Some of the work that reaches me isn’t from scratch; it is projects abandoned halfway. The first question is always the same: can it be saved, or should it be rewritten?
What gets reviewed in three days
- Is there a code repository, and is the history kept?
- What state is the database design in? Can the data be saved?
- How old are the dependencies, and are there security holes?
- How much of it actually works, and how much only sits on the screen?
- Do you hold the deployment pipeline and server access?
The “cheaper from scratch” threshold
If understanding the existing code takes longer than rewriting it, the threshold has been crossed. The clearest sign is a small change breaking three other places. The data side can usually be saved; the interface is usually rewritten.
How a takeover runs
First a written status report: what exists, what doesn’t, the risks and a recommendation. Then a gradual transition — the system keeps running while it is renewed piece by piece. Switching everything off and on in one night is the riskiest method.
In a stalled project the most expensive thing is not the code but the lost time; the longer the decision waits, the bigger the cost.
Was this useful?
Rate it — you can change your vote later.
4.7· 160 ratings
