Your team wants to rewrite a legacy service. How do you evaluate the proposal?

Summary

Look at evidence — where incidents and change costs concentrate — weigh the business cost of a feature freeze and the typical risks of rewrites, and prefer an incremental strangler approach that delivers value along the way unless the evidence strongly favours a rewrite.

Longer Answer

Rewrites usually take longer than estimated and lose business rules encoded in old code. Questions to ask: what specific problems would it solve? Are they concentrated in a few modules? Can we add tests and extract those modules incrementally? What happens to the roadmap meanwhile? Record the decision in an ADR, give the proposers ownership of the new parts, and set a revisit trigger.

Key Points

• Evidence of where pain concentrates
• Business cost of a freeze; rewrite risks
• Incremental strangler approach by default

Deep dive: Capstone: Rewrite or Refactor

What interviewers look for

Balanced, evidence-based judgement and stakeholder handling.

Red flags

• 'Rewrites are always bad' or 'always good'
• Deciding on code aesthetics