]. We are intere-=-sted in similar distinctions with respect to another form of unsolvable terms, namely runtime errors such as (1+ "foo") or aWindow.computeSalary(). Let " denote such errors. Most type s... |

4 |
Fer-Jan de Vries. Meaningless terms in rewriting
- Kennaway, Oostrom
- 1999
(Show Context)
Citation Context ...ul notion of subsumption: otherwise since error generation is the basic observation, an exception handling construct would make all programs incomparable; similar requirements are made for example in =-=[10, 22]. Bec-=-ause of this requirement, " is subsumed by any term and therefore is the top element in the subsumption ordering. In consequence the semantic structure is a lattice, like in the original work of ... |

3 | Typing Record Concatenation for Free - emy - 1992 |

1 | Currently available at http://cuiwww.unige.ch/~dami - Report, Geneva - 1997 |

1 | van Oostrom and Femke van Raamsdonk. Combinatory reduction systems: introduction and survey - Klop, Vincent - 1993 |

1 | Boudol. Lambda-Calculi for (Strict) Parallel Functions - erard - 1994 |

1 |
Pierre-Louis Curien and Matthias Felleisen. Fully Abstract Semantics for Observably Sequential Languages
- Cartwright
- 1994
(Show Context)
Citation Context ...ul notion of subsumption: otherwise since error generation is the basic observation, an exception handling construct would make all programs incomparable; similar requirements are made for example in =-=[10, 22]. Bec-=-ause of this requirement, " is subsumed by any term and therefore is the top element in the subsumption ordering. In consequence the semantic structure is a lattice, like in the original work of ... |

1 |
A Comparison of Record Calculi
- Dami
- 1997
(Show Context)
Citation Context ...to check a restricted class of "applicative contexts" and the proof of the laws above becomes easy. Complete developments and more detailed discussions for record extension constructs can be=-= found in [12, 14]-=-. 6 Interpreting Types This section illustrates the usefulness of both subsumption and labelled reductions for the semantics of types : subsumption is a natural foundation for interpreting subtyping, ... |