JavaRosa currently uses kxml to parse XML. Certain operations perform poorly O(n²), drastically slowing down parsing large documents. A successful proposal for this project must:
- Objectively benchmark poorly performing operations to identify root causes and to be able to measure improvements
- Evaluate potential changes for risk/reward trade-offs
- Sequence changes to ensure high-value changes can be safely made (e.g., increased test coverage, comparing output, feature toggles)
The selected student will implement the proposed changes.