{"title":"形式化有类型部分求值器的正确性属性","authors":"Noriko Hirota, K. Asai","doi":"10.1145/2541568.2541572","DOIUrl":null,"url":null,"abstract":"This paper presents our experience of formalizing Danvy's type-directed partial evaluator (TDPE) for the call-by-name lambda calculus in the proof assistant Coq. Following the previous approach by Coquand and Ilik, we characterize TDPE as a composition of completeness and soundness theorems of typing rules with respect to the semantics. To show the correctness property of TDPE (i.e., TDPE preserves semantics), we further define a logical relation between residualizing and standard semantics, following Filinski. The use of parametric higher-order abstract syntax (PHOAS) leads to a simple formalization without being disturbed by fresh names created during TDPE. Because of the higher-order nature of PHOAS, it also requires us to prove manually a core property that corresponds to the main lemma of logical relations, which appears to be difficult to prove in Coq.","PeriodicalId":153056,"journal":{"name":"Programming Languages meets Program Verification","volume":"14 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2014-01-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":"{\"title\":\"Formalizing a correctness property of a type-directed partial evaluator\",\"authors\":\"Noriko Hirota, K. Asai\",\"doi\":\"10.1145/2541568.2541572\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"This paper presents our experience of formalizing Danvy's type-directed partial evaluator (TDPE) for the call-by-name lambda calculus in the proof assistant Coq. Following the previous approach by Coquand and Ilik, we characterize TDPE as a composition of completeness and soundness theorems of typing rules with respect to the semantics. To show the correctness property of TDPE (i.e., TDPE preserves semantics), we further define a logical relation between residualizing and standard semantics, following Filinski. The use of parametric higher-order abstract syntax (PHOAS) leads to a simple formalization without being disturbed by fresh names created during TDPE. Because of the higher-order nature of PHOAS, it also requires us to prove manually a core property that corresponds to the main lemma of logical relations, which appears to be difficult to prove in Coq.\",\"PeriodicalId\":153056,\"journal\":{\"name\":\"Programming Languages meets Program Verification\",\"volume\":\"14 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2014-01-11\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"2\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Programming Languages meets Program Verification\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1145/2541568.2541572\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Programming Languages meets Program Verification","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/2541568.2541572","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Formalizing a correctness property of a type-directed partial evaluator
This paper presents our experience of formalizing Danvy's type-directed partial evaluator (TDPE) for the call-by-name lambda calculus in the proof assistant Coq. Following the previous approach by Coquand and Ilik, we characterize TDPE as a composition of completeness and soundness theorems of typing rules with respect to the semantics. To show the correctness property of TDPE (i.e., TDPE preserves semantics), we further define a logical relation between residualizing and standard semantics, following Filinski. The use of parametric higher-order abstract syntax (PHOAS) leads to a simple formalization without being disturbed by fresh names created during TDPE. Because of the higher-order nature of PHOAS, it also requires us to prove manually a core property that corresponds to the main lemma of logical relations, which appears to be difficult to prove in Coq.