Case study
2026
BuildForge
A ground-up rebuild of the Path of Building character planner in Rust, keeping the original calculation engine so every number stays in exact parity.
- Client Personal project
- Discipline Desktop App
- Duration Ongoing
- Budget band Personal project
The problem
Rebuilding a tool a community already trusts carries one risk that outweighs all the others. If your numbers disagree with the original by even a little, nobody uses it, and you will never be sure which of you is wrong.
Reimplementing years of accumulated damage and defence calculations from scratch is how you guarantee that outcome.
What I did
A Rust workspace behind a Tauri shell. The interface and the passive skill tree are written fresh, because that is where a rebuild actually pays off.
The calculation engine is not rewritten. The original tool’s Lua engine is vendored and driven headlessly through mlua, so every damage and defence figure comes from the same code the community already relies on.
That is a deliberate refusal to be clever. Rewriting the engine would have been the interesting problem and the wrong decision.
Outcome
A planner whose numbers are in exact parity with the original by construction, rather than by testing and hoping.
Knowing which parts of a system to leave alone is most of the skill in a rebuild. It is the same judgement that decides whether a client’s project needs replacing or just fixing.