| 作成者 |
|
|
|
|
|
| 本文言語 |
|
| 出版者 |
|
| 発行日 |
|
| 収録物名 |
|
| 巻 |
|
| 開始ページ |
|
| 終了ページ |
|
| 出版タイプ |
|
| アクセス権 |
|
| 関連DOI |
|
| 関連URI |
|
| 関連情報 |
|
| 概要 |
The typed Lambda-mu-calculus is known to be strongly normalizing and weakly Church-Rosser, and hence becomes confluent. In fact, Parigot formulated a parallel reduction to prove confluence of the type...d Lambda-mu-calculus by "Tait-and-Martin-Löf" method. However, the diamond property does not hold for his parallel reduction. The confluence for type-free Lambda-mu-calculus cannot be derived from that of the typed Lambda-mu-calculus and is not confirmed yet as far as we know. We analyze granularity of the reduction rules, and then introduce a new parallel reduction such that both renaming reduction and consecutive structural reductions are considered as one step parallel reduction. It is shown that the new formulation of parallel reduction has the diamond property, which yields a correct proof of the confluence for type free Lambda-mu-calculus. The diamond property of the new parallel reduction is also applicable to a call-by-value version of the Lambda-mu-calculus containing the symmetric structural reduction rule.続きを見る
|
Mendeley出力
