Users often don’t need Turing completeness in their DSL, since by definition they’re already intended to be limited in scope, and can get some real benefits by avoiding it, from more predictable resource usage to curtailing misuse. In my experience, most DSLs end up being Turing complete by default. While there are many factors that go into writing a good DSL, one that’s rarely discussed is how capable it needs to be. A good homegrown DSL can save you an enormous amount of time, but anyone who has had to interact with a bad one knows that it becomes an exercise in frustration. Initially, these are probably DSLs associated with tools others have written, such as Skylark rules or SQL queries, but as a project grows it can become useful to write your own DSL that’s suited to your specific needs. Almost every large software project ends up using a number of domain-specific languages (DSLs).
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |