typesDVIA

typesDVIA is a framework for classifying and managing data types within digital information architectures. It provides a vocabulary and set of rules intended to improve interoperability between systems that exchange structured data. The framework treats types as first-class entities that can be described, combined, and constrained, enabling consistent schema generation and validation across tools.

Origin and scope: The concept emerged within the DVIA family of methodologies, an initiative focused on aligning data visualization, interaction design, and data modeling. typesDVIA extends DVIA by formalizing type taxonomies and their metadata, to support both humans and machines in understanding data semantics.

Core concepts: It defines several type categories: primitive types such as string, number, boolean and date; composite types such as records or objects with named fields; parametric types including generic containers like List<T> or Map<K,V>; and constrained types augmented with validation rules. Types may carry metadata such as descriptions, units or provenance and participate in relationships such as subtyping and compatibility.

Operations: Tools and workflows using typesDVIA can annotate datasets, generate schemas, perform validation, and translate types into target formats (for example JSON Schema, SQL DDL, or GraphQL). The framework also supports mapping rules to reconcile differences between source and target type systems.

Examples: A Person type with fields name (string) and birthDate (date); a List<Person>; and a union type CatalogItem = Book | DVD illustrate simple and polymorphic compositions.

Limitations and reception: The abstraction can be complex and may impose additional learning overhead. Adoption depends on ecosystem support, and inter-system mappings may require bespoke rules.

See also: data types, type theory, schema languages, data modeling.