A DSL for Payroll Calculations
summary
DATEV builds a domain specific language in MPS (on top of kernel f) that encodes payroll business logic, generates Java, and runs CI to deploy microservices; it handles 13 million payslips per month and transfers 170 million datasets per year, preserves four years of historical rules, manages roughly 3000 fields, supports versioned rules with a rule finder (no solver), unit tests in-model with a tracer, and can invoke the official XML-to-Java tax algorithm as a jar.
AI-written summary · about
chapters (15)
- 0:00Introduction and context
- 1:27Domain complexity challenges
- 3:34Need for a payroll DSL
- 6:44Technical requirements for DSL
- 8:28DSL demo and structure
- 14:14Testing and debugging rules
- 17:47Handling versioning
- 22:21Visualization and tooling
- 26:11Integrating external algorithms
- 29:20CI/CD deployment workflow
- 31:56Lessons learned and best practices
- 35:10Utilizing KernelF
- 38:47User experience and support
- 40:35Project challenges and benefits
- 43:29Q&A session
related talks
48mYesterday SAP for $10M, Today Java Microservices for ₽10M
23mBuilding scalable systems in a complex compliance world
57mPatterns of SaaS: Refactoring Databases
09mA DSL for My NoSQL Database
36mData pipelines engineering made simple with Scala
42mIT Around Data Management: Finance Industry
52mData Testing with HEDDA.IO
29mSelf service data governance