Data-oriented programming (DOP) is a paradigm that models programs as a series of transformations on data, guided by four key principles.
An example of a GitHub crawler is used to explore how to model a complex page data type with various states and required operations.
Java records are the ideal choice for modeling immutable data, and compact constructors can be used to ensure deep immutability for collection fields.
Sealed interfaces create a closed type hierarchy, allowing you to precisely model "either/or" relationships between different record types representing distinct data states.
By combining a precise type hierarchy with validation in record constructors, you can prevent the creation of invalid object states within your domain model.
Operations are moved out of data records and implemented externally using switch expressions with pattern matching to deconstruct and act on the data.
Switching over sealed types is safe because the compiler enforces exhaustiveness, issuing a compile error if a new subtype is added but not handled.
Data-oriented programming provides a more ergonomic alternative to the classic visitor pattern and is useful for modeling complex return types.
tree-IT GmbH
Bad Neustadt an der Saale, Germany
Bitpanda
Vienna, Austria
29:16Balkrishna Rawool
48:42Dominik Przybysz
27:49Ron Veen
58:19Michael Koppmann
27:39Ron Veen
30:21Ron Veen
30:28Ron Veen
28:13Erick Wendel
.gif?w=240&auto=compress,format)
Jobs that call for the skills explored in this talk.
Opus
Manor Park, United Kingdom
MOIA GmbH
Berlin, Germany