by Till­mann Ren­del, Julia Trieflinger, and Klaus Os­ter­mann

In Pro­ceed­ings of In­ter­na­tional Con­fer­ence on Func­tional Pro­gram­ming, pages 269–279. ACM, 2015.

This pub­li­ca­tion is re­lated to the Uroboro re­search pro­ject.

Ab­stract

De­func­tion­al­iza­tion and re­func­tion­al­iza­tion es­tab­lish a cor­re­spon­dence be­tween first-class func­tions and pat­tern match­ing, but the cor­re­spon­dence is not sym­met­ric: Not all uses of pat­tern match­ing can be au­to­mat­i­cally re­func­tion­al­ized to uses of higher-or­der func­tions. To rem­edy this asym­me­try, we gen­er­al­ize from first-class func­tions to ar­bi­trary co­data. This leads us to full de­func­tion­al­iza­tion and re­func­tion­al­iza­tion be­tween a co­data lan­guage based on co­pat­tern match­ing and a data lan­guage based on pat­tern match­ing.

News