Comments on: When Refactoring Code is a Waste of Time

By: Ivan Porto Carrero

Ivan Porto Carrero — Fri, 12 Sep 2008 03:53:14 +0000

I must admit that I’m guilty of refactoring but Idon’t go in and refactor the whole lot just the things I touch and when I see there are obvious problems with that code.

If I inherit a project from somebody else I tend to leave the things that work alone but when I find lots of duplication and other smells I cannot resist and refactor that stuff out and then still only in the classes I actually touch. Is that the refactoring cowboy you’re talking about? The last thing I want to do is piss the people around me off

I don’t care much for programming style as I can follow most styles of development not that I like all of them but they pay me so I’ll follow the standards.

By: Ivan Porto Carrero

Ivan Porto Carrero — Fri, 12 Sep 2008 03:53:00 +0000

I must admit that I’m guilty of refactoring but Idon’t go in and refactor the whole lot just the things I touch and when I see there are obvious problems with that code.
If I inherit a project from somebody else I tend to leave the things that work alone but when I find lots of duplication and other smells I cannot resist and refactor that stuff out and then still only in the classes I actually touch. Is that the refactoring cowboy you’re talking about? The last thing I want to do is piss the people around me off
I don’t care much for programming style as I can follow most styles of development not that I like all of them but they pay me so I’ll follow the standards.

By: Rob Bazinet

Rob Bazinet — Wed, 10 Sep 2008 22:57:50 +0000

@Dan it is a hard urge to resist but when the client is paying for it, it is hardly ever worth the effort. If the code works, be happy and move on.

@Micheal - same here, I bill clients and touching code that already works doesn't make them very happy.

Great comments guys. I expected to be flamed actually.

By: Rob Bazinet

Rob Bazinet — Wed, 10 Sep 2008 22:57:00 +0000

@Dan it is a hard urge to resist but when the client is paying for it, it is hardly ever worth the effort. If the code works, be happy and move on.
@Micheal – same here, I bill clients and touching code that already works doesn’t make them very happy.
Great comments guys. I expected to be flamed actually.

By: Michael Eaton

Michael Eaton — Wed, 10 Sep 2008 17:47:47 +0000

@Dan:

"It's a dull conclusion, but it comes down to judgment. Sometimes it's really worth it, even if it takes a while."

That's true. In my case though, I bill my clients by the hour so sometimes I can't jump in and refactor *everything* even if I want to.

By: Michael Eaton

Michael Eaton — Wed, 10 Sep 2008 17:47:00 +0000

@Dan:
"It’s a dull conclusion, but it comes down to judgment. Sometimes it’s really worth it, even if it takes a while."
That’s true. In my case though, I bill my clients by the hour so sometimes I can’t jump in and refactor *everything* even if I want to.

By: Dan Bernier

Dan Bernier — Wed, 10 Sep 2008 15:51:31 +0000

I once wanted to do this very badly. I'd inherited code w/ a TON of complexity, and a lot of it seemed needless. I wanted to rip through the code with a machete, and tear about half of it out, but I held back. After living with it for a week or two, I talked it out w/ a senior developer (I was pretty green), and he said sure, clean it up. The resulting code was -much- better. The original coder had been reading GoF, and made himself a gratuitous Pattern salad. With all the "extra" out of the way, you could see what the code was supposed to do.

OTOH, I'm guilty of making fastidious code-hygiene fixes, and introducing bugs. It's frustrating to take a much prettier bunch of code with a tricky bug, and revert it, because the original code at least works right, as far as you can tell.

It's a dull conclusion, but it comes down to judgment. Sometimes it's really worth it, even if it takes a while.