http://ayende.comAyende @ RahienCopyright (C) Ayende Rahien 2004 - 2021 (c) 202660One concept many people on the "business" end of our profession don't understand is sunk costs. Whatever resources have been used up to this point is gone and are never coming back, so they should look forward and say "what will all these bug fixes from this time forward cost me v. what benefits do I get by re-writing". If more people on the business side did that, life would be much easier. Then again, on the development side we often tend to underestimate the time it'd take to rebuild an applicaiton. The "Oh, that shouldn't take more than a few months" one off said to a manager without really thinking it through gets us in a lot of trouble. http://ayende.com/4223/hey-you-developer-learn-a-bit-of-business-speak#comment8http://ayende.com/4223/hey-you-developer-learn-a-bit-of-business-speak#comment8Fri, 02 Oct 2009 15:59:21 GMTRight. Sorry. http://ayende.com/4223/hey-you-developer-learn-a-bit-of-business-speak#comment7http://ayende.com/4223/hey-you-developer-learn-a-bit-of-business-speak#comment7Fri, 02 Oct 2009 12:50:49 GMTRafal, Pay attention to the context in which this was written, we aren't talking about rewriting the whole app. We are writing about rewriting a fragile _component_ inside an application http://ayende.com/4223/hey-you-developer-learn-a-bit-of-business-speak#comment6http://ayende.com/4223/hey-you-developer-learn-a-bit-of-business-speak#comment6Fri, 02 Oct 2009 12:48:46 GMTI reckon I only ship 1/3 of the code I write. The other 2/3rds consist of early iterations of the feature, bits of 10 minute prototypes, experimentation with language features in Notepad etc. My code is smaller and simpler (thus less buggy) by doing this. http://ayende.com/4223/hey-you-developer-learn-a-bit-of-business-speak#comment5http://ayende.com/4223/hey-you-developer-learn-a-bit-of-business-speak#comment5Fri, 02 Oct 2009 10:21:13 GMTUsing the issue database or specifying lots of new features are nice ways of convincing the business that a rewrite is the correct path. Yet i have personally experienced a situation where "the business" felt so heavely invested in a project that they just kept throwing time and money at bug fixes and hack extensions and simply did not (want to) realize that the product had to be restructured from the ground up to ever realy become a success. Ofcours in this case the "bad code quality" extended across the architectural design, all the way up to features beeing poorly thought out. Also i like to whine about rewrites :) http://ayende.com/4223/hey-you-developer-learn-a-bit-of-business-speak#comment4http://ayende.com/4223/hey-you-developer-learn-a-bit-of-business-speak#comment4Fri, 02 Oct 2009 09:13:34 GMT@Dave, I oversimplified a lot, just because I have heard too many programmers whining that their project is a crap and should be blown up and started from scratch. Your approach is rational and you are well prepared for convincing your co-workers that some unusual actions need to be taken - this is a completely different case. http://ayende.com/4223/hey-you-developer-learn-a-bit-of-business-speak#comment3http://ayende.com/4223/hey-you-developer-learn-a-bit-of-business-speak#comment3Fri, 02 Oct 2009 08:56:00 GMTOf course you're not gonna rewrite the full solution. The original post was about fragile code that often tend to produce bugs when it's altered because it's already changed so many times. Wanna talk about business speak? I just say it in statistics and graphs. Using the issue database I keep a sore of how many issues a certain feature or component get during it's lifetime, the time between bugs and how long it takes to resolve those issues. So when my graph shows management that the time between issues for a feature or component decreases and the time to solve it increases they ask me how I can reverse that trend. And that's the point I can convince management it's time to rewrite that feature or component. And remember: That all your tests pass doesn't mean your have high quality code, it only tells you that it works. http://ayende.com/4223/hey-you-developer-learn-a-bit-of-business-speak#comment2http://ayende.com/4223/hey-you-developer-learn-a-bit-of-business-speak#comment2Fri, 02 Oct 2009 07:16:11 GMTBut do you really think rewriting something from scratch just to fix some bugs or messy design is a good idea? Throw away everything and start with nothing but a faint idea an eagerness to work? No wonder manager's dont believe that, it's a nonsense... I'd rather collect ideas for new functionalities that are impossible with current architecture, or design a completely new GUI with another technology, and then decide to start from scratch because the gathered changes are so massive that it's easier to rebuild everything than modify existing code. http://ayende.com/4223/hey-you-developer-learn-a-bit-of-business-speak#comment1http://ayende.com/4223/hey-you-developer-learn-a-bit-of-business-speak#comment1Fri, 02 Oct 2009 06:20:19 GMT