Ask HN: What's the state of the art on data mapping?
- XML -> JSON - RDF/XML -> JSON-LD - JSON -> HTML (rendering) - HTML -> JSON (scraping) - Relational DB -> Objects (ORM) - FOAF -> Schema.org (ontology mapping) - Photoshop -> Sketch - Imperial -> Metric - JPG -> PNG - REST -> GraphQL (proxy) - French -> English (translation) - TypeScript -> JavaScript (compiling) - XSL -> XSLX (upgrade) - Model -> ViewModel
It seems like for every new pair of data, we reinvent new tools and techniques from scratch. The best ideas in one domain have to be rediscovered in another. Manual effort is required to find the right conversion tool and execute the conversion. Data integration is a nightmare, and costs companies millions of dollar. Consumers are stuck in data silos and proprietary formats, and can't move their data out.
Don't we have some general theory on data mapping, that we could use to formally map between all these different structures and format, and automate the whole conversion process? Something along the lines of semantic mapping or parsing expression grammar?