This repository is a practical before/after example for modernizing a 10+ year-old public-sector admin system.
The goal is not to redesign everything at once. The focus is to reduce user resistance, preserve familiar work patterns, and improve consultation management and map-based selection flows for real business users.
- Target: legacy admin consultation and map-based workflow screens
- Main users: field operators and public-sector staff, especially users in their 40s to 60s
- Goal: reduce friction, improve task efficiency, and lower input errors
- Core methods: conversational consultation UI, OpenLayers map selection, report automation
- Developers modernizing old Java/JSP admin systems
- Teams looking for UI/UX examples that respect existing user habits
- Portfolio builders who want to explain a practical before/after redesign
- Conversational consultation management UI
- Map-based parcel selection to reduce manual input errors
- Automated operation report flow
- Performance notes for PL/SQL and query optimization examples
- Install Java 11+ and Maven.
- Build from the project root.
mvn clean package- Check sample DB/config files under
resources/when available.
The gallery below shows how the interface changes from dense legacy screens to more guided task flows.
- Keep base text around 16px to 18px for readability.
- Use 44px to 48px minimum control heights to reduce click mistakes.
- Use explicit labels and step-by-step progressive disclosure.
- Preserve enough contrast between text and backgrounds.
- Explain both the cause and next action in error messages.
:root {
--font-size-base: 18px;
--line-height: 1.4;
--button-min-height: 48px;
--space-section: 24px;
--accent-color: #2b7cff;
}- Report bugs or setup problems in Issues.
- Share public-sector or legacy UI improvement ideas in Discussions.
- Documentation, accessibility, and screenshot explanation improvements are welcome as pull requests.
No explicit license is currently provided. If this project is intended for open reuse, adding MIT or Apache-2.0 is recommended.





