3 ms·
When is a BIG Rewrite the answer?
- lugg 12y agoOver the years my opinion on this sways between never, and whenever you feel like it. Never: its far too risky, usually ends up not fixing half of the issues in the first place. Better off taking scorched earth refactor approach at the class / file level than system level. Whenever: the more often you do something, the better at it you should get, the faster you will get, and the better your designs will get. I think there really is likely some middle ground in there. The one thing I caution against is not to refactor while doing bug fixes unless it is part of fixing the bug. Overcomplicates the situation dramatically, make a note to refactor next time you're replacing a feature / have some downtime (lol downtime)
- dragonwriter 12y agoWhen you can't do a component-wise rewrite and the cost of not fixing the currently-inadequate components justifies the cost and risk of a big rewrite. But make sure when you do it you build with a loosely coupled architecture so that this is never an issue again -- any system that will force you into deciding when a big rewrite is necessary is badly designed.