A cryptarithmetic puzzle is presented where letters represent unique digits in a mathematical equation that must be solved.
A high-level solution is designed with a generator and evaluator, based on a grammar that defines expressions as constants, variables, additions, or multiplications.
Product types are implemented with records and sum types with sealed interfaces, which can then be deconstructed using pattern matching for `instanceof` and `switch`.
An expression is modeled as a sum type using a Java sealed interface with record implementations for constant, variable, addition, and multiplication.
A static `evaluate` function uses a pattern matching switch expression to recursively calculate the value of an expression tree, ensuring compile-time safety.
The main puzzle solver logic is built by creating a sealed interface for constraints and using a generator to check all combinations against them.
A unit test is created to assemble the puzzle's variables, expressions, and constraints, which is then run to find and verify the final solution.
The combination of algebraic data types for data modeling and pattern matching for behavior is defined as data-oriented programming, a powerful pattern for writing readable code.
28:24Nicolai Parlog
27:49Ron Veen
28:48Georg Dresler
58:19Michael Koppmann
27:39Ron Veen
30:21Ron Veen
38:07Lothar Schulz
30:28Ron Veen
.gif?w=240&auto=compress,format)
Jobs that call for the skills explored in this talk.
Joh. Berenberg, Gossler & Co. KG