Refactoring vs. Rewriting

Question

With legacy code, how do you know when it's best to rewrite over refactoring?

Answer 1

One factor to look at: would it take less time to rewrite than to disect and refactor the code?

Answer 2

Discussed extensively:

http://stackoverflow.com/questions/146124/what-are-the-limitations-of-refactoring

http://stackoverflow.com/questions/144734/when-is-it-good-if-ever-to-scrap-production-code-and-start-over

http://stackoverflow.com/questions/35233/rewrite-or-repair

http://stackoverflow.com/questions/21857/when-do-you-know-its-time-to-rewrite-an-application

http://stackoverflow.com/questions/78806/refactor-mercilessly-or-build-one-to-throw-away

Answer 3

And here:

http://stackoverflow.com/questions/35233/rewrite-or-repair

Answer 4

I think the question is, how do you tell which would take less time, a complete rewrite or a refactor of existing code.

The answer probably has something to do with a different question: can you, upon glancing down the code, immediately see what chunks can be separated out into usable modules, and upon testing a positive answer by trying a few small cases, did the program break?

If you could not immediately see a way, or the program broke when you tried a few small cases, it was probably badly enough written that it needs a rewrite.