Comments on: Greenfield Already! When is enough enough? http://halfbakedbits.com/2009/01/greenfield-already-when-is-enough-enough/ Small Ideas Factory Sat, 15 May 2010 10:38:11 +0000 hourly 1 http://wordpress.org/?v=3.0 By: Matt Farina http://halfbakedbits.com/2009/01/greenfield-already-when-is-enough-enough/comment-page-1/#comment-290 Matt Farina Fri, 09 Jan 2009 17:41:03 +0000 http://halfbakedbits.com/?p=204#comment-290 <p>OK, I agree that all codebases are crap. But, depending on the context of the crap there is a time to rewrite it so it's less crappie. The crappiness level to rewrite for me is when it's cost effective to your business objectives. For example, if your system requires a lot of support to teach users how to use it because it's not intuitive you might rewrite it (or that part) because the overall cost will be less in the end and it will be easier for users.</p> <p>Also, making a more extensible architecture in and of itself is not a reason to rewrite. Some things just don't need that and adding it can go against the business goals. If you make something more extensible but never use it though you get a performance hit that's no good.</p> <p>Now, you should never rewrite just to rewrite. For example, if you have crap code that works beautifully but you can't stand to look at it you should leave it be. It may be ugly but it works. Rewriting code means dealing with bugs and all kinds of other goodies. The decision to do that shouldn't happen lightly. It can hurt like a truck running over your foot.</p> OK, I agree that all codebases are crap. But, depending on the context of the crap there is a time to rewrite it so it’s less crappie. The crappiness level to rewrite for me is when it’s cost effective to your business objectives. For example, if your system requires a lot of support to teach users how to use it because it’s not intuitive you might rewrite it (or that part) because the overall cost will be less in the end and it will be easier for users.

Also, making a more extensible architecture in and of itself is not a reason to rewrite. Some things just don’t need that and adding it can go against the business goals. If you make something more extensible but never use it though you get a performance hit that’s no good.

Now, you should never rewrite just to rewrite. For example, if you have crap code that works beautifully but you can’t stand to look at it you should leave it be. It may be ugly but it works. Rewriting code means dealing with bugs and all kinds of other goodies. The decision to do that shouldn’t happen lightly. It can hurt like a truck running over your foot.

]]> By: Brett Veenstra http://halfbakedbits.com/2009/01/greenfield-already-when-is-enough-enough/comment-page-1/#comment-289 Brett Veenstra Fri, 09 Jan 2009 17:32:10 +0000 http://halfbakedbits.com/?p=204#comment-289 <p>I agree when the codebase is crap, although that can quickly get muddy. I generally think all codebases are crap, save for my flavor of the day. :)</p> <p>More extensible architecture? I'd like to hear more detail before I weigh in. Currently, I'd disagree, as I stated above:</p> <blockquote> <p>What we get “paid for” is actually dealing with a lot of crap code, some written by us, the rest written by others.</p> </blockquote> <p>What kinds of things did you have in mind?</p> I agree when the codebase is crap, although that can quickly get muddy. I generally think all codebases are crap, save for my flavor of the day. :)

More extensible architecture? I’d like to hear more detail before I weigh in. Currently, I’d disagree, as I stated above:

What we get “paid for” is actually dealing with a lot of crap code, some written by us, the rest written by others.

What kinds of things did you have in mind?

]]> By: Matt Farina http://halfbakedbits.com/2009/01/greenfield-already-when-is-enough-enough/comment-page-1/#comment-288 Matt Farina Fri, 09 Jan 2009 13:35:59 +0000 http://halfbakedbits.com/?p=204#comment-288 <p>Two more cases: - The original codebase was crap. - The codebase needs a more extensible architecture (tied to the first one in some cases).</p> <p>I recently rewrote a codebase because it was that bad. I usually say leave it if it works but there is a point where it's more cost effective to rewrite a codebase rather than support it.</p> <p>Joomla! and it's 1.5 release is an example of the 2nd point.</p> Two more cases: - The original codebase was crap. - The codebase needs a more extensible architecture (tied to the first one in some cases).

I recently rewrote a codebase because it was that bad. I usually say leave it if it works but there is a point where it’s more cost effective to rewrite a codebase rather than support it.

Joomla! and it’s 1.5 release is an example of the 2nd point.

]]>